Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: simpson36 on August 04, 2009, 07:58:52 AM
-
'Living document' will be revised and expanded occationally but link will remain the same, so just hit it again for lastest updates.
This is only a documentation of my experiences with the products and is not an endorsment or 'buying giude'. Comments welcome.
http://www.thecubestudio.com/ServoDriveReview.htm (http://www.thecubestudio.com/ServoDriveReview.htm)
-
Thanks for the post. Good comparative info I must say. I know it took you time and $ and sharing is
appreciated.
RICH
-
Thanks for the write up. I have used Geckos past with smaller servos motors and just purchased a Viper 200 to test on some larger 100V DC Servos.
I'll post my testing on the Viper 200 when I can.
-
That's a good read. Thanks for sharing your knowledge and experience.
-
Thanks for the write up. I have used Geckos past with smaller servos motors and just purchased a Viper 200 to test on some larger 100V DC Servos.
I'll post my testing on the Viper 200 when I can.
The choices that I am aware of in that price range are:
Viper 200 170V 20A $240 +$25 ship
Granite VSD-E 160v 14A $280 + ?? (ships from Finland or New Zealand)
Rutex 2020 200v 20A $230 + $12 ship
CNCdrives 180v 40A $170 + $22 (ships from Hungary)
It would be great to have reviews available for these products.
-
I'm satisfied with the Rutex 2010 from a performance standpoint for my 4th axis project, but I am still in the hunt for an inexpensive servo drive that meets all of my wish list items to use on the X,Y,Z axis. I'm using my little X2 as a test bed to find the servos and drives I want to use on my upcoming X3 project.
So I have added a CNCdrives model called Whale3 to the lineup and ordered the drive. I've started adding some info on the drive to the review.
FYI, Gecko on Friday announced the new 320X. They provided me some early specs which were not very impressive, but the final released version is much better. Most of the problem areas have been fixed, but unfortunately, the drive will still bounce between encoder counts while not moving. Gecko refers to this as 'singing'. I have a different word for it. The encoder section of the manual is still incomplete.
CNC drives also has a brand new 160V 35A drive coming out called 'Dugong'. No word yet on whether this will replace their current 180v 'Mammut' model (where do they get these names? ::))
-
First off thanks for taking the time and expense for this test.
My experiences with Gecko 340 and Rutex drives is exactly the same. The 340's with 500 count encoders would fault unless I used the slowest of accelerations and even then It was just unworkable. I suffered with this for over a year and finally replaced them with 4 990 drives. It was a night and day difference. The machine just ran like a champ. I did find this link about tuning that helped me get my arms around it a bit better. http://www.eaglemotion.com/home/R9xHbdi/pg5.htm (http://www.eaglemotion.com/home/R9xHbdi/pg5.htm) Although I'm still not sure if it's as good as it can be. The new drives are great in that you can set the PID values with the software.
My only real complaint with Rutex is availability. Tom at Rutex US has great customer support and has helped me many times.
Derek
-
Good link, thanks!
A couple things about the Rutex I don't like are having to run thru a full time 1 ohm resistor which carves 20V off my top end, and the fact that there is no protection from BEFM.
Both Rutex and Viper have the reputation of burning out from BEFM. The more expensive Granite drive does have a protective resistor feature. Rutex warns against setting the drive's amp limit to its full spec. ??? I have it set to its full rating anyway, but I'm not particularly confident in it not only because of the warnings, but because I have read so many stories about burnt out drives. I have not tested the 4th axis at full speed with a heavy chuck and workpiece yet because I am worried that with a serious 'flywheel' being decelerated from 3,000+ motor RPM, I may get smoke. Larken obviously expects this to happen since they have the Viper's power components socketed for 'easy replacement' and they advertise their drives as 'repairable'.
Mo' better, methinks, to not have them burn up in the first place, which brings me back to the Dugong model from CNCdrives.
I have been talking to CNCdrives quite a bit and found out that the Dugong model does replace the old Mammut model. It has a ton of nice features, including automatic tuning, but what I am really interested in is the automatic brake resistor feature (above a set limit, it bleeds off excess voltage to the resistor ). That would protect the drive and the motor against getting burnt up. The Dugong looks looks like it may be the ideal drive for the 4th axis, so I went ahead and bought one along with the Whale3. With the Dugong watching over itself and the motor, I should be able to 'put the boots to it' and get back to breaking stuff ;D
The CNCdrives ship to me in the next couple days . . from Hungary . . so no telling how long till they get here.
-
Added review of two more drives, both from CNCdrive. Same link.
The Whale3 and the Dugong . . . NEW WINNERS!!
Anyone contemplating buying servo drives will want to read this.
-
The Dugong looks like it might be what people have been waiting for, a decent drive capable of running the DC motors present on the 80's CNC Knee mills :)
I looked at the specs of the older drive (Mammut) and it seems that the continuous current is 40Amps but there is no mention of peak, so I presume that is both continuous and peak? Does the drive just fault if the 40Amp is reached? Is this the same for the Dugong albeit 35Amps?
Hood
-
I will check that out and post back. Last night I finished connecting my new control box to the mill and got the faults and resets working. To monitor faults, I have lights on the box for faults (and it stops MACH) and simple reset buttons also on the control box.
Reset on Whale3/Dugong is just grounding a 5v reset line so it could be done from MACH also. Rutex requires cycling the 24v digital supply so it would need to have a relay rigged up. I'm just using an on/off toggle for that.
As to over-current, by way of example, the Gecko does not fault at max amps, it just limits it there. That is what I *think* the Dugong does also, except that if the overage lasts a certain amount of time, it will then fault, but let me make sure and post back. As far as I know at this point, you cannot turn off faulting on the Dugong.
After to performance and the cool software, the big advantage I see in the Dugong is the automatic handling of the braking resistor (the one they sent me is friggin' HUGE!). I am aware that the user can and *should* build an external braking circuit and I have seen a couple of schematics for how to accomplish that, but I suspect few people actually do because I never read "my braking circuit failed', but I read over and over "my drive burned out. The built in pre-engineered braking circuit is perfect for 'electronics for Dummies' level guys like myself. *Theoretically* it should be impossible to burn up a Dugong drive. I am going to beat up on it as much as possible, but I do not have a big enough motor to really get it's attention.
-
Sounds like they are nice drives and are coming closer to the functionality that I have with mine. Mine would cost 10 times the amount these do though so they do seem a great deal.
I would imagine if they limit the current then the drive would fault, not on overcurrent perhaps but following error. This would be easily sorted with tuning I would think, as long as the rapids/accel were still acceptable, due to detuning, which they likely would be :)
I like the protection of the shunt resistor, still not quite as functional as my drives but again they are a fraction of the cost so cant possibly grumble about that and they certainly sound like they are way ahead of the competition.
Hood.
-
The Dugong limits current to 35amps. It will allow a small momentary overcurrent (no spec on that).
It does not fault (stop the drive) for overcurrent situations, although it does have an phsical LED on the drive and the tuning software has a virtual LED to indicate current limiting is active.
The drive will fault if the overcurrent exceeds 40amps and is not controllable by the drive, for example with a short circuit.
I also learned that the fault lines and reset lines of the drives can be combined so that a fault on one drive will stop all of them and all can be reset with one line.
Yesterday, I was messing around with the Dugong and I had Mach running a test program while I simultaneously watched the servo drive diagnostics on a laptop. changing a PID parameter and then immediately rerunning the same test program IN MACH is definitely the hot setup. 8)
Earlier, in my office, I had Mach running a test program and the Dugong tuning software running simultaneously on one of my render slaves (Core2 duo). Can't say if all of that power was needed, but both apps ran fine together and MACH never hickupped at 60k kernal speed, so it is definately doable.
So far, so good.
-
Sounds an excellent drive :) Thanks for the info.
Hood
-
Here is a quick review of the Viper 200.
My Colchester / Ailam Crusader control went U/S so required a rapid conversion to Mach + new drive system.
I contacted Larken and explained I needed the drives quickly. I ordered 3 - 2 for the machine + 1 spare.
These were despatched swiftly with all connectors / jumpers and a programming lead included.
The PC boards apear to be wlll constructed and finished, I could not see any mounting or soldering defects. Nice...
It took a few weeks to design and build the new control system, Viper manual is up to date, (manual should recomended / detail using fans / additional heat sinks if end mounting).
The drives were commissioned as per the manual, powering up the logic side and checking coms - no problem there.
Check of the PSU - OK 130VDC
Power onto the drives, YES, we have movement.
Tuning is easy, if not a bit of a guestimation exercise - the manual does give some help but its a bit thin (fortunatley my business partner has designed small digital servo drives so understood the PID settings and generally how to tune by feel.)
The tuning process took about 1 hr per axis, checking the stiffness of the drive and ensuring stability. All done manually, i.e twist the pulley by hand and check the movement of the axis on a dti.
Overshoot is the most difficult thing to detect and tune out.
The drives have been given a good work out for a week now, they maintain accuracy and run very cool, and have not needed to be reset.
Well pleased.
The Anilam control had max rapids of 4m/min...quick test with the Vipers we had this up to 6m/min (and could have got more) but having just got the lathe working again and not wishing to break it, we set the max rapids at 4m/min
The only thing I would like to see is mounting recomendations in the manual and a graphical means for seeing the motor position vs demand signal to aid tuning.
The CNC Drives Dugong looks interesting - might try these to replace my stepper driven Bridgeport drives...mmmm.. decisions decisions.
Either way, the Vipers do a good job. ***Recommended***
ATB
Derek
-
Thanks for following thru with that, Derek. A downspec version of the Viper200 is what I *thought* I was getting because that is the manual Larken was providing as the Viper95 manual when I made my purchase. Larry Kenny informed me that he has updated the manual and the Web site and I added that info to the review. It is important that people keep in mind that the Viper200 and the Viper95 are not just different versions of the drive, as is the case with other models. They are totally different products. The Viper100/200 had the features I was looking for. The Viper 75/95 did not.
Just as an interesting note: because of my own mistake, I killed the Whale3. But with detailed instruction from the vendor, I was able to repair the drive! ;D This is what happened:
I was testing all of the servo drives with a variety of motors and one was a 35V NEMA23 which I ran with 72V thru a 1ohm wound power resistor in series with the motor. At first power up I had the leads reversed and when I changed the wiring in a relatively inaccessible area, I ended up with the motors leads connected to each other and the Whale3 outputs connected to either side of the resistor. Not good. The drive tolerated this abuse for (as a guess) about 20 seconds before vaporising two traces on the board. I corrected the wiring and fixed the traces, but on power up, the fuse on the main PS popped indicating a dead short in the Whale. I have a little bit (enough to be dangerous) of prior experience with MOSFETS and knew that they fail shorted, SO that was my "electronics for Dummies' level diagnosis. I removed the MOSFETS and sure enough three of the four were shorted. :'(
So informed CNCdrives what I had done and told them I did not expect a warranty repair, but would appreciate any advice they might be willing to share. To my surprise, I got a very detailed instruction on how to trouble shoot the problem, track it thur the drive, and what parts to replace (including part numbers) and in what order! I'm no electron whiz, but I CAN follow instructions so I bought the parts ($6 total). I've done plenty of desolder/solder stuff, but never any SMD, so I watched some instructional vids, bought a temp controlled iron and some other SMD tools and practiced on an old VGA card and some memory sticks. Parts came, fixed the drive and its running fine!
(http://www.thecubestudio.com/RapairedBoardBottomWEB.jpg)
(http://www.thecubestudio.com//RapairedBoardTopWEB.jpg)
-
The drives have been given a good work out for a week now, they maintain accuracy and run very cool, and have not needed to be reset.
The drives should not need to be reset unless something goes wrong. And sooner or later, something goes wrong. What is important is that the drives are able to initiate an e-stop in Mach. I strongly suggest you set up and test that function before you have need for it.
According to the manual, the Viper200 should be able to be reset by cycling the control power without killing the HV, so that part shouldn't be an issue.
-
Good job on the repair, see, like me, you also like to do a bit of platic welding at the same time ;D
Hood
-
The drives have been given a good work out for a week now, they maintain accuracy and run very cool, and have not needed to be reset.
The drives should not need to be reset unless something goes wrong. And sooner or later, something goes wrong. What is important is that the drives are able to initiate an e-stop in Mach. I strongly suggest you set up and test that function before you have need for it.
According to the manual, the Viper200 should be able to be reset by cycling the control power without killing the HV, so that part shouldn't be an issue.
Hi Simp'36,
Yup, nice job on the repair, hope those tiny wenie little critters with 471 marking don't need changing, just look at the size of them compared to the 8 pin DIL package... we'll need a soldering iron the size of a pin !
I have done some SMT work myself and hand soldering that stuff is almost as hard as the strain on your eyes.
Re: the Viper 200s, my estop circuit is hard wired (mushroom head crash stops and "very"end of limit travel switches) at the moment, nocks out the spindle VFD as well.
Next step is to wire in the 1st satage travel limits and the Viper fault signal to the BOB Estop I/P . This will disable the step / direction OPs to prevent any further movement.
The "very" end limit will kill everything in the case that a drive has run away.
that is the philosophy anyway :-)
Those Dugongs do look good though ;-)
ATB
Derek
-
Yup, nice job on the repair, hope those tiny wenie little critters with 471 marking don't need changing, just look at the size of them compared to the 8 pin DIL package... we'll need a soldering iron the size of a pin !
Yep, them there were the SMDs that needed to be replaced. I got away without changing the 'legged' SMD IC chips nearby as they were OK. I got instruction on how to check them out with a scope, but I only have a software scope and I have used it, but don't really know what I'm doing with it. So the SMD ICs 'check out' came with power on and see if the motors move . . :-X
As to the e-stop topic, anyting is possible, but an outright runaway crash into the table stops is pretty unlikely with reasonably current electronics, methinks, as the drives (not sure about the Viper, but *probably*), have encoder sensors that will shut down the motor if the encoder dies or becomes scrambled or otherwise unreadable. This is separate from a normal servo 'following error'. The more likely scenario is a broken tool or stalled spindle nowhere near the table stops. If the machine keeps on truckin', things can go from bad to worse in a hurry . . . like the spindle crashing into a fixture or something equally fun . . none of which will be caught by a table stop, but would most likely cause a following error which is what you want to stop the movement and possibly (but not necessarily) the spindle.
Seems like you've got a plan scoped out, though. A caveat with the Dugong and Whale3 is that the error line doesn't have enough power to drive the BOB input reliably, so a little circuit has to be constructed for the interface. Post back if you have any similar issue with the Viper and I can post the schematic for the little driver circuit I devised. I would post it now except that CNCdrives has suggested a better alternative to what I came up with and I would want to check that out first.
-
Interesting develpoment . . . the Rutex self destructed.
Unlike the Whale . which I murdered with my own mistake, the Rutex just exploded for no particular reason. In the process, the PS rectifier was also fried.
The AC line-in fuse on my PS popped off and I was able to track it down to one of the bridge restifiers being shorted. Once I located the problem it was relatively simple to replace. However, on power up, the Rutex oddly was moving the motor continuously in one direction.
I reset the drive and zeroed the encoder and turned the HV back on and . . POP! Fire and smoke from under the Rutex . . . . and . . . the rectifier was fried again. Hard to say who threw the first punch.
The Rutex exploded a MOSFET and has thus far resisted all attempt at repair.
Each drive has an independent fuse on the positive DC power (20A in the case of the Rutex). Despite the exploded MOSFET and the shorted Rectifier, that fuse did not open. I replaced the rectifier again and the mill is back running fine with the two CNCdrive servo drives and a stepper back on the Z axis for now.
I don't know if this is enough info for anyone to formulate a theory about what happened. What I don't understand about the event is how the Rutex somehow *appears* to have fried the rectifier. Is that possible?
-
Update:
I reported the drive burn-out and asked for information about repair on both the Rutex web site and after getting no response, there, also by email directly to Ton Eldridge, the Rutex rep. Thus far no response their either.
An interesting discovery that I made while researching the parts needed for repair, is that the MOSFET controller is an obsolete part. One of these controllers is burned out on the drive along with the MOSFETS. Mouser has no stock and Newark says that after their stock is gone, there will be no more available. Neither of these major suppliers lists a replacement or equivalent part.
The controller is an ST brand L6386. It would seem there are two conclusions to draw:
1) the well known Rutex 'burn out ' issue is NOT confined to older models or even a single current model
2) it seems reasonable to assume that even the current model drives will soon be unrepairable for lack of available parts once the obsolete stock is used up.
-
Another update on the Rutex drive failure:
I received an e-mail responce from Tom Eldridge, not as as a result of my official contacts, but as as result of posting here in the Mach forum.
He states in part: "When a customer asks for warranty service, I normally send out the replacement the day I receive the bad drive from them. If I am comfortable with the customer, I will send out a replacement if they call me right away, even without payment and even before I receive the blown drive. If you had called me right away, I'd have done that for you too. "
Note: This statement is consistent with reports I read and obtained from Rutex users, as noted in the review. I did not call Rutex. I used only e-mail communication.
Mr. Eldridge goes on to add this condition to the warranty offer: "Normally I would just warranty this for you, but if you put the current setting higher than I recommended and if the drive solder runs are damaged he may only want to warranty them half way...in other words, you would have to pay for half of a drive to receive a new one from us."
Note: The drive is rated at 20A. The recommended setting above which the warranty is cut in half is 7A. This reccommendation, to my knowledge is 'published' only in a single old posting in an obscure Yahoo forum. The drive continues to be rated at 20A in the product literature.
Thus I have apparently voided my full warranty not by attempting to repair the drive, but by using it at it's rated capacity. Fortunately, I have successfully repaired the drive myself, but as others have done, I will have to 'derate' it and find an use that requires only 7A max.
-
Talk about the fine print. 20A vs 7A is more like 1/3 of it's rating.
-
I agree, if that is truly the case then they are deceiving people and everyone should be told that.
Hood
-
Vladimir, top dog at Rutex, sent me an explanation as to why it is 'OK' to have to limit the current, claimed that this doesn't effect operation because the drive will still somehow provide the max amps . . . . which makes no sense to me and is not consistent with my experience with the product. For me the drive had distinctly different behavior with the amps limited to 7 as 'recommended' by Rutex. I could not get a 3.5A cont, 20A peak motor to accelerate using that setting.
However, setting the drive to it's rated 20A resulted in the best performance of all of the drives by a small margin . . but only for three weeks until poof.
The boss man offered full warranty replacement instead of the half warranty offered by the US rep. Since a refund if the product 'does not meet customer expectations' is part of the warranty, I asked for that and returned the drive altogether and got a full and prompt refund.
I have decided not to review the Granite drive. I want to make it clear that this is not due to some known failing of that product. The focus of the review is DC brush drives and for that purpose, the Granite is simply too expensive to justify. The eagerly anticipated 'dual DC drive' is now reality, but so it the resulting limitation to 20A per motor, leaving a situation where one can still purchase two competing equivalent products cheaper than one Granite drive. When you include the 'install kit' and the extraordinarily high shipping cost, the drive is simply impossible for me to justify for DC servo use given the alternatives. You mileage may vary.
If a user contemplates using brushless motors, either DC or AC, then the cost of the Granite is more palatable. Granite has only one "universal, I can do everything" product, so for low end application such as stepper or DC brush servo, you still pay for the enormous capability, even though you will not benefit .
So, after several months, thousands of reads and dozens of emails, the review is therefor completed and closed. Vendors can still provide new information or corrections and I will continue to do updates.
-
Thanks Simpson.
Brett
-
On the basis of this (and a couple of other) thread, I bought 3 x Dugong drives plus a HD BOB from CNCDrives to retrofit a Bridgeport Series 1 Mk2 Interact.
I've kept the original SEM servo motors which are rated at about 140v, 25A. The rating put them outside the capabilities of Gecko's which were my first choice.
The Dugong drives have proven to be fantastic. Unbeatable for the money. The config software is a bit clonky - but works well enough. I give them a whole hearted recommendation!
I added AMT102 capacitive encoders set to 1000 counts per rev which the drives multiply to 4000 giving 0.000625mm per count.
I've been using the mill with Mach and the new drives for a month now. I've been a Mach3 Turn user for a couple of years and never used Mill before. I felt right at home with it after about 5 mins - brilliant!
To give the Dugongs a real test, I have decided to use one for my spindle drive controller. The mill has a 5Hp DC motor which is pretty big physically. I'll add another AMT102 to the spindle then run it with step & direction from Mach.
I'm using a Smoothstepper so I should be able to run it pretty fast.
If they can run a spindle - they can run anything! I'll keep you posted!
Si
-
Forgot to mention, the reason I want to try one on the spindle is because the original Mentor2 drive let the smoke out when I started the spindle with the spindle lock engaged.
I would have thought it would just throw an over current error - but the smoke got out first! I have been quoted $1800 to fix it - so it's going in the trash!
On the plus side, this will leave the whole of the mill running on single phase!
Si
-
Good to hear :)
Hood
-
I've kept the original SEM servo motors which are rated at about 140v, 25A. The rating put them outside the capabilities of Gecko's which were my first choice.
I've been trying to buy these motors on ebay and I keep getting sniped each time. What do you think of the motors in general?
To give the Dugongs a real test, I have decided to use one for my spindle drive controller. The mill has a 5Hp DC motor which is pretty big physically.
For that big brute motor, you should include a BIG braking resistor on the Dugong. The drive uses the resistor automatically . . a nice feature . . and it should be fairly indestructible so long as it has the big resistor to dump to.
Incidentally, I have added the Granite Devices VSD drive to that review and I will be adding the MR-C and J2S series Mitsubishi industrial drives as well once I have some more experience with them (and get some time). The review link does not change, so hitting it again always gets the latest version.