© 2009, 2010  www.theCUBEstudio.com


This review is a ‘living document’ and is being updated as new information becomes available. The link does not change, so to follow the progress, just hit the link again to get the latest version.


This review and comparison is a documentation of the author’s experience in using the mentioned products. This is a ‘user review’ and not a ‘buying guide’ nor is it intended as a representation of suitability of a particular purpose or an endorsement of any product.  The author was not compensated by any of the vendors.





Revision 00   07/01/2009  initial release  

Revision 01   07/05/2009  added info on Leadshine and Tek10

Revision 02   07/15/2009  added partial review info on Leadshine

Revision 03   07/16/2009  completed review of Leadshine.

Revision 04   07/18/2009  completed review and summarized.

Revision 05   07/22/2009  Added AMC and Granite to lineup (future tests).

Revision 06   07/24/2009  Added DMM to lineup. Updated summary with Rutex response.

Revision 07   07/27/2009  Updated pricing on Granite drive.

Revision 08   07/30/2009  Edited for publication.

Revision 09   08/03/2009  added Update to Rutex setup, deleted AMC drive from lineup.

Revision 10   08/05/2009  added vendor comment from Leadshine

Revision 11   08/07/2009  added CNCdrives Whale3 to lineup and started some review sections.

Revision 12   08/11/2009  added CNCdrives Dugong to lineup Vendor comment from Larken (Viper)

Revision 13   08/13/2009  removed DMM from lineup. Update to Viper research. Update CNCdrives tech support.  Highlighted all updates.

Revision 14   08/26/2009  finished review of Whale3 and Dugong drives .

Revision 15   09/02/2009  added update on Whale3 to Warranty and Tech Support section

Revision 16   09/04/2009  added update to CNCdrives in Researching section.

Revision 17   09/09/2009  added CNCdrive comment to Vendor Comments Section, Added update to Granite drives lineup description.

Revision 18   09/24/2009  added update on Rutex drive burn out. See Warranty section and conclusion section.

Revision 19   10/28/2009  Review finished and closed. Update to Granite drive review status. Vendors may continue to provide updates for inclusion.

Revision 20    05/25/2010 Review reopened to include Granite Devices VSD-EX NEW WINNER!

Revision 21    06/24/2010 Categorized drives and added ‘Best Choice’ for drives below 100V, added updates to Dugong and VSD-XE drives under various headings.

Revision 22    08/06/2010 Added AC Servo category and two Mitsubishi drives and motors. Updated ongoing issues with the Granite VSD-EX and Dugong drive.







- DC Brush Servo Drive Review and Comparison -



Gecko 340              ( 80V    20A   $148 + $14 ship from US)

UPDATE: Vendor provided info: replaced by Gecko320X as of 08/07/09 new price $114)


Rutex 2010    ( 100V  20A   $149 + $17 encoder interface + $11 ship from US)


Viper 95            ( 95V   12A    $174 + $6 install kit + $25 ship from Canada)

UPDATE: vendor provided info:  price reduction to $148


Leadshine 810       ( 80V    20A   $119 + $12 ship from US)


CNCdrives Whale3    ( 80V    20A   $107 + $22 ship from Hungary)


CNCdrives Dugong    ( 160V    35A   $170 + $22 ship from Hungary)


**Granite VSD-EX       ( 160V  40A  $290 + $23 install kit + $25 tuning cable + $60 shipping)


***DMM Dyn2-B-8      (removed from consideration due to small fixed following error) 


****TEK10                      ( 80V   25A       $309 + $?? ship)


*AMC DX15C08C GE1  (removed from consideration due to OEM CAN interface)


-        - AC industrial servo drives Review and comparison -


Mitsubishi MR-Cxx drive with matching HCPQxx Motor

Pricing info and performance review to be added soon


Mitsubishi MR-J2S-xxA drive with matching  HC-KFSxxK motor

Pricing and performance review to be added soon









* This drive is about $800 new, but readily available used for $50 to $250. It is included as a yardstick for comparison to the high end.

   The product that I have for review is an older OEM used drive which uses the CAN interface only, and a functional CAN-to-PC converter is over $100 putting the drive out of the price point being reviewed, so this drive will not be reviewed.

   Interestingly, AMC claims that ALL Digiflex drives have an RS232 interface, but obviously that is not the case. If a used or surplus drive with an RS232 interface becomes available, I will review it.


NOTE the following paragraph no longer is applicable: ** By request, Granite drive may be added later. The drive cost is $240 or $290 depending on model at current exchange rate. You also need a special cable $27 and an install kit $24. Shipping is from Finland or New Zealand and varies apparently between $12 and over $100. I have asked (on their CNCzone forum section) if there is a US vendor, no response. The drive is rumored to control two independent servo motors, but there is no official note of that on the company web site. The company states that stepper motor support is ‘preliminary’ so it makes sense to wait until that is completed before testing.


UPDATE: Granite drives has announced that their drive will support two independent servo motors with a user installable firmware update.

                  The product is being sold by Keling Technology but I was unable to determine if Keling has the new 160V models or the older models.


UPDATE: the Granite product is no longer listed on the Keling site. 


NOTE the following update is no longer applicable: UPDATE: I have decided not to review the Granite drive. I want to be very clear that this is not because of any specific problem with the product. My focus in this review was DC brush motors and the Granite drive (inclusive of ‘install kit’ and extraordinarily high shipping cost) is simply priced out of that specific market, in my opinion. The use of brushless drives either DC or AC is a different story and if that were to focus here, the higher cost of the Granite would not be an issue.


UPDATE:04/19/2010,  I will now review the Granite devices newest drive model.  Granite has upgraded the drive to 160V and has a new XE model that provides a peak of 40A for a single DC servo motor. In addition, the drive has been updated to handle two separate DC brush motors, effectively halving the cost of the product. In the works is new firmware that can operate a stepper in closed loop with an encoder and reach speeds as high as 7,000 RPM. This capability is new and as yet unreleased. The stepper functionality is in  beta testing. 



***By request, DMM drive may be added later. This AC brushless drive is bundled with a 250watt AC NEMA34 motor, cables and software for $250 plus $30 shipping from Canada to the  US. This product seems like a ‘must have’ product, but so far, I can find nobody at all who has used this product. Their web site is new and very sparse. They have only recently created a presence on Ebay and show no sales of the complete setup yet. It does have a money back guarantee, so I may review it anyway.


UPDATE: DMM drive will not be reviewed due to lack of interest together with some performance details noted in a close read of the specs, notably a very small fixed following error. 


UPDATE: 04/29/2010, DMM now has a 400watt motor. The following error is now adjustable, but remains at only 128.




****TEK10 will not be reviewed because the specs show it to be basically an overpriced Gecko. Trip limits are fixed at 128 or 1024 and it is tuned with an Oscilloscope and trim pots like the Gecko. No software tuning.





GECKO 340  the Gecko site is a treasure trove of information and calculations for both steppers and servos. There are several excellent tech papers on the site and the product manuals are available for download so that you can read about the products prior to purchase. I found no inaccurate information or deceptive claims on the Gecko site. The Gecko product line is extremely popular and information on any of the products is readily available on many forums.


RUTEX 2010 the Rutex site likewise has a great deal of general and specific information on their products and downloadable manuals and tech papers. Rutex suffers from a deserved bad reputation due to problems with the 2020 model drive and perhaps its predecessor as well that have a tendency to burn out far below rated capacities. Rutex acknowledges the problems with the previously mentioned drives, and claims  that the  2010 reviewed here does not have that problem, but it is very difficult to dig up any information, reviews or opinions on this model. I found no inaccurate information or deceptive claims on the Rutex site.


VIPER 95 the Viper site has very little in the way of general or specific information. When you click on the link to the Viper 95 manual, what you get is the manual for the Viper 100/200 with is a totally different product. Unless you notice this, you will be reviewing information that does not pertain to the product you are going to buy.


False Claims:


Larken makes the false claim that they are the only manuf  to offer a money back guarantee. Both Gecko and Rutex offer a money back guarantee and state so on their web sites. DMM offers a money back guarantee on their Ebay listing. Leadshine offers a money back guarantee, although not officially.


UPDATE: this claim has been removed from the site.


Larken claims that they are the only manuf to offer a 60 day warranty. That is technically true as Gecko, Leadshine, and DMM  offer 1 year and Rutex and Granite offer two years, but the implication is clear that the competitors do not offer good warranties.


UPDATE:  this claim has been removed from the site.


Inaccurate/deceptive information:


Larken shows a matrix comparing their product to Gecko and Rutex. The Rutex product they are comparing has been out of production for 4 years and the current Rutex models exceeds the Viper in nearly every category.


UPDATE: the Rutex drive has been removed from the matrix altogether, so there is no longer a comparison to the Rutex product. . The Gecko was updated to show a 340X model (actually it should be 320X)


Leadshine 810 the AMT website has a vast amount of information on general CNC as well as specifics for their products. Downloadable manuals for the drives as well as the tuning software, and some informative downloadable tech papers on CNC setup and various calculations.  AMT also sells matched servo motors and the drives can be pre-tuned to them by the factory. . Very nice for nubees!  I have not had time to search the web for reviews/opinions/support of the Leadshine drives. I will update when I have done so. I fund no inaccurate information or deceptive claims on the AMT website.


CNCdrives Whale3 the website has a main section and a ‘shopping’ section. I could not get to the main section from the shopping section without discovering and entering the main URL: www.cncdrive.com  This is only significant if you get to the site thru a link to the shopping section because most of the documents are in the main section and you won’t be able to find them. The site is straightforward with adequate product descriptions and downloadable documents and software for the products. There is a paper on general PID tuning, but nothing else in the way of general CNC literature and no formulas or tech papers. They do have a series of videos demonstrating the products.

I found no misleading information or deceptive claims. 


Both the US and Hungarian websites are so far out of date that they still show discontinued products as ‘new’. And the actual new products are not shown on the site.


UPDATE: the US and Italian web sites now show the current products.


CNCdrives Dugong: same as Whale3. Dugong is a brand new product just starting to ship and not shown on the web site.



Granitedevices VSD-XE   This site is professionally done with quite a bit of general info as well as detailed info on their products. Links to other interesting resources are spinkled around. All manuals are downloadable as well as firmware updates and marketing materials.

They have very recently added online purchasing.  E-mails to the sales group with technical questions about the product were answered within hours. Typically I only put one question in an e-mail because often if more than one question is posed in the same e-mail, the responder answers the first question and  none of the others. In the case of Granite, I posed three questions and all were answered satisfactorily.

I found no misleading information or deceptive claims.






GECKO 340:  is easy to purchase from several different sources as well as from the Gecko site. On-line purchase and payment is available as well as stock status and product is in your hands in a couple of days . .  well packaged and often with a little stuffed Gecko. Shipping and tracking information was provided via E-mail.


No additional parts are needed to install the Gecko drive. The Gecko needs only the motor power and the normal 5V control signals available from the BOB or parallel port.


RUTEX 2010  is a little more complicated in that the you can only enter an order on the site and they ‘contact you’ via email. They did contact within 24 hours with a total and the product was shipped the same day as the paypal invoice was paid. Product arrived in a couple days. Rutex does not routinely e-mail shipping information or tracking numbers but will provide the information if you ask for it.


Rutex includes one of the two required special connectors to install their drive.  And you may have to buy a 24V power supply for the Rutex, unless for some reason you already have 24V available (I didn’t).


VIPER 95 proved quite inconvenient to purchase. You can only order via phone or e-mail. An e-mail order went unanswered for 4 days until prodded with follow up e-mail and then a paypal invoice simply appeared which included a $25 shipping charge. There were more days of silence after that invoice was paid until prodded again at which time an e-mail from Larken stated the drive would be shipped the next Tuesday. On the following Thursday, another e-mail from Larken stated that the drive had not been shipped because they were waiting on parts (for the drives they supposedly had in stock) and that the parts were due in the following Tuesday. The drive was shipped the following Wednesday and arrived a few days later with no invoice or packing slip in the box. The lack of a packing slip will cause great deal of trouble if you need to return something to Canada.


Larken charges extra for the special connectors required to install their drive. The Viper makes its own logic power from the motor supply voltage, so motor power is all it needs.


Leadshine 810: is easy to purchase from the AMT website or ebay. Online purchase and payment is available at either place. I ordered the 810 from the AMT website and the product was shipped same day from a US warehouse with tracking number provided via e-mail.


As with the Gecko, the Leadshine 810 has clearly identified screw terminals.


The drive arrived promptly, but the packaging was minimal, especially for USPS shipping. The drive comes in its own close fitting box with no padding whatsoever and that box was stuffed into a USPS ‘small box’ with a tiny strip of thin bubble wrap. The drive was mistakenly shipped without the RS232 tuning cable that is supposed to accompany first purchases.


CNCdrives Whale3:  Ordering is available on the site via a ‘contact the vendor with an order form’ approach not unlike Rutex. CNCdrives site offer on-line payment of the order if you are using Paypal only. I entered the order and within 24 hours had a response that the Whale2 had been superceded by the new Whale3 (not on the website as of this writing). They provided a link to the documentation on the new drive and responded within hours to a few questions about the new model. I am in the process of purchasing this drive so I will finish this section after I have received the drive.


Although both the website checkout and the e-mail invoice they send show a cost identified as the ‘total’ price. After payment is made, they come back with an e-mail stating that you have to pay a shipping charge of US $22. This charge would cover up to 10 Whale3 drives ordered at one time.


They claim that they are unable at this time to get their website updated to show shipping charges.


CNCdrives Dugong:  same as Whale3, except shipping charge would cover 4 drives.  The Dugong is a brand new product and is in limited supply. They have promised to ship one Dugong and one Whale3 to me by the end of this week (08/14/09)


UPDATE. The drives shipped on the date promised and arrived 6 days later via USPS. Packaging was adequate.


Granite devices VSD-XE: Online ordering was added to the site less than two months ago and has an interesting quirk. You can choose between being invoiced and paying immediately using Paypal. I chose paypal and was surprised to see close to $400 additional shipping charge added.  Backing out and selecting to be invoiced resulted in an immediate e-mail of an invoice with the correct amounts.  The  e-mailed invoice contains clear instructions for paying with a bank wire transfer or Paypal. Granite’s international shipping charge US$60 is by far the highest of any of the reviewed products. Granite states that the shipping is insured. While there is no indication of it on the site, I recall from previous research that the charge covers more than one drive, so to confirm this, I placed 4 drives in the cart and the shipping remained at US$60. The shipping cost increased to $67 for 5 drives.

A  US$31 ‘install’ kit contains a mounting face plate and parts to make needed cables as well as fiber optic cables to accomplish remote monitoring. A US$30 USB cable is required for tuning the drive. Fortunately, only one is needed for as many drives as you have, and this cable is a stock FTDI part and can be purchased directly from FTDI for about $17.

The drive was shipped 4 business days after the order and an e-mail notice was provided.


Package arrived on May 14, 2010. Total time was 16 calendar  days after order.  Packaging could only be described as adequate.





Setup and Features:


GECKO 340: this  drive is very easy to install and configure. The drive is tiny and all connections are on the same edge. Wires connect with simple and clearly labeled screw terminals. The manual has complete accurate diagrams, however, if you use an encoder which exceeds the Gecko’s ability to supply encoder power, you will be on your own to get it wired correctly. In such a case, you have to provide external power and the manual is devoid of information or diagrams on how to accomplish this. Several e-mails to Gecko support netted some useful  information, but after repeated exchanges, the information required to accomplish the hook-up was still not complete.  Thru trial and error, I eventually discovered that the Gecko had to be separately grounded back to the BOB in order to read the encoder with external power.


The Gecko manual does not explain the relationship between the encoder count and the step multiplier, and therefore the formulas provided can be misleading. The manual also makes some references to additional requirements to run ‘small’ motors, but does not quantify what ‘small’ means, so that information is not useful.


The Gecko has no tuning software available for it and it does not communicate at all with a computer. The drive is tuned in real time via three pots on the back edge of the drive. The parameters are explained well in the manual and the real-time tuning is quick and easy with clearly audible feedback from the motor as you turn the pots. There is a fairly large range of adjustment within which the motor is happy, which makes the use of an oscilloscope a luxury rather than a necessity.


No additional parts, connectors or tools are needed to install the Gecko drive. However, Gecko recommends using an oscilloscope to tune the drive.



RUTEX 2010:  Rutex offers a mother board which eliminates both the mounting and wiring requirements and the drives are clearly designed with that in mind as mounting and wiring is somewhat complicated and requires the purchase of some odd connectors before you can even start your install.  They provide the part numbers and a suggested vendor for the connector, but it is aggravating that they do not include the 40 cent connector that they know you will need instead of making you pay the 40 cents plus $5 shipping or (in my case) drive 45 minutes round trip to Fry’s electronics to buy the 40 cent connector.


Unusual interface:  Rutex chose to incorporate an SPI interface for communication between the PC and the drive for the purpose of tuning. This uses the same parallel port that is used to actually operate the drive in service, so you have to figure out how to do both jobs with the same port. The manual describes the cable you must build to connect directly to a parallel port for tuning purposes, but I doubt anyone would find that very logical as you would still have to later ‘unbuild’ that cable and connect the drive you your BOB. Fortunately, you can start right off connecting the drive to the BOB and provided you use the specific pins that the Rutex tuning software requires, you can communicate from the PC to the drive for tuning purposes using the existing CNC connection. However, this will still prove problematic with more than one Rutex drive to tune as you would have to connect each drive in turn to the PC for tuning and then reconnect to the BOB. Small tweaks to the tuning would prove very inconvenient.


UPDATE: Rutex claims that when using their motherboard, each drive can be tuned in turn by selecting it in the tuning software. I do not have the motherboard and cannot confirm this. There is nothing in the documentation about it.


Rutex provides tuning software that has a complete set of checkboxes to turn the various features on or off and set all of the tuning parameters.  The software has several test modes that run the motor thru its paces and the software graphs the behavior of the drive, thus eliminating the need for an oscilloscope. The software works well, but does have bugs and some room for improvement. The SPI interface chosen by Rutex is unstable and during tuning, the link goes south frequently causing the drive to need resetting and loosing whatever parameters you were working with that you had not yet saved.


UPDATE: Rutex acknowledges the unstable SPI interface and claims to be working on and update that will fix this.


There is not a lot of info available from users or from Rutex Tech support offering specific suggestions of reasonable values to use for specific motors or even for motors of a certain size or power. Rutex tech support seems as baffled as everyone else on exactly how to tune the drive, or what the mysterious Kd index is for. It is basically all guesswork and tuning the Rutex can only be described as an arduous hit or miss process that will consume hours of your time and much of your hair.



VIPER95: This drive proved very difficult to install as the connectors come off all different sides of the drive and the drive has several different styles and sizes of connectors, only one of which is easy to use.  You much purchase and pay separately for the several special connectors that are required to install the Viper. If you do not notice this during your purchase, you may have to endure a long delay and another hefty shipping charge waiting for the connectors before you can even get started installing.  The connector pack comes with some very tiny contacts that must be crimped onto the wires and then inserted into the equally tiny connectors. Fortunately I already had a crimping tool so I did not have to buy the $40 tool recommended by Larken to make up the required cables. Still, I had a difficult time dealing with such tiny pins, many of which were only making intermittent contact after assembly and had to be disassembled again and ‘tweaked’ to make good contact. As with Rutex, this practice of not supplying required special connectors and in the case of Larken, not providing a pre-assembled connector with pigtails is just unfathomable to me.


Trying to install the Viper, you quickly discover that the manual you were provided is not for the product you have in your hand. The manual shows pins that do not exist on the product you have and there is basic information missing; for example the polarity of the power connections and the orientation of a row of jumpers, requiring several emails to Larken just to get enough info to get the product installed.


With things finally wired up, I discovered that despite Larken’s claim, the Viper does not support the latest E6 model US Digital single ended encoder.  After numerous e-mails back and forth with Larry Kenny (the owner of Larken) he eventually stated that there was ‘a way’ to make the encoder work, but would only say that the diagram would be in the ‘next manual’ and did not provide the information to get the encoder working. I continued the evaluation with a small 300 line encoder that did work with the Viper.


The Viper faulted constantly during tuning and to reset the drive required powering down the main power supply and waiting for the caps to drain. More e-mails to Larry Kenny  confirmed that this is the only way to reset the drive. When I informed him that the manual I was provided for the Viper 95 from his web site clearly showed the drive could be reset via cycling the 24V control power, and that killing the power to the whole CNC machine, including the steppers, was a deal breaker, he then said that there was ‘a way’ to reset the drive from one of the connectors, but again he failed to provide that information, claiming it would also be in the ‘next manual’. He provided dates by which he would send the ‘next manual’ but those dates came and went with no manual showing up, so I was unable to configure the planned encoder, nor reset the drive without constantly having to power down the entire power supply.



Tuning the Viper is done thru an RS232 connection using the crude DOS text based HyperTerminal program included with the PC operating system. Commands to the VIPER are simple single letter commands that return decimal or Hex values.


The continuous e-mail exchanges with Larken needed to get this product going serendipitously uncovered the fact that there are several undocumented parameters that are required to do the tuning. One of these critical parameters is the following error. The following error can only be set to a few thousand counts, but even if that was enough, the value is not saved in the EEprom of the Viper, therefore the Viper75/95 cannot be successfully tuned for applications where the low default value is not acceptable.


UPDATE: Larken claims that a firmware bug preventing the parameter save has been fixed. I no longer have a Viper drive and cannot confirm the fix. Firmware updates require the drive to be returned to Larken in Canada at customer expense.


There is no software at all, let alone graphing software and the manual is silent as to how to use an oscilloscope to peer into the drives performance, so in effect the Viper is a tune by the ‘seat-of-the-pants’ process at best.



Leadshine 810:  This drive is supposed to ship with a ‘special cable’ needed to connect it to a PC for tuning. For the signal side connectors , you will need a jeweler’s screwdriver, as the terminals are tiny. Like the other drives in this review, the screw terminals are actually on plugs so once all of the connections have been sorted out, the drive can be simply unplugged for relocation or repair.


The Leadtech tuning software functions exactly as described. In contrast to the Rutex, the Leadshine drive stays connected at all times and exhibited no annoying bugs. On the other hand, the software is ‘minimum essential’ not having the rich feature set of the Rutex offering. There is only a single trapezoidal motor test, for example. You can choose among several different parameters for the software to graph and also select the speed, acceleration, number of loops to perform and the time period to measure for the motor test routine.


The Leadshine 810 has a full set of features, but few can be toggled or set to specific values. Notably absent is any setting for servo loop period. Both Rutex and Larken warn that this is the most basic critical parameter to set before all else and in the case of the Rutex, I found that the loop period has a decided effect on the whole performance envelope.  Yet, like the Gecko, the Leadshine 820 has no setting for this. I’ve asked Leadshine Tech support about this.


UPDATE: see entry under the performance heading for the resolution to this issue.


The manual shows a set of jumpers that must be set correctly inside the drive. To get inside the drive, you remove the wiring plugs and two screws holding the lid to the base. In so doing you have to break a QC inspection label. Inside there is a row of pins, none of which has a jumper on it and none of which are labeled. Tech support eventually answered that are no jumpers inside the drive. No explanation as to why the manual instructs the user to go into the drive and set the jumpers. 


CNCdrives Whale3:  This new model only supports differential encoder wiring, but they provide a tiny translator board free with the drive to convert single ended to differential. The Whale3 uses RJ45 connectors on the interface converter board and also on the drive so that a normal twisted pair ethernet cable can be used between the motor and the drive. The drive provides plenty of 5V power for all of the encoders I have looked at. HV is thru screw terminals on a plug like the other drives reviewed here. The control signals are on a separate RJ45 connector allowing the use of a normal ethernet cable also. Some BOBs  have an RJ45 connector as well, but I don’t know yet is there is any chance of using an unmodified cable between the drive and a BOB, but they do have their own BOB with enough RJ45 plugs to run 4 axis plus another for their own relay board and another to take inputs The drive uses separate 12V power for the logic side, also coming thru the RJ45 connector. The drive can be reset via a pin on the interface without removing the HV. The drive comes with PC based graphic tuning software which communicates with the drive thru a standard USB cable and remains active in real time during operation of the drive for the purpose of monitoring the following error (and perhaps other functions)


UPDATE: The drives arrived and bolted right into the pre-made holes in a new mounting plate. The ability to rely on a vendor’s product dimensions (and mounting holes)  is important if one wants to pre-build a control box prior to having the physical drives in hand.


CNCdrive includes, free of charge, a small circuit board to convert a single ended encoder to differential. The board has screw terminals for the encoder wires and an RJ45 for the cable to the drive.  A manual clearly shops how to connect the encoder and a standard CAT5 or better network cable goes between the interface board and the drive. It woks perfectly with a few different encoders that I tried. I did not connect a differential encoder directly to the drive.


The power and motor wiring are done to screw terminals on a plug fitting that is included with the drive. Terminals are clearly marked.


The drive communicates with a PC for tuning purposes with a standard A-B type USB cable. The tuning software is a little bit complex to install because you have to set up a virtual COM  port in the PC by installing Microsoft .NET and also a utility program that creates a ‘virtual’ COM port. Clear instructions are provided for how to do all of this and the utilities automatically install just like any other driver.


The tuning software, ‘Servoconfig3’  is new and under active development, but it functions as described and exhibited no glitches during the tuning process. It does not interfere with the operation of other USB devices. The documentation, as has been typical of all of these products, is inadequate when it comes to guidance for tuning the drive. Basic explanations of the various parameters are provided and a couple of example screens and you are then set adrift. CNCdrives delivers  the drives pre tuned for their own motors. The vendor has had automatic tuning for certain previous models and claims to be planning to add that capability to these new drives.


The tuning is accomplished by setting PID parameters and doing a servo move which is then automatically graphed. This iteration process continues until the drive is tuned. The software has only one mode to move the motor and you can set only the distance moved. I found this to be incredible Spartan in contrast the rich set of diagnostic moves available in the Rutex software. Then it occurred to me that since this tuning software reportedly can show following error in real time while MACH is running, perhaps it could also tune with MACH running.


Incredibly, while MACH is up and running and controlling the drive, the tuning software can ALSO be running and tuning the drive in real time. This capability allows you to use MACH itself to perform the needed diagnostic moves. This proved extremely convenient and accurate as any change in the tuning parameters could be immediately tested in MACH.


The RJ45 ‘main’ connector is well documented. I had no problems connecting it up by simply cutting 8” or so off the ends of an old Ethernet cable and hooking the appropriate wires to my BOB. Most RJ45 connectors are clear plastic so it is easy to determine which wires go where. Ethernet cables are available anywhere and are very inexpensive for the suggested  CAT5 quality. 


The error line does not have enough power to operate an LED nor a pin on the BOB reliably. The solution is to use a transistor to power the LED and BOB pin using 5V from the BOB and letting the ERROR line simply trigger the transistor.


The drive has a reset pin . .  and that’s about all they say about it. It is up to the user to figure out how to use it. The drive resets when the reset line is grounded, so a simple momentary contact puss button is all that it needed on the control box to reset the drive. You do not need to shut down either the motor power NOR even the digital power (5V) to reset the drive. The drive can also be reset via the tuning software.




CNCdrives Dugong: similar to Whale3 and uses the same tuning software. Power connectors are much larger and include separate terminals for the braking resistor. Perhaps the most important feature of this new drive is the automatic brake resistor. HV is monitored and above a set limit, power is dumped to the braking resistor thereby protecting both the drive and the motor. This should prevent the burnouts that happen with the Rutex and Viper drives.  The drive can live with a 1 ohm resistance which eliminates the need for a voltage killing series resistor as is needed with the Rutex. No sizing calculations for the braking resistor are included in the documentation.


CNCdrives claims that brake resistor calculations will be added to a future updated manual. Absent that info, CNCdrives states that they are going to provide a properly sized resistor with my order without charge.


The Dugong has individual screw terminals for all HV power. This not only has the current capacity, but provides a very secure connection and is easy  to attach large gauge wire because you can simply use the appropriate size lug for the wire you are using and not have to worry about trying to stuff large wires into small gripping holes, and the wires cannot get unplugged or vibrate loose.


Tuning is done identically to the Whale3 using the same software.


UPDATE 06/24/2010: Since the introduction of the Dugong over a year ago,  CNCdrives had been continuously promising to add new features to the drive, but thus far, no new features have been added. The company has become unresponsive to queries about the status of any new features.


Granite devices VSD-XE: Granite now has a special parallel Break Out Board that appears to provide an easy, clean and correct connection of up to 4 drives.  For this review, I am adding a single Granite drive to an existing setup that already has two BOBs and therefore did not purchase the special Granite Devices BOB. The Granite BOB comes at a relatively high cost, but for a new install with multiple drives, it may be worth the price in time savings booth in setup and ongoing maintenance.


The drive uses the ancient DB25 connection for the motor. This connector was not designed for power and in order to use it in that capacity, Granite uses multiple pins in each power line to get the current capacity. This odd scheme is made a bit easier by the inclusion of a small PC board that the user is required to solder to the DB25 pins. This PC board groups the DB25 pins into 4 heavy duty solder points to carry the motor power.  These little boards are supplied with the drive and not the install kit. The install kit has the DB connectors and covers as well as the header type connectors and ribbon cable needed for the logic interface. The DB connectors for the motor power and also the encoder do result in a very slim profile for the drive.


The mounting points on the physical drive match the dimensions given in the documentation, so control boxes can be pre-built using the Granite manuals. A very nicely made stainless steel mounting plate is included in the install kit for mounting the motor connector side of the drive to a control box. Fiber optic cables are included in the install kit that will allow the LEDs to be monitored remotely. This is an interesting approach to remote monitoring. The fiber optic scheme uses no power, is rugged, simple and is immune to noise.


A typical screw type connector is provided for the drives power source for both motor and logic.


Note: the HV and 12V logic grounds are connected internally so users need make sure that their HVDC and the 12V logic source can have a common ground.  The documentation does not discuss this issue and a question to tech support resulted in this comment: “It is recommended to have either 12V or HV supply floating to avoid ground loops.” Which is obviously not going to be the case if the grounds are connected inside the drive. Additional questions cleared up the confusion and apparently the comment applied to the condition of the grounds BEFORE connecting the to the drive. Once connected to the Granite drive, the grounds would be common.


Everything about the Granite product is of the highest quality and also carries a corresponding high price.  For users who desire ‘the best’ and are willing to pay a premium price to get it, this drive will be the one to have. Only the Gecko,  Leadshine and AMC were comparable in quality.


Unfortunately Granite provides only ribbon cable for the  header connectors. This would be fine if using their Break Out Board, which uses a similar scheme,  but for connecting to a third party BOB, users may want to consider purchasing crimp pin or solder type header connectors so that heavier color coded and shielded wire can be used.


The tuning software is reasonably easy to install and functions well. Like many devices, the Granite drive is not a true USB device but uses an RS232 to USB converter made by FTDI. A driver is installed in the computer that created a ‘virtual’ COM port and communicates over the USB to special circuitry in the target device. Granite uses a cable with the requisite circuitry encased in one of the connectors.


However, there is a very significant caveat to be aware of; the drive cannot be tuned while the CMD cable (connection to the BOB) is plugged into the drive. It must be unplugged before the drive can communicate with tuning software or a ‘communication failed’ error is generated. I discovered quite by accident that unplugging the CMD cable cleared the problem. Assuming I had something connected incorrectly, I questioned tech support for the solution and was surprised to be informed that this behavior is known to them,  that there is no work around. Granite claims that this behavior is noted in the documentation, but I did not see this issue described anywhere in any of the manuals. If one is doing a new install and follows the initial setup sequence detailed in the ‘getting started’ manual, then coincidentally this issue is avoided, but in my view, that does not constitute being ‘covered in the manual.’


Another caveat, although an understandable one, is that the max and continuous amps cannot be set higher then the overlimit amps in the tuning software. Users are likely to run into this restriction because the max and continuous amps settings occur before (above) the overlimit amps setting. The overlimit setting must be set above the desired max befor the software will accept high numbers in max and continuous amp settings. SO one needs to skip ahead to the overlimit, set it high, and then return to the running limits.


A  potential problem is with third party break out board compatibility. I have two Break Out Boards; a CNC4PC brand on LPT one and a Homann Designs brand on LPT two. The Granite VSD-XE refused to function with the Homann Designs board. After connecting to available pins with no success, I then removed the functioning Dugong drive’s connections from the Homann board and hooked up the Granite drive to the same pins, again with no success. The Granite VSD-XE worked perfectly when connected to the CNC4PC break out board.


UPDATE 06/24/2010 UPDATE: The VSD-XE works fine with Homann designs latest MB-02-V6 breakout board.   


Overall, the Granite XE was the most time consuming and complicated installation of all of the drives. To some degree this is unavoidable because the Granite product can control so many different types of motors and has so many features that the available choices in connections, set ups, and tuning parameters are many times more than any of the single purpose drives in this review. 


The granite drive has several unique features that, for clarity, are covered in context in the performance section.


UPDATE 08/06/2010: Some claimed  features of the VSD-XE simply do not function at all. After many communications back and forth about the specified ‘position reached’ function, Granite Devices finally admitted that this function does NOT function if the drive is used with Step and Direction input. It is note worthy that in my initial contacts with Granite prior to purchasing the drive, I made it very clear that this particular feature was critical to my application and that I would be using Mach3 and step and direction input.






GECKO 340 operates in a mode that Gecko refers to as ‘singing’. Their explanation for this annoying behavior is that the motor is ‘bouncing ‘ back and forth between encoder counts as ‘required’ to maintain its position. None of the other drives reviewed do this ‘singing’. The Gecko has a fixed fault trip of about 128 counts. When that error is reached, the drive faults and the motor stops. There is no option to change this behavior.  With my chosen 1800 line encoder, the result was constant faulting for what would be harmlessly tiny errors induced by acceleration or deceleration, that would have been quickly made up in an instant, had the Gecko not faulted. Gecko is working on a new replacement model for this drive that will have a fault count that is modifiable over a small range, but the current Model 340, in my opinion, is unacceptable for anything over a 300 line encoder. 


The Model 340 has a maximum read frequency of 250k which limited my motor speed to 2080RPM with an1800 line encoder.


The Gecko 340 runs both 36V 3A and 72V 8A motors  smoothly and quietly (except when holding) with crisp instantaneous stops.


UPDATE: The replacement model 320X has following error adjustable in steps up to 2048. Frequency has been doubled to 500k. Note that this is from the spec sheet only. I have not tested this new drive. 



Rutex 2010: once tuned, this drive is nothing short of spectacular in performance. The holding is sharp and tight and the drive pushes my motor speed to it’s max 3,300 RPM which translates to 400,000 encoder counts per second. Stops are instant with no overshoot or lag even with accelerations that shake the whole machine, and the following error even at full speed with an 1800 line encoder was between 3 and zero counts. While the 36V 3A NEMA  23 motor had no problems,  the 72V 8A NEMA 34 motor was very rough and ‘growly’ until a significant amount of the mystery ‘Kd index’ was introduced. Note that while Kp and Kd run into the thousands, ‘significant’ for Kd index is a numeric value of 2. While there is a brief cryptic explanation in the manual, nobody really seems to know exactly what the purpose of ‘Kd index’ is,  and many of the users have it set to zero. In my case, Kd index smoothed out the motor’s slow speed running without compromising the crisp stops and sudden changes in direction.  


The faulting on the Rutex 2010 can be set all the way into to 30,000 or disabled completely. This is an important feature for my application driving a 4th axis that serves as both a holding indexer and also a high speed lathe spindle. The drive can also be set to fault if it detects and encoder error. This is not so important in my application, but is a critically important feature is you are pushing a mill table or gantry with the servo motor. This feature would theoretically protect against the servo ‘running away’.


It is very unfortunate that Rutex chose to use the same port for tuning as for running. This prevents any real-time reporting or polling of the drive during operation, effectively defeating my plan to read status during operation and make decisions based on that status.


As mentioned, Rutex has a less-that-stellar reputation for reliability, but their US rep claims that this was confined to a specific model and that the 2010 is robust and meets its specifications completely, although he still posted advice not to set the current limit anywhere near the drive’s maximum rating, which is inconsistent with the first claim. I set the current limit to the full 20A spec of the drive. The motor max draw is 38A. So far no problem. 


Viper95: It is hard to say what the performance potential of this drive is since tuning it requires entering  ‘shot-in-the-dark’ values into undocumented, yet critical parameters, and there is no graphical representation available to display the actual behavior in real time. However, I did get the drive tuned well enough to say that it performs very well driving a relatively large (for the review products.) motor. The Viper 95 is not useful in my application since it will not accommodate, nor store the very large following errors required for use in combination ‘lathe/index’ head, and powering down the whole machine including the steppers (which will then require re-homing) is a definite no-go.  However, the drive would be a good performer on a normal table or gantry axis. It does a very good job holding and displays crisp stops and is smoother than the Rutex at slow speeds. The drive refused to run my servo motor up to full speed without faulting, and the smallest error I could achieve in high-speed steady state even using the specific new ‘forward’ feature was in the 30 to 50 count range.  Here again that would probably not be an issue for use driving a normal machine tool axis.



Leadshine 810:  This will not be an apples to apples as I am evaluating the Leadshine using only the 36V NEMA23 motor. However, that motor is very close to the motors that Leadshine sells for use with this drive. The performance of the Leadshine DCS 810 can be described as lackluster. The holding is abysmal and there is no combination of settings I could find that would make it hold or stop sharply without overshoot. The drive should have had no problems at all with a small NEMA23 motor, but it could not run the motor in the same crisp fashion as the other three drives in this review.


Leadshine tech support has confirmed that there is no way to reset this drive other than shutting down the main power supply. The software manual shows that a fault can be cleared on a special error screen, but this does not work.


There is no setting for the servo loop period, which both Rutex and Larken say is a critical setting. However, the Gecko seems to work fine without a settable loop time, yet the loop time setting in the Rutex has an extreme effect on the behavior of the motor, so as with the ‘singing’ behaviors mentioned earlier, the manufacturers contradict each other.  I have asked Leadshine tech support about this and got a curious answer that the drive is adjusted via PID which is ‘more powerful’, which makes no sense.


UPDATE: More queries to Leadshine tech support resulted in an additional response that the tuning software that they provide does not tune ALL of the parameters (one of which presumably is the loop time).  For that you need the ‘special’ version of the software. They provided a link to the ‘special’ version, but my evaluation is completed and the drive has been returned to Leadshine so I will not be doing anything with the ‘special’ version.


During my testing, the DCS 810 drive began to show an encoder fault. Checking the encoder on the Rutex (which displays the encoder output in real time) showed it to be working fine. It was not possible to clear the error  via the tuning software, even though the software has an ‘erase error’ function. A differential encoder was tried with the same result. Only after  powering off the main power 10 or 15 times would the drive reset the fault. It would then run the ‘roadrunner’ program for a while and fault again. Note that this is an encoder fault and not a following error fault. i.e. the drive thinks there is something wrong with the encoder.


CNCdrives Whale3: The Whale3 performance is excellent. Even without spending the time required to get a really refined tune, after just a basic tune for each motor, this drive runs both a 72V 8A and a 55V 5A motor in equally excellent fashion. Both motors are silent when holding, run very smoothly at slow and high speeds and both motors are pushed to their absolute maximum RPM.  Stops are crisp and following error is only a count or two at full speed.


The Whale3 could not achieve maximum acceleration from the 8A motor because of the drive’s20A limit. The larger motor can draw nearly 40A and the 20A limit was apparent when the acceleration were set very high.


CNCdrives Dugong:  This is a big drive with big performance. I only tested this model with my largest motor, the 72V 8A Keling. With the drive’s 35A limit, the motor acceleration and deceleration is crisp. Setting extreme accelerations and speeds in MACH, very far beyond anything remotely useable in machining, generated large following errors on acceleration, but the Dugong quickly caught up and ran the motor at full speed with a error count of only 1 or 2 with an 1,800 line encoder.


The motor was silent while holding and ran smooth at all speeds.


The test motor is on my 4th axis and after an initial tune, I put the drive belt on to see how the drive would do pushing a load. The 4th axis was sitting on the floor in my office. It is heavy. I gradually increased the acceleration and ran a test program. I did not get to the max acceleration because I had to stop the test at a point where the 4th axis began to jump around on the floor from the sharp acceleration/deceleration.


The Dugong needs only a 1 ohm armature resistance so no series resistor is required and therefore the motor gets full voltage at all times. This alone made a very noticeable difference over the Rutex drive which required a 1 ohm series resistor, cutting 20 volts off at MAX torque.


Additionally, the automatic braking resistor will allow me to safely test the 4th axis to it’s ultimate potential; spinning a chuck with a heavy workpiece. The large voltage generated on decelerating this ‘flywheel’ mass will be automatically dumped to the braking resistor if the drives voltage limit is exceeded (unlikely).


UPDATE:05/14/2010, A few months ago and after the review of this drive was completed, I discovered that when using a step multiplier, the drive exhibits some bad behavior. The motor runs roughly at low speeds and cycles thru odd pitch changes while it is running. Related to this are two other anomalies.  During tuning, the step response also cycles every other step with sometimes considerably different traces. These first two behaviors have no practical effect, however, the third symptom can be a serious problem. Sequential small jogs (.001”) also are not consistent, cycling between a larger and smaller actual movements. This behavior was reported to CNCdrive and they indicated that this would be fixed promptly. Later they claimed they could not reproduce it. I then sent an audio file demonstrating the clearly audible pitch changes and also screen captures of the alternating step response traces in the tuning software. Then they acknowledged the problem but numerous queries over the next several months about progress in fixing the problem were answered with that same response each time;  that it was impossible to correct because they could not reproduce it.  Eventually, many months after the problem was reported, CNCdrives provided a beta firmware that improved the behavior slightly,  but the drives dithering was increased significantly.


UPDATE: 06/24/2010, The above mentioned problem has not been corrected and CNCdrives has become unresponsive to inquiries about it.


UPDATE: 08/06/2010, Still no corrections to this problem or other issues with the product. No response from CNCdrives.


NOTE: the CNCdrive smaller  ’Whale’ drive does not exhibit this behavior.


Granite devices VSD-XE:


The performance of this drive is on a different level from the others in this review. Only the Rutex is competition for the excellent performance of the Granite drive. The best way to describe the performance of this product would be to take the best performance, best features and best tuning software from each offering and combine them into one product.


The Granite VSD-XE has completely smooth running and silent holding performance, even with large step multipliers. None of the other drives have this combination, probably due to a pair extraordinary features that only the Granite product has;


First, the Granite has a selectable ‘input smoothing’ feature which makes the drive run with very large step multipliers as if it were running with none. This is an extremely useful feature specifically for MACH3 because it effectively eliminates the need for the smoothstepper or equivalent products in order to get the step frequency high enough to run the motors at full speed with high res. encoders. The drive is just as smooth with a 12x multiplier as with 1x even at low speeds. 


Second, the Granite has a separate fully tunable ‘holding zone’ that the drive with use whenever the ‘position reached’ condition is active. The size of the zone, in encoder counts, is user configurable. The zone has its own PID settings separate from the main PID tuning.


Unfortunately, like the Rutex, the Granite cannot be tuned while connected to the CNC control, which I find to be inconvenient, but the Granite tuning software is better in most respects and is much better documented and has no unexplained ‘mystery setting’ like the Rutex.


The Granite has none of the bad behaviors of the Dugong with step multiplication active. The motor runs smoothly and silently at any speed, does not dither while holding and most importantly, it has even consistent steps while jogging. 


The Granite is the only drive in this review that can be disabled on command and continue to track the encoder. Once re-enabled, the drive uses a separate, user configurable set of parameters to return to the correct coordinates, so the machine does not have to be rehomed in the middle of a job. In my view, this feature alone makes the expensive Granite drive worth the price. The ability to disable the drive makes it suitable for use with a 4th axis or a mill spindle drive. These applications need to have the spindle released in order to conveniently change tooling or center a workpiece in a 4 jaw chuck, or to access the adjuster on a boring head, for examples.


Another unique feature of this product is an external pin signaling ‘position reached’. This feature makes the drive perfectly suited for 4th axis duty where the spindle will be locked when it is stationary for heavy duty applications like holding a trunnion table of offset drilling.


Any one of the above features, unique to the Granite drive (of the drives reviewed here) in my opinion, make the drive worth it’s high cost . . provided of course that you have a real need for these features. In my 4th axis application, I need all of them. 


UPDATE: 08/06/2010, As noted above, it has come to light that certain features of the VSD-XE are non-functional. While this is known to Granite Devices, they have made no mention of this in their documentation or web site.


 The XE version of the drive has a very high quality specially machined ribbed aluminum heat sink which increases the drives ratings. For extreme duty, the drive can be easily and effectively fan cooled.







Warranty and Tech support:


Note that no mention was made of this review to any of the vendors. Vendors were not given the opportunity to provide products for testing and all drives were purchased on the open market. Therefore the experience and behaviors described in this review are what any new customer might expect.


Also note that drives at this price point are considered by many to be ‘hobby’ level products. However, although I personally am unaware of any of these products being used in production environments, they are definitely used by commercial shops. Therefore the standards I applied are based on my 30 years of dealing with industrial suppliers, the vast majority of which employ very knowledgeable, competent and professional people in both their marketing and engineering support groups.


GECKO 340: Gecko offers a money back guarantee for a reasonable period and a generous warranty.  While Gecko products have the reputation as being top tier in reliability, I had the occasion to use the warranty for a failed Gecko203V and the service was exceptional.


Tech support is available all over the place with Gecko specific forums and any Gecko questions are quickly answered. My tech support questions directly to Gecko, both  pre and post sale have been answered by Marcus, whom I believe is the designer of the drives. I found him to be very responsive, knowledgeable and very matter-of-fact in his answers.


RUTEX 2010: Rutex apparently is an on again off again company with extended periods of inactivity in their history. For a couple of years they have had a US rep that maintains products in stock for immediate shipment or replacement under warranty and does tech support for questions asked directly on the Rutex site. That rep is Tom Eldridge and his forum posts go back several years so there is some credibility, in my view, to Rutex’ claim to be making a concerted effort to improve their reputation and visibility in both the industrial and hobby markets in the US. I have read numerous instances where drives have been replaced no question asked including those that have been out of production for several years. I have had my e-mail questions promptly answered, although not always with a useful explanation of some of the more esoteric features of the drives that I suspect only the designer himself fully understands. Rutex needs to produce more comprehensive documentation on the PID tuning and especially the mystery Kd index parameters and include sample reasonable values as a guide rather than leave their customer base wondering if 100 or 9,000 is closer to the correct setting.


UPDATE: The Rutex 2010 drive suddenly burned out for no apparent reason. It seems to have taken the PS with it. On power up, the motor on the Rutex ran continuously in one direction. I connected the tuning software, reset the drive and the encoder and then turned the motor power back on. At that point there was a pop and smoke from behind the Rutex drive.


I have reported the failure on the Rutex Web site and asked about repair service and got no response. I then emailed directly to Tom Eldridge, the Rutex rep. No response so far.


See additional comments in the conclusion section.




VIPER 95:   A warrant of only 60 days does not inspire confidence. Considering that as of this writing, Larken still does not even provide a manual for the VIPER75/95 products that they claim to have ‘sold hundreds of’, it is surprising to me that the owner Larry Kenny can have a sarcastic and condescending manner in dealing with the inevitable flood of questions that come his way just in order for a customer to get the product working at all. He claimed that he had the solutions to the failings of the Viper95, but refused to provide those, instead simply insisting that the information would be the ‘next manual’ and leaving me hanging with a non-operational product for over a month, and making broken promises as to when the ‘next manual’ would be delivered. He ultimately stooped to name-calling, saying that my questions were “smart ass” and that I was creating a “pissing contest”. I found his behavior unprofessional, and ultimately I chose to return the drive for a refund, which he issued promptly on receipt of the returned product.


Leadshine 810: there is no mention of a money back or satisfaction guarantee on the AMT website or the product manual, although they did refund my purchase price including the shipping cost when the drive proved unsatisfactory for my purposes.  The Leadshine 810 drive carries a 12 month warranty.


Leadshine tech support was initially slow to respond to questions. When I suggested that I would need to return the drive the response got much better. Ultimately, both the US marketing rep and the Leadshine tech support people displayed a genuine interest in resolving the issues even after they knew that I planned to return the product. I made some specific suggestions on how they might better accommodate the machine tool market and they were quite receptive, even to the point of passing some of the suggestions along to management and reporting back that actions were going to be taken to address the issues. While the specific product, in its current form, was not acceptable for my purpose, the behavior and professionalism of the Leadshine and AMT personnel were in line with the higher end industrial vendors that I am accustomed to dealing with.


CNCdrives Whale3 and Dugong: The written 15 day money back guarantee shows a lot of confidence in the product, and eliminates the risk (other than one’s time and shipping cost) of trying the product. A one year warranty is offered without a lot of disclaimers. Pre-sale support was provided within hours and ALL questions were answered in the first response. The vendor is unaware of this review and I will update when I have had a chance to use their tech support.


I have had numerous discussions with the designer of the CNCdrives products and he is extremely knowledgeable about the products, about PID tuning, about DC servo motors, about CNC in general, about competing products,  and also demonstrates a rather  in-depth knowledge of Mach3. I have asked multiple questions in a single e-mail and so far, ALL questions have been answered in the first response. Thus far, his demeanor is professional and responses are complete, useful, and  usually same day.



UPDATE: During my testing of the Whale3 drive, I made a wiring mistake that left the Whale3 motor outputs wired to either side of a 1 Ohm power resistor. The drive withstood this abuse for about 20 seconds or so until two traces vaporized in the power output side of the board. I fixed the wiring problem and repaired the drive’s  PCB but restoring power to the drive resulted in immediately  blowing the fuse on the main power supply. This indicates failed MOSFETS (big power transistors), so I removed them from the drive and checked them out. Three of the 4 were shorted. I informed CNCdrives that I had killed the drive due to my own error and therefore I would not be expecting a warranty repair. When I explained that I was going to attempt the repair myself,  they were extremely helpful in providing detailed information and advice that I needed to successfully repair the drive. The Whale3 is now back in the control box humming along nicely.


UPDATE: 05/25/2010   Improvements, bug fixes and new features promised over 6 months ago have not materialized. Two months ago, CNC drives claimed to be working on fixes and new features, with a new beta firware imminent. However after a couple of weeks passed with no contact, I queried the progress and was told that the firmware was not being worked on during that time because CNCdrives had sold out of all of the Dugong drives and had none to test with. New drives were not expected for an additional two weeks, making a total 4 week unannounced delay in the ‘imminent’ firmware release. Eventually a beta firmware was provided which improved the problem, but also caused the drives dithering to increase dramatically.


UPDATE: 06/24/2010: At this writing it has been over 9 weeks since I tested this firmware and reported the results to CNCdrive. During that time several inquiries about progress have gone unanswered. 


UPDATE: 08/07/2010: Still no resolution and no response from CNCdrive.


Granite devices VSD-XE: The basic drive carries a one year warranty while their flagship XE comes with two years. Granite web site claims that they have a zero tolerance policy for hardware failure and software bugs. There is also a claim to have ‘Premium Support’ which lasts the lifetime of the product. Pre-sale questions were answered completely within hours. However, questions asked after the purchase have gone unanswered. I asked via e-mail if the drive’s ‘position reached’ indicator remained active when the drive is disabled.  No e-mail response was provided, so I started a discussion on the Granite section of the CNCzone forum. 8 days after the e-mail question a response was made on the forum indicating how the ‘servo ready’ indicator functioned. So I asked again on the forum the same question about the ‘position reached’ indicator function but there has been no reply.


UPDATE: 05/14/2010 Apparently not all of the Granite e-mail addresses are monitored. Support emails send to the support e-mail address go unanswered, but if sent to the sales department get prompt responses.


All questions were eventually answered and response has been consistent and timely since I began addressing the support questions to the sales e-mail address.


UPDATE: 08/07/2010: As mentioned above, it is clear that Granite devices is aware of certain features that simply do not function and also they are aware that the drive cannot be tuned while connected to the CNC controls. No mention of these limitations is mentioned in the documentation and it takes numerous queries to Granite tech support to finally secure an admission that these limitations exist. In the case of the tuning limitation, Granite Devices claims that the limitation is stated in the manual, but it is not.




The Gecko drive works very well with small motors provided you have a low count encoder. It is by far the easiest to connect and tune. Gecko is about to come out with a ‘significantly improved’ replacement, but the couple of specs I got from Marcus were not significant improvements in my opinion. However, no conclusion should be drawn about the new 320X drive until the full specs are published.


UPDATE: the 320X has been released and specs are available on the Gecko website.


The Viper was returned after much frustration with both the product and the support. The Viper 95 performs very will if you stumble across the right settings, but it does not store the following error setting, limiting it’s applications with a high resolution encoder, and the Viper’s high price and high shipping cost combine to put it significantly higher than any of the other drives. If Larken fixes the firmware bugs and produces the promised manual with the ‘secret’ solutions included, that might relieve some of the frustration of dealing with the product, but as of this writing, this choice should be avoided.


UPDATE: Larken has added the Viper75/95 information to the existing manual to create essentially a ‘universal’ document. Diagrams for external modifications to fix the encoder interface and to rig a hot reset are included. This information was not available at the time this review was written, so I have not tested these fixes. I will not be retesting any of the Viper drives. The main power polarity is shown, but the orientation of the rows of jumper pins is still not identified.


The Leadshine is the least expensive, but it also has the worst performance, being unable to stop the motor crisply without overshoot. The only way to reset the drive after a fault is to power off your main power supply, which is problematic for use on a machine tool. Leadshine is introducing a new lower cost 810’S’ version of the drive which uses only single ended encoders and has a reset pin, however, a close read of the manual reveals that the reset is ONLY applicable to a following error and all other errors still require a complete power down of the main supply. The Leadshine 810 , in its current version, also should be avoided for use on machine tools.


Of this bunch, the Rutex is the clear winner (see update below), with the possible caution that reliability might be an issue, based on Rutex reputation. This product is not without problems and requires a lot of patience to get it working, but the performance is exceptional and it is the only drive in this review that does not have constant faulting. The Rutex has good price, availability, and the performance is exceptional. The only issue is the inadequate manual, which does not provide sufficient information or examples to get the drive tuned quickly and easily. The tech support, as I mentioned earlier was not able at this time to provide useful insight into the cryptic settings. The interface chosen is unstable and does not stay connected reliably during tuning. There is no such issue with running the drive however, and thus far the drive has performed perfectly. If you can find someone using the same motor that you plan to use and you can get the drive settings from them, and you pre-purchase the special connector that Rutex should have provided, then the only obstacles to the Rutex install would be eliminated. 


UPDATE: the RUTEX is now the number two drive in this review having been outdone by the new WHALE3 and Dugong models.


UPDATE: as mentioned in the warranty section, the drive has burned out. Rutex has thus far been unresponsive. As mentioned in the review, Rutex has a reputation for their drives burning out. Rutex acknowledged this, but claimed it was limited to a specific mode and that the 2010 was fine. However, they also warned not to set the drive at it’s full rated power, which I found quite suspicious.


Based on this experience, the well known Rutex ‘burn out’ problem is not confined to specific older models. This drive blew out the MOSFETs and one of the MOSFET controller chips. The controller chip is an L6386 and is listed as ‘obsolete’. Mouser electronics has no stock and Newark electronics has a note that when their stock is gone, they will not be getting any more. There is no replacement or equivalent listed at either of these major suppliers.


The conclusion is that the burn out problem is not confined to older models as claimed by Rutex, Rutex is unresponsive to customers with burned out drives, and even the new models use obsolete parts so  it will be only a short time until the drive are unrepairable due to unavailable parts.  Based on these findings,  I have to recommend that Rutex products be avoided.



The Whale3 and Dugong drives from CNCdrive are now the top choices in this lineup.


UPDATE: 06/24/2010: the Whale3 remains the ”BEST CHOICE” among the smaller (100V/20A or less) drives in this lineup, but after adding and reviewing the recently upgraded Granite VSD-XE, it is now the TOP CHOICE . See UPDATES under other headings for the latest info on the previous TOP CHOICE, the Dugong drive.


Performance  is equal to the Rutex, but there are no significant detractions like mounting problems, oddball connectors, buggy interface and software and the like.


Documentation is accurate but needs improvement in the area of completeness. This is understandable given that these are brand new products and the documentation is available in several languages. The web site, on the other hand is abysmal and far out of date. This has no effect on the quality and performance of the products, but certainly cannot be good for business and in some ways it could begin to be a detriment to existing customers due the lack of downloadable documentation and software updates for the new products.



Support is exemplary. By contrast,  Larken is completely unacceptable and Rutex, while having good intensions, is more adept at making excuses for the existing shortcomings in the product features and documentation, as well as their business practices,  than they are at actually fixing or improving anything. CNCdrives tech support has thus far proved to have a  professional demeanor matching that of Leadshine, and  fast response with USEFUL answers to ALL questions asked, rivaling the service of Gecko.  As a company, CNCdrive are keenly interested in customer feedback and express a genuine interest in improving the products and documentation whenever possible, but do not provide false dates or unrealistic expectations that they then cannot meet.


UPDATE:05/25/2010: While support from CNCdrives was excellent initially, and they demonstrated a keen interest in improving the product. Good intensions are admirable, but the actual results are that the product is in essentially the same condition as it was at its introduction 11  months ago when the original review was performed. After the first unsuccessful beta firmware release 4  weeks ago, CNCdrive became unresponsive and has provided no updates or ETA on promised improvments.


UPDATE: 06/24/2010: Another month has passed (now 9  weeks after the first unsuccessful beta firmware test) and still there is no new firmware and no responses from CNCdrive. The conclusion that I reach is that it has been discovered that the promised fixes, improvements and new features are not doable.



Cost comparison shakes out like this:  Rutex charges $17 for the encoder interface which CNCdrive provides free. Larken does no offer the interface, but it can be purchased form encoder vendors for a similar price. Adding the Rutex 2010  ($148) or Viper95 (recently reduced to $148) drive and interface together, either drive is nearly $60 more than the Whale3. The price difference between the Dugong and the comparable Rutex2020 of Viper200 is also significant.


Overall, while performance is the number one priority, the Whale3 and Dugong drive’s intelligent layout, standard connectors, USB interface and solid tuning software make them far less frustrating to deal with in terms of installation, setup and maintenance.




Granite devices VSD-XE:  . . the NEW TOP CHOICE.  After being added to the lineup, evaluated and compared to the competing products, the Granite VSD-XE  earns the top spot.


Granite and Gecko are vendors in this lineup that have continued development and made significant improvements to their products. (note that all of the vendors are aware of this review and all have been invited to provide updates and/or product announcements) Since this review was first undertaken, Granite has completed and released previously announced features like 160V capability, their unique dual DC brush motor capability, and a motion library. Announced, but not available at this time is a sophisticated series of stepper motor control schemes.


Tuning changes cannot be conveniently tested under CNC control as can be done with some other drives, because like the Rutex, the Granite drive cannot be tuned while connected to the CNC. This makes it very inconvenient to tweak tuning since the control box would need to be opened and the drive unplugged from the BOB before it could be tuned, and the USB unplugged and the BOB reconnected to test the new setting under CNC control.


UPDATE:06/24/2010  after spending some time with the product, I have found the above mentioned (and undocumented) restriction to be an extreme inconvenience. The inability to tune this drive without first opening the control box, physically disconnecting the ‘command’ cable, then connecting the USB cable, (and then going thru a special power cycling sequence each time  to connect to the drive for tuning) has proven to be a very time consuming and restrictive.  For example, there is a status screen in the tuning software that shows the state of various flags and indicators which is functionally useless since they would only have meaning if monitored in real time while connected to the CNC with the drive operating a motor. The lack of documentation on the problem together with tuning software features that would only be useful In real time, lead me to conclude that real time tuning with the CNC controls connected was the intent, but that for whatever reason, was discovered to be unworkable. Granite gave no indication that this would ever be corrected.


While the Granite VSD-XE is my TOP CHOICE for this review, that choice is weighted heavily with performance and features of the VSD-XE that the competing products simply cannot match in their current configurations. However, users requiring a 100+V/35A drive who do not need the features of the Granite product, and are not gong to use step multipliers, would most likely be happy with the less expensive previous winner, the Dugong from CNCdrives in Hungary. 


In both performance and features, of the products that were evaluated in this lineup, the Granite XE has no equal. I have made the argument that for use with DC brush drives, the significantly higher cost of the Granite XE is not justifiable since you would be paying for capabilities (AC, Brushless, Stepper, etc. that you would not be using. However, the cost of the XE in many cases can now be literally halved for medium sized DC brush motors because the VSD can now handle two motors.


Previously, from a pure performance standpoint, the best  showing came from the Viper and the Rutex, but problems or limitations with each drive mitigated this advantage. When price, tuning software, performance, support, and reliability were all considered together, the Dugong was the leader,  even though it may not have been the best in every category. However, as with any review, it should be remembered that the product’s performance is delivered under a specific time frame and conditions, which is fluid and can change over time.


Between the original review 10 months ago and today, the Dugong has stagnated and retains its deficiencies, Granite has continued development and added very significantly to their product offering, cutting the effective cost in half for certain applications and providing increased capacity and capability. 


On performance alone, the Granite XE is equal to the best drives in this lineup. When step multipliers are used, the Granite significantly exceeds the previous winner; the Dugong.


In features, it has no equal in this lineup. In this group of products, of those that were evaluated, the Granite VSD stands alone in having some of the features of the mainstream industrial drives costing much more, so the XE is also the winner on features.


On price, Granite has leveled the playing field with their dual drive capability. Only the Gecko remains less costly for two motors, but the Gecko has much smaller ratings, so a direct comparison is not reasonable. When used to run two DC brush motors of medium size, the Granite VSD-XE is competitively priced. Shipping remains extremely expensive for those in the US, but I have checked this out and Granite has no control over the extremely high cost. Those are the rates. Same package is $14.00 from US to Finland, but $60 from Finland to the US. If 4 drives are purchased at the same time, the shipping remains the same and becomes less of a disadvantage.


Reliability is unknown at this point and will be covered with an update if the drive proves unreliable. However, the drive has about every electrical protection one can think of and the mistake that I made which killed the Whale drive would not have affected the Granite XE.  The VSD has overcurrent protection that would eliminate the likely cause of the Rutex failures.


The manual’s advises to limit the power during initial setup. I did this by placing a 3A ‘safety’ fuse in the HV supply to the drive. Once the drive was functioning and the settings ratcheted up to 10A cont 38A max, predictably the next step response test blew the 3A fuse. My impatience led to putting in a new 30A fuse before the CAPs had bled down fully, discharging the cap into the VSD drive. Certainly this blunder has the potential to immediately fry many drives, but the XE came thru unscathed.  


Note that the VDS-XE was tested only with a single large DC brush motor for this review.  AC motor, BLDC motor, Stepper motor and the dual DC brush motor capability were not tested and no conclusion should be drawn from this review about performance in those applications..


UPDATE: 06/24/2010: Despite Granite Devices website claim that they have a ‘zero tolerance’ policy for errors, the actual product has non-operation features and extreme limitations that make tuning extremely difficult if not impossible because the motors behave differently in use than they do with a single tiny jog and there is no practical method  to conveniently test tuning setting made with the tuning software, making the tuning process a very tedious and arduous process involving repeatedly unplugging and plugging the cables.


Additionally, Granite Devices has been claiming for months that their stepper drive firmware would be out in ‘two weeks’ and it is still not released. They state that fixing the tuning issue is ‘not a priority’ and that making the advertised fatures actually work would be ‘a good idea for the next firmware release’.


While the Granite VDS-XE has undisputable the best performance of any of the DC brush drives in this review, I would not recommend it unless there is a specific feature that is not only critical to the intended application, but that the functionality of that feature can be confirmed in advance and made a condition of the sale of the product.




Comments from the Vendors:


Tom Eldridge, from Rutex tech support, provides this rationalization of why they do not supply the special connector:

I just want to note that we do not expect customers to make the connector that you mentioned you thought should have been supplied.”I tell every customer who inquires about a drive that I don't think already has a mother board to get one, and I also offer to take it back after they are done using it for tuning.  If I did not mention this to you, it is because I assumed you already had a mother board.”  Mr. Eldridge sums it up with this statement;  “ . . . . people make so many mistakes trying to wire things, and then I have to support them and help them find their own wiring mistakes.  So...using the Rutex mother board is the only smart way to go.  That's why I don't supply the 16 pin connector.”   



Paul, from Leadshine Tech support has this to say about the products inability to reset.

“For the error clearing and reset problem, I think most of the errors would rarely occur if all the components in a system are best designed. Otherwise, the system should be redesigned if errors always appear. So the ability to clear the error without turning off the power is not so necessary for a well designed system.”



Larry Kenny, owner of Larken provided these statements in response to this review. This is typical of his tone.

“Thanks Mr Simpson for the lovely review, after answering your 16 or more sarcastic and rude emails, i finally said just return the drive which I then gave you your Full money back.   I'm not sure what your issues are ,but no one deserves to be someones door mat just because he sells them a servo drive.”


“Yes, i've sold hundreds of Viper 100/200, but the V75 is only been available since 4 months. I have since fixed the T paramenter save bug which you pointed out (thanks). As far as the delay in getting you the manual, I am a single parent and loaded down with sales, r&d and taking care of my kids, so yes, relentless, stubborn emails can be frustrating (most people don't argue like you did).”



Balasz from CNCdrives provided this response to the inability of the drive’s error line to properly signal a standard 5V  BOB:


“However, our new BOB already includes the buffer and the error line on this BOB is connected to an opto onboard. The new BOB will be priced also for hobbiist, so I think to use it together with our drives will be the best and easiest solution.”


Tero Kontkanen from Granite Devices has this to say about the mission critical ‘ Target Reached’ function not working on the drive: 


“ I believe you would need value that states when servo is settled within given tolerance. Is this correct? There is no such output (at least for now). Thanks for pointing out this too. It sounds to be a good addition to the next firmware release.”


And in response to my suggested possible solution of using a mechanical switch (as I had to do with the Rutex drive) instead of constantly plugging and unplugging ribbon cables:


“Yes, you can place switches to CMD cable lines but you'll need switches also for USB cable to prevent it from interfering the lines in non-config mode. Our future drives will have separated lines for both links so simultaneous use will be much easier possible.”



Send questions, corrections or comments to  steve@thecubestudio.com