[Splint]

Hi,
I am having an issue with my 2010 drive setup. Whats happening is then I try and run the motors with the rutex vb6 tuning software the motors run but shudder. I've played around with various tuning settings but regardless of the settings the same shuddering behavior is always there. I have all but eliminated the possiblity of radio frequency intereference with the encoder signal as I am now using audio video cable which was reccomended to me by an electrical engineer who is experienced in industrial robotics, and I'm running it far away from the motor power wires. I ran the same process with another pc with identicle results which I think I can safely assume that the signal from the pc is fine. I've powered the motor power wires direct from the power supply (bypassing the rutex drives) and they run strong and smooth. I've yet to varify the data cable from the pc the the rutex motherboard is ok, I think it is but I'll try and get hold of another to try it out.
Appart from that the only think left is the rutex drives and motherboard. Each drive has the same behavior. When I run the step response sometimes the motor runs away sometimes it dosen't. When I hit the stop button it continues to run, I have to turn off the power supply to make the motor stop.

Any ideas?

Thanks
Splint

[Torchhead]

It does sound like a tuning issue to me. The drives ship basically un-tuned. If the gain and dampening are at the default values they will barely drive a load. If you can grab a motor axle and move it with the power applied (motor corrects after it has rotated about 1/4 turn) then it's not accepting the tuning parameters you are entering. I have never been able to get the software tuning to work but I don't have a motherboard so communication was in question. Untuned motors are "spongey" meaning you they respond slowly or delayed with a physical twist of the motor.

You also have to have the encoders in phase with the motor signals are it will run off on it's own. You might try just reversing the motor power leads (leave the encoders hooked up as they are) and see if it changes the behaviour.

TOM CAUDLE
www.CandCNC.com

[Splint]

O.K. I beleive I have found the problem, but the issue is not fully solved....

I did some experimenting to try and iscolate the the problem and in doing so I found that cutting shortening the cable that runs from the R2210 line driver to the R2010 driver board made the motor run normally. I was able to tune the motor to what I regard as being quiet good (no shuddering, a nice curve in the tuning software screen and running quiet smoothly). The shortened cable was about three inches long and is unsheilded but works fine for testing purposes.

I was under the impression that the R2210 line driver boosted the signal such that up to a few hundred feet of cable could be run between the line driver and the servo controller card. I cannot find any documentation to support this but from memory that was the suggested maximum length. I suspected that there was a voltage drop in over the length of the origional cable (8 meters) I was using (power from the servo driver to the line driver) causing the line driver not to work. I modified the 3 inch cable so the power and earth remained at three inches and the signal wires were 8 meters - The result was the origional problem I was having from the start, upon closer examination it appeard that the signals (a and b) were being partially lost along the way. The motor would run away and shudder but the encoder readout on the tuning software was not proportional to the motor rotation.

I did also use some sheilded cat 6 cable which was sourced from a factory which uses it without problems on motion control equipment in a production environment. That cable had 4 pairs sheilded with an overall sheild. Same thing, motor ran away, encoder readout not proportional. That cable was ADC Truenet CAT 6 S/STP LSZH4P AWG26 20 meters in length.

I'm pretty confident that it is a signal loss issue and not a noise issue.

Does anyone have experience with running the R2210 line drivers or know what the true length limits are? Alternately can anyone suggest an alternate signal booster that will allow me to run up to 10 meters of cable to the servo drive?

Thanks
Splint

[Torchhead]

I'm pretty confident that it is a signal loss issue and not a noise issue.

So lets do a little math. Let's say the wire size you are using is 24 ga. The average current across that will be 25ma or less. The resistance of 24 ga wire is about 1 ohm at 10 meters If you start out with 5 volts at the motor (differential encoder end) then your loss on voltage across 10 meters (39 ft) of wire is E=I X R or E = .025 * 1 or .025 volts. 25 milivolts drop. That is pure DC loses. The impedance of UTP is around 100 ohms. To get the HF loss of the signals you need to know the frequency of the pulses. Some rounding of the signal will be evident but the driver end should be able to handle the cable impedance and send a clean enough signal to be recognized by the receiver. Not knowing what Rutex has at the receiving end and what impedence matching they are doing it would be difficult to profile the HF (leading edge) roll-off characteristics.


In a differential drive system the DRIVER is at the motor (differential encoder) and needs a source of +5 referenced to the RECEIVER power supply. That normally comes through the cable. The receiver is at the other end. If you try to pump single ended through the cable you will get little or no signal so any length of wire will kill it.

We use differential signaling with Gecko 320/340 series on our EZPlug cards. We supply a differential Driver (pigtail) that plugs into a USDigital E5S or E2S single ended encoder and makes it a balanced differential signal. Chips on the EZplug card sort it back out and into the single ended Gecko. We can easily run 15 to 20 meters in a noisy environment with no problems. In theory you should be able to run 25 meters or more with no problems.

Shielding on a cable does not help the voltage loss. It just keeps noise from being injected from the outside world and only then if the shield is not connected on both ends. It does add to the capacitance of the cable so the HF characteristics are even worse than unshielded UTP!

[Splint]

In theory you should be able to run 25 meters or more with no problems.

Are you suggesting that the line driver isn't working correctly? I've got three of them and they all behave the same way.

[Torchhead]

A proper working differential balanced line can drive 5 volt signals up to about 30 meters. A 12V differential drive can do 100 meters on CAT5. A 48V differential drive can do thousands of feet (example ISDN digital phone lines). I'm just telling you what a properly designed driver can do.....Since they all do the same thing it's valid to conclude:
1. The encoders are the wrong style OR
2. They aren't hooked up right OR
3. The 2010 can't supply the correct or sufficient DC volts to run the Drivers OR
4. A miracle happened and all of them went bad at once
5. All the above

TOM CAUDLE
www.CandCNC.com

[Jim Scott]

I had this problem or at least on very similar when setting up my drives . While tuning the motor would at random start a violent shudder and get worse until I had to cut the power. I spent hours trying to "tune " this out before I realised I had a genuine fault. I did all my setting up and testing with a drive card fitted to slot 1 of the motherboard as suggested in the manual . I lost about 2 weeks of eveings changing cables ,motors encoders power supplies and monitoring with a scope before I tried the damn thing in a different slot on the mother board . Guess what ? it ran smooth as a breeze in the remaining 3 slots .The motherboard was replaced under warranty .
Jim.

[Splint]

Thanks for the suggestion Jim,

the problem in this case is consistant regardless of whatever slot, card or line driver combination I use. The only time i can get the motors to run as they should is when I run the short (3") lead from the line driver back to the servo drive. This suggests to me that the A B signals are breaking down over the 8 meters which they had been running. I know for a fact the the circuit is correct, the only difference between the working setup and the not working setup is the length of the signal wire. I can only assume that the line drivers are not working or there is a design/manufacturing fault in the 2010 drives causing them to not be able to read a slightly weak signal. From my understanding the line drivers should amplify the signal such that a few hundred feet of cable can be used, perhaps the line drivers are not working at all. They are a Rutex product so there should be little possibility of incompatibility. I've emailed Tom Eldridge from Rutex US asking him to read through this thread and see if he can help but I've yet to hear back from him...Tom, if you're out there...

Cheers
Splint

[rutexus]

Wow,

You are certainly having problems getting set up. Thank you all for your help in this problem. I am impressed by the advice given.

In supporting Rutex drives for the past 4-5 years or so, I have found that most problems like this occur from inadequate shielding on the encoder feedback. It is not just shielding however that is important. It seems as though even with shielding, it is important that the encoder wires be kept away from the wires running to spindles, the motor power wires themselves to the servo motor and any other noisy signals.

The Rutex drives respect the shielded pair configurations, and utilize popular line drivers, the 26c31 and 26c32 matched pairs of drivers. Yes, they run on 5 volts, and I know users often use them 5 meters. I don't hear from too many users running them 10 meters, but that does not make me think they are not used at those distances.

I had one user tell me that in order to eliminate his encoder feedback noise issues, he had to shield the servo motor wires to within 3/4 inch of the R2020 drive connections, but that was the most stringent case I can remember.

Be sure NOT to connect the shield on the motor wiring to the motor case. I would recommend you grounding the motor case. While I have not seen this cause problems with Rutex drives, I have seen ungrounded motors cause problems with Servo Dynamics drives, Elmo drives, and others.

Although this sounds entirely crazy, I shield the motor wires, but do not connect the motor cable shields to anything on either end.

Time, and time again, the problem turns out to be noise picked up from the motor or encoder wires. I'd try different motors to ensure that the motors themselves are not introducing noise if you have swapped out everything else. Divide and conquer is always my tactic in these cases. Normally the problem is that there is some part of the configuration that it is inconvenient to swap out, and the customer holds off from trying that. I cannot remember a rutex drive having this kind of problem inherently. I'd press on and isolate the problem by moving things or swapping things out.

Obviously you have tried the drives, and R2210 adapters already. If you need another mother board to try, I have them in stock, and I will warantee any of our products.

If this does not help, give me a call at 573 341 1528, or 573 368 7399. I always prefer phone support to email.

Vladimir just returned yesterday from Eastern Europe where his ailing parents still live. For the last two years he has really been struggling to help them through the autumn of their lives. This time he was over there caring for them for the last two months. It is hard on him. Before this last trip, he finally introduced the R2040 drives. I have them in stock. They work well in my tests on a lathe with Servo Dynamics analog drives on it. Although I ran out of stock of the R2010 drives just recently, I expect to have more in stock in a couple of weeks. They are in production again.

I hope this helps.

Tom

[Splint]

Hi Tom,

thanks for the reply. I'm really at odds with this now. I came back to it after a while to run through and be clear on exactly what happens under what conditions and I started off with the short (3") cable between the line driver and the servo drive card which had shown a lot of promise of working normally, but when I ran it the whole issue came back in full effect, I have absolutely no clue as to why all of a sudden it was working ok and then not.

I purchased a 24volt 1amp power supply to power the motherboard since the previous board power was on a winding on the same coil as the 60 volt motor supply. I thought an independant motherboard power supply might help but it didn't.

I tried wrapping aluminium foil around the entire length of the encoder cable, wrapping the power and earth (twisted pairs) as a pair then a and b (twisted pairs) as a pair then all together, the sheilding was not grounded but it made no difference.

I got two 12 volt batteries, hooked them up in series and used them as a motor power supply (completely eliminating the mig welder internals I had been using which I suspected may have been a cause of noise), unfortunately this made no difference.

I earthed the case of the motor to the motor power supply, no difference.

I powered both the motor and the electronics from the 24 volt batteries, no difference there.

I've tried the motor (Elecrtrocraft E543) with the Renco encoders as origionally supplied and with US Digital E2 encoders, no difference.

Ive tried a range of lengths of sheilded (8 meters and 20 meters) and unsheilded (3",10,15 and 20 meters) encoder cables, both near and away from other electrical cables, no joy there although the 3" unsheilded cable seemed to work ok temporarily.

I've tried a couple of different cables from the port on the PC to the motherboard, no difference.

Iv'e tried a couple of different computers to run the tuning software, no difference.

It's not an environmental issue as the problem has been consistant from when I built the machine at my parents house to when I moved it into my own garage.

I've made sure there was nothing stupidly wrong like connectors not properly pluged in etc.

It appears to me that by looking at the encoder position readout in the tuning software and moving the motor shaft that the line drivers either cannot deliver a clear signal or the drives cannot interpret a clear signal. I dont know if this is caused by electical noise or something else.

Tom, I've not been able to contact Vlad since before he went to look after his parents, do you know if it would be possible to bundle up one of my motors, the looms and the Rutex products send them to Vlad and see if he check over the whole thing come up with any answers?

Cheers
Splint

[rutexus]

Hi Splint,

I think you get the prize for spending the most time working on your problem. I wonder whether you may have damaged the boards with your power supply. If the secondary shared any wires in common, this is very possible. If the secondary coils were completely separate from one another that should not have been a problem. If the returns were not connected properly, they surely would have hurt the drives.

Anyway, you may pack it up and ship it to me or Vlad, and as long as you pay the shipping, I'll check it out for you for no charge. I have some more R2010 drives coming in this week from Vlad, and then soon I will have another large shipment of them coming in.

Tom

[Splint]

Thanks Tom,

damage to the boards is possible but it seems odd that for a short period the drive I was testing with actually came good and worked pretty well then went bad again, I really can't work that one out. When it's good it's good, when it's bad it's possesed by the devil. When I first started investigating the problem it was in a bench testing situation and there was just an occasional kick in the motor, when it's on the machine and driving the axis it really goes nuts, but I did have it on the machine when it was good and got it running quiet nicely for a while.

With respect to the power supply setup, there is a single secondary winding which runs to the capacitors and then branches off into dedicated fused cables to each board each with their own earth running back to the capacitors. At no point has there been any daisy chain type arrangements.

I'm happy to cover freight charges, I'll try and get in touch with Vlad and see if he can check them out, that way I don't need to worry about international shipping and Vlad is probably in a better position to be doing any repairs if required.

Best regards
Splint

[rutexus]

Splint,

Now I'm even more suspicious of your power supply. Please fax me a schematic of the power supply. Fax: 573 341 2672. Put to Tom's attention.

Tom Eldredge