Another DMM drive losing counts to Kanalog

[mmurray70]

Been having a few issues with gradually loosing position on my DMM servo/drive on my 4th axis. Its being controlled analog with encoder feedback to kanalog. Everything works great with spindle off, did a bunch of testing/tunning and everything was spot on. But after running a few parts I noticed things are drifting off a little with spindle on.

Im using really nice double shielded encoder cable with shield connected to servo drive ground. And its only a short run of about 3ft. Kind of surprising this is causing problems as im using the same cable in much longer runs on my other 3 encoders (not DMM drives) and its working perfectly. I see another thread with similar problems on a DMM drive a month ago. I wonder if the encoder output from the DMM drives is causing problems?

Any suggestions on what else to try? Should I try connecting the shield to Kanalog ground instead of servo drive ground or is this a bad idea? Also the 220v single phase supply for this servo drive was taken from 2 of the 3 terminals of the spindle VFD, could noise be comming from these power lines and messing up the drives encoder output signals? I could take power directly from transformer output but its more difficult and only a couple feet away anyway.

Similar Threads:

• Need Help!- Losing quadrature encoder counts from DMM servo drive to Kanalog
• Need Help!- Partner 4 w/ centurion 4 losing servo drive
• need help, spindle drive, losing active light, ovl1, ovl2, error..
• KFLOP + Kanalog to JF H2N-TD 3Phase 220V Servo Drive Wiring Questions
• Servo drive enable connections - KFlop/KAnalog connection?

[Jim Dawson]

I would only connect the shield at the servo end, not the servo ground, but rather the encoder output shield (the D-sub shell). This works well with my DMM drives, I've had no problems. Daisy chaining the power off of the VFD might not be the best idea.

[mactec54]

Been having a few issues with gradually loosing position on my DMM servo/drive on my 4th axis. Its being controlled analog with encoder feedback to kanalog. Everything works great with spindle off, did a bunch of testing/tunning and everything was spot on. But after running a few parts I noticed things are drifting off a little with spindle on.

Im using really nice double shielded encoder cable with shield connected to servo drive ground. And its only a short run of about 3ft. Kind of surprising this is causing problems as im using the same cable in much longer runs on my other 3 encoders (not DMM drives) and its working perfectly. I see another thread with similar problems on a DMM drive a month ago. I wonder if the encoder output from the DMM drives is causing problems?

Any suggestions on what else to try? Should I try connecting the shield to Kanalog ground instead of servo drive ground or is this a bad idea? Also the 220v single phase supply for this servo drive was taken from 2 of the 3 terminals of the spindle VFD, could noise be comming from these power lines and messing up the drives encoder output signals? I could take power directly from transformer output but its more difficult and only a couple feet away anyway.

I doubt that yours is the same problem, his was a chip problem, from his control side

You have a wiring Problem, you should have a Power Filter before your VFD Drive and a separate main Power Supply for the Dmm Servo Drives, nothing should be connected to the VFD Drive, ( This is crazy stuff ) so you need a power distribution Terminal Block before the Power Filter, from this you can split up as many 240v main Power connections as you need, twist all these 240v power wires as this helps with noise as well

How do you have the spindle wired, Shield and Ground connections are important

[TomKerekes]

Hi mmurray70,

Should I try connecting the shield to Kanalog ground instead of servo drive ground or is this a bad idea?

I think that is a good idea.

Please also see this wiki article:

Regards

[mmurray70]

I doubt that yours is the same problem, his was a chip problem, from his control side

You have a wiring Problem, you should have a Power Filter before your VFD Drive and a separate main Power Supply for the Dmm Servo Drives, nothing should be connected to the VFD Drive, ( This is crazy stuff ) so you need a power distribution Terminal Block before the Power Filter, from this you can split up as many 240v main Power connections as you need, twist all these 240v power wires as this helps with noise as well

How do you have the spindle wired, Shield and Ground connections are important

Spindle wiring is all stock, machine is a 1994 Fadal 4020. None of the spindle power lines (in and out) are shielded. I didnt really like the idea of taking off the VFD either, but this is the only source of 240v in my machine. There is only one 240v three phase output from the transformer and it goes directly to the VFD within a couple feet of cable. There is no filter either, the VFD manual mentions line filters, but there were none installed on this machine. Do transformers act as line filters? Could it be that there is no line filter because the machine has its own transformer?

Is there any advantage of connecting directly to the transformer or wouldnt this be the exact same thing since its all directly connected just a couple feet away? The transformer uses weird push in, spring loaded connectors so this is why chose to connect to the VFD instead, especially considering there was no filter or anything in between.

What are my options for filtering the supply to the DMM drive? I installed one of these on the control power side of the DMM drive: https://www.digikey.com/products/en?keywords=817-1207 Is this suitable? I dont think it was a perfect match to what the manual suggested. Is there something I can use to filter the main Dmm drive supply line as well? Would an isolation transformer help or would the noise pass right on through?

Tom, Ill try changing shield ground to Kanalog and see how that works.

[mmurray70]

Just tried moving shield ground to Kanalog and no improvement. Should I try the parallel resistors on inputs next or do something with my power supply?

[TomKerekes]

Hi mmurray70,

Just tried moving shield ground to Kanalog and no improvement

Did you verify the shield was isolated on the Drive's end?

Should I try the parallel resistors on inputs next or do something with my power supply?

Since the resistors are easy you might try that next.

I don't suppose you have an oscilloscope?

Regards

[mmurray70]

Hi mmurray70,

Did you verify the shield was isolated on the Drive's end?

Since the resistors are easy you might try that next.

I don't suppose you have an oscilloscope?

Regards

Yes I did dissconnect sheild from the drive end. Ill try the resistors next. No I dont have access to a scope but I always wanted an excuse to buy one. I might take a look around at used ones. Any suggested models to look at? Are the computer based ones any good?

[TomKerekes]

Hi mmurray70,

No I dont have access to a scope but I always wanted an excuse to buy one. I might take a look around at used ones. Any suggested models to look at?

I find Rigol brand to be good quality and reasonably priced. $260 and up depending on speed, number of channels, and memory depth. The DS1104Z has 4 channels, 1Gsamples/sec, 100MHz bandwidth, 24 million points of storage.

https://www.rigolna.com/products/dig...1000Z/ds1104z/

Are the computer based ones any good?

I'm not familiar with any.

Regards

[mmurray70]

I tried adding the resistors and the problem is considerably worse. Even looses position now with spindle off, and drifts off much faster with spindle on. So I guess this would suggest that the pulses are on the low side and resistor dropped it down further and made things worse right? Pulses not high enough to be read?

Do you think this would be suitable for a scope? https://www.amazon.ca/Hantek-DSO5072...C6QP2RFVDQYG86 This has good reviews, good price, and 2 day delivery.

[TomKerekes]

Hi murray70,

I tried adding the resistors and the problem is considerably worse. Even looses position now with spindle off, and drifts off much faster with spindle on. So I guess this would suggest that the pulses are on the low side and resistor dropped it down further and made things worse right? Pulses not high enough to be read?

Strange. I suppose you might be right. You might check the levels with a Voltmeter. You connected 2 resistors, one for A and one for B correct? From + to -. Not to GND correct?

Do you think this would be suitable for a scope? https://www.amazon.ca/Hantek-DSO5072...C6QP2RFVDQYG86 This has good reviews, good price, and 2 day delivery.

Yes that looks good. I would guess the noise might be something like 100ns pulses. Without a scope it is like working blind.

Regards

[mmurray70]

Yes I connected two 160 ohm resistors. One across A+ and A- and one across B+ and B-, nothing to ground. Not sure why it made the problem worse. I checked the voltage tonight and it was 3.70 or 0.60 to ground depending on state.

I went ahead and bought that scope, should have it by Friday. Might need some help with setup. I bet the signal is a mess.

So what are my options for filtering that 240v supply line to the DMM drive anyway if I do find a noisy signal? Would this be a good choice? https://www.digikey.ca/product-detai...631-ND/2136886

[TomKerekes]

Hi mmurray70,

Yes I connected two 160 ohm resistors. One across A+ and A- and one across B+ and B-, nothing to ground. Not sure why it made the problem worse.

Neither do I.

I checked the voltage tonight and it was 3.70 or 0.60 to ground depending on state.

That should be correct.

I went ahead and bought that scope, should have it by Friday. Might need some help with setup.

Cool. Can't wait to see what you see. I'd scope the encoder signals at ~1us per division and 5V per division to look for spikes. One thing to be careful about is that some scopes connect the probe ground to line input earth ground. I usually don't want this and "float" the earth ground of the scope. Otherwise be aware you will be earth grounding whatever you connect the probe ground to.

Would this be a good choice? https://www.digikey.ca/product-detai...631-ND/2136886

I don't have much experience with those but seems like that might be oversized current wise to what you need. I think the higher the current rating the less noise attenuation you might get.

Regards

[mmurray70]

Cool. Can't wait to see what you see. I'd scope the encoder signals at ~1us per division and 5V per division to look for spikes. One thing to be careful about is that some scopes connect the probe ground to line input earth ground. I usually don't want this and "float" the earth ground of the scope. Otherwise be aware you will be earth grounding whatever you connect the probe ground to.

Can I check to be sure its isolated with a simple multimeter between probe ground and earth ground and look for high resistance?

I don't have much experience with those but seems like that might be oversized current wise to what you need. I think the higher the current rating the less noise attenuation you might get.
Regards

Ok ill do more reading and try and figure out exactly what amperage i need. The DMM manual actually suggests a 50amp breaker or contactor somewhere which made me think big, its gotta be overkill for a 750w drive though.

[durundal]

Ok ill do more reading and try and figure out exactly what amperage i need. The DMM manual actually suggests a 50amp breaker or contactor somewhere which made me think big, its gotta be overkill for a 750w drive though.

FWIW I have this main power chain: 240V 50A breaker in panel --> power cord --> 30A breaker in cabinet (from when I only had a 30A wall circuit in a previous location) --> contactor --> individually fused 3x DMM drives and on power up the 30A breaker will trip about once every 5-10 power ons. They have never tripped the 30A breaker during actual use, I think they must draw a big inrush current on startup. There is also a Leadshine (to be replaced with a 4th DMM) and VFD but it still can trip the breaker even when those have their fuses pulled. I don't recall if 2 DMMs is enough to trip it but I don't think it ever did with just a single drive.

[TomKerekes]

Hi mmurray70,

Can I check to be sure its isolated with a simple multimeter between probe ground and earth ground and look for high resistance?

Yes. You will probably measure either 1 ohm or less or 1 Meg Ohm or more.

Ok ill do more reading and try and figure out exactly what amperage i need. The DMM manual actually suggests a 50amp breaker or contactor somewhere which made me think big, its gotta be overkill for a 750w drive though.

It seems hard to find specs on inrush currents for line filters. But it seems TDK allows up to 50X the rated current for a few milliseconds:
https://product.tdk.com/info/en/cont...778880096.html

[mmurray70]

Hey guys, I got a hold of DMM on the phone today and may have found the problem. Looks like It could be my fault, I skipped a step when installing. They said I need a connection from Kanalog ground to pin 5 of the DMM encoder output connector. And I left this out. I just connected the A,B and Z signals and that was it. He was sort of surprised it worked at all without it. Does this sound right to you Tom? Nothing I can damage by connecting this is it?

Also, I dont have any spair conductors left in my encoder cable. Should I run a single ground wire (DMM said this was fine), use the sheild as a ground wire, or buy another cable with more conductors? Would be nice and neat if i could use the shield but im not sure if this is a good idea or not. Ill wait for you guys to advise on whats best.

Also I told him about the possibility of noisy supply line and he said if this doesnt fix the problem it would be a good idea to put a filter on the supply line. He recommended the same type as used for the logic power supply (Schaffner FN610-1 or equivalent), but larger. He said a 10a capacity should be fine for a single 750w drive.

[TomKerekes]

Hi mmurray70,

Hey guys, I got a hold of DMM on the phone today and may have found the problem. Looks like It could be my fault, I skipped a step when installing. They said I need a connection from Kanalog ground to pin 5 of the DMM encoder output connector. And I left this out. I just connected the A,B and Z signals and that was it. He was sort of surprised it worked at all without it. Does this sound right to you Tom? Nothing I can damage by connecting this is it?

No GND could certainly be a problem. Even though the signals are differential, there is a limited amount of common mode offset that is allowed. Connecting a GND shouldn't cause damage

Also, I dont have any spair conductors left in my encoder cable. Should I run a single ground wire (DMM said this was fine), use the sheild as a ground wire, or buy another cable with more conductors? Would be nice and neat if i could use the shield but im not sure if this is a good idea or not. Ill wait for you guys to advise on whats best.

Using the shield is probably not a good idea. Separate wire should be ok.

Regards

[mactec54]

Hey guys, I got a hold of DMM on the phone today and may have found the problem. Looks like It could be my fault, I skipped a step when installing. They said I need a connection from Kanalog ground to pin 5 of the DMM encoder output connector. And I left this out. I just connected the A,B and Z signals and that was it. He was sort of surprised it worked at all without it. Does this sound right to you Tom? Nothing I can damage by connecting this is it?

Also, I dont have any spair conductors left in my encoder cable. Should I run a single ground wire (DMM said this was fine), use the sheild as a ground wire, or buy another cable with more conductors? Would be nice and neat if i could use the shield but im not sure if this is a good idea or not. Ill wait for you guys to advise on whats best.

Also I told him about the possibility of noisy supply line and he said if this doesnt fix the problem it would be a good idea to put a filter on the supply line. He recommended the same type as used for the logic power supply (Schaffner FN610-1 or equivalent), but larger. He said a 10a capacity should be fine for a single 750w drive.

The Ground may work on the outside but it, could also be a noise problem should be a cable with the correct number of conductors, IGUS will have the cable you need

A 10A capacity would not be fine, the Drive has up to a 20A rating so you would need at least 25A Filter

[mmurray70]

Hey guys good news looks like its working with a ground wire connected! Sorry for taking up your time and missing this obvious mistake.

I didnt like the idea of another single wire dangling out of the db9 connector so I went to order another cable, and when I looked at that info on the cable im using they consider the braid outside as a 7th conductor, an uninsulated ground wire. And its already connected to Kanalog ground now, so I decided to try it as a ground and seems to be working fine. So ill see how it goes. If i have any more trouble ill post back.

Big thanks to all that helped!