Set up using AMC AB15A100 drives, brushed DC motor Prototrak Plus with Encoders

Todd & tommylight many thanks for your help. I have disconnected (left in the temps, so hook up Rigol, anytime) the Rigol o-scope, and see no real noticeable change in the trace. I plan to work (never moving Pot4), Pot1 & 3, try to keep pot 3 below 7.0 turns (as you have stated this is a high number of turns) & work Pot1 so the motor does not get nosies (usually over 10.5), with the intent the trace has sharp corners and level "cruse".
I am open for suggestions, and very grateful for the replies!

Many thanks!

Interesting, now it is nicely behaved. I think that would seem to imply that it is a noise issue on the encoder lines, and some how hooking up the scope helped clean up the signal enough to not bother the servo drive, or there is a wiring connection problem and what ever you did to connect the probe leads, improved that connection so it isn't currently causing a problem.

Not being an electrical engineer I really don't have a clue what the answer is, but I'm sure to someone who knows what they're doing the answer and solution is probably obvious.

I'm not sure what is going on with your scope trace but I don't think the settings are right.

If it stays as nice as the last plot, it is probably good enough to move on to tuning the position loop.

Todd, many thanks! Attached is a trace, taken just now, the corner is not so sharp, but balanced (symmetrical) top to bottom (when trace is below the "zero" table is moving to the right), the settings are SW2=1.147; Pot1=10.0; Pot3=6.75, this are same settings as you referenced in prior post. I have tried Pot 1 over 11.0, with Pot 3 low, as to not get spike, the spike will 1st appear when the table moving to the right. I am willing to do what is necessary to get this correct, but if some sacrifice at top feed rate is necessary, its ok. I still have the "temps" connected & 7i77 card is out of cabinet. I will wait for your response before putting the electrical cabinet back together.

Many thanks!

You may just be able to call it good enough. I'd be inclined to try tuning the position loop in Linuxcnc with it as is. If you can't get satisfactory results with the position loop, then you could come back to this and increase pot1 to get a faster response, then increase SW2 to tame the overshoot caused by the higher pot1 setting. The AMC instructions are to turn up pot1 until it buzzes (I think when it is at rest, enabled but no commanded movement.) Then back off 1 turn. Then fix any overshot with SW2 (with the test we were using).

Todd, tommylight, piasdom, the sponsors, moderators, and the many members that organize the forum and offer suggestions, I am ever so very grateful for your help, especially, the members who have welcomed me and others like me who are not gifted with the electrical and programing side of motion control, I offer my sincere thanks.
I am embarrassed to report that the "x" axes has apparently what ever the issue is has changed to a permanent issue, not intermittent. I tried to start the (the P, I, D, FF1 etc...) tuning, I did not change anything in the electrical cabinet, so all SW1 settings are OFF, the SW2 & SW3 are the same for both drives. The .ini file (the .ini & .hal files are attached) MIN+FERROR=2.0; the MAX_ACCELERATION=1.2 (the "Y" axes is 10.0) there is some small difference.
When I started I had some traces that were in the correct direction, the angle at the start/end was less than 45*, with a small "cruse" area, small wavy trace. I never took a pic because I was thinking I could do better. Now, it appears as soon as I use F2 the "X" axes motor will start & run fast until it errors out, the belt is off, it was very shocking the 1st time it happen when belt was on.
I have really studied all the connections associated with "x" axes encode wires, used volt meter to check continuity all the way to the plug at P1 on A-M-C drive & at TB3 on 7i77 card.
If I need to use Rigol o-scope I will skin some wire off the plug at the A-M-C, P1 connector and if needed pull terminal out of plug and slide a section of shrink tube over the bare wire. This will allow the 7i77 card to remain mounted in the el cabinet.
The LEDs all look proper on drives & 7i77 card. The pic of the "Configuration" screen for the "y" axes has 3 extra lines: OUTPUT_SCALE; OUTPUT_MIN_LIMIT & OUTPUT_MAX_LIMIT. I was going to try tuning the "y" axes but noticed the difference.

Many thanks

Edit: The DRO seems to work proper, the only setting that had a value in the "calibration" screen was the FF1=1.0, all others were 0.0.
 

It would seem to most likely be an encoder wiring issue. For some reason the X-axis drive isn't seeing a proper encoder signal. Although if the DRO in Linuxcnc is working and behaving properly, it would indicate that the encoder is at least working. So that would indicate that the wiring between the encoder and Linuxcnc is at least ok. You should carefully examine your wiring between the encoder and the drive. Is it ran away from and not parallel to strong noise sources (such as the drive and motor power cables)? How are the encoder cables laid out, does it go encoder=>7i77=>drive or is there a splice point somewhere?

Todd, thank you!!! I was away for most of the day. The wires start at the encoder (to match the US Digital encoder, I soldered and shrink tubed a matching plug with a short length of wire, inside the motor cover) go inside the electrical cabinet (used the existing plug with (new contacts on the inside only, soldered on leads to 7i77) plug) pass through the electrical cabinet.
The signal coming from the encoder travels to the plug/pass through at the electrical cabinet to 7i77 card, at this screw connection a 2nd wire goes to the P1 on A-M-C drive. If I were to revisit this, I would solder both wires together, then put under the screw, now each is independent.
When you talk about keeping away from the "parallel to strong noise sources" how far "away" is a reasonable distance? We are talking about all the "low voltage wires", or especially the "encoder wires"? I have used solid 14 copper wire, twisted the white with the black wire for each motor going from the capacitor and try to keep this by itself.
The 1 splice (to put on the new U S Digital end) is inside the motor cover, the encoder wires are in their own metal spiral wound conduit, so if I were to need a few inches of wire I will need to cut off the conduit or get a longer set of wires.

Tomorrow I will study the plug that goes through the el cabinet (I will look closely to see if any have backed out part way) in addition to the encoder area. It is OK to have the low voltage wires close to the encoder wires, just not the high voltage?

Many thanks

At your screw connector on the 7i77, it is actually better that you did not solder the wires together before inserting. Soldered wires in a screw terminal is bad juju. The solder will tend to creep with time in a pressure fitting, work loose and cause problems. In a screw terminal always use either bare wire or a crimped furrele. (It is less of an issue with a low current application, but a high current one it is really bad.)

How did you make your connection to the P1 connector on the drive? The wire crimp contacts in that type of connector can be difficult to do without the proper crimping tools. (I've made bad ones even with the proper tools.)

Your #1 suspect should be the the encoder wires and ground wire(s) between the 7i77 and the P1 connector on the drive. Unplug the P1 connector and check the continuity of the wires between the terminals of the P1 and the 7i77.

Check to see if there is continuity (0 ohms) between pins P1-2 and P1-11 on the drive. If there is no resistance they are common ground points and only one of them needs to be connected to the 7i77. If they are not common (resistance is more than 0) then you should have P1-2 connected to the 7i77's TB5-3 and P1-11 connected to TB3-3. If they are common, having both connected will create a ground loop. In that case you should probably disconnect P1-11 and keep P1-2 connected to TB5-3.

Yes, it should be ok for low voltage wires to be ran together. If low voltage wires need to cross a high voltage wire, they should do so at 90 degree angles. But avoid running them parallel in close proximity.

Todd, Thanks! Today I replaced the control wiring, soldered the 2 wires together where they meet (encoder & 7i77 TB3 & wires to TB5 & P1 driver) under a screw, the wire was re-purposed from flat screen TV so it was in a larger bundle that was twisted as a set of wires in the TV. The twisted bundle stays together as a single wire for each axes & tied off to one side so can't get close to the DC current, there is some 110 volt closer to the bundles, around 3/4" away.
As I have just powered up and the "x" axes has just a small movement (about 1/8 turn, of less) every 1001, 1002, etc.. count without the belt on, FF1 is 1.0 all others 0.0.
Just installed the belt, the arrow keys work on command to move "x" & "y" with only the FF1=1.0.
I plan to try the tuning (with FF1 set low as per tommylight's servo tuning post) again, set P as high w/o oscillation; add FF1 then go back to P add some then try FF1 again. Never touching the SW2 or 3 or Pot settings.

Many thanks!