AC and DC Servo projects with Kaflop+Kanalog

[Myk Bates]

Interesting response. I have my + REF In from my drive to Kanalog 0 On JP11. So I have been using channel 0 and trying to get the drive working from there. When I switched to channel 1 and clicked enable with the default settings the drive took off running. I then changed the axis mode output type to DC servo and got no response from the motor at all. Switching to Step Dirrection on channel one will cause the drive to run but the encoder pulses show up on channel 0. Changing to Step Dirrection on channel 0 does nothing. I'm at a loss for what is going on here.

[Al_The_Man]

I am using the TEST/OFFSET pot on the drive to get it closer. I have tried to get the motor to respond using the step response screen but have had no luck.

You should not use the test/offset pot once the drive is nulled.
IOW the TEST is for offline testing and then the OFFSET should be nulled with a zero input to the drive (+-ref shorted) to set zero drift, do not adjust after this.
Al.

[Myk Bates]

I must have misunderstood. When I read "OFFSET/TEST used to adjust any imbalance in the input signal or in the amplifier." I thought that was what I was doing. The motor was turning slower when I gave it a -50 DAC command then when I gave it a 50 DAC command. Adjusting the pot seemed to fix that. I am probably going about this all wrong though as I have never set up a servo motor before today.

[Myk Bates]

I still cant get the motor to respond to input other than DAC commands. I wired everything up as per my drawing posted earlier but here is a picture too. Maybe someone can point out something that I am not noticing.
...
...
...
or not. After multiple attempts I cant seem to load a 397kb JPEG. It always worked before they made things look different here...

[TomKerekes]

Hi Myk,

It would be helpful if you provided more specific feedback on how well you understood what you were to do, what you did, and what happened.

It isn't clear but I will move on assuming the DAC is controlling the motor properly even though you didn't tell us what happened or what tests you performed.

I think you are probably right that offset pot should be used to center the torque around zero. Set the DAC to zero and adjust the offset pot to get zero torque. It is sometimes impossible to get zero torque/motion. That is an issue with analog signals. It doesn't really matter because the servo will adjust to whatever is required to hold position.

If the DAC output works and the Encoder reading works then the Servo should work if properly configured.

There is an example starting configuration with extremely low gains. Go to the Config Screen, select Axis Channel 0, Push "Load Channel" and select:

KanalogInitialPID.mot

This configuration assumes you are wired to DAC 0 and Encoder 0.

Then go to the Step Response Screen and push "Zero" and "Enable". Does the Motor Servo and hold position (even though very weakly)? If so what do you observe? If not what do you observe?

Regards

[Myk Bates]

Here is where I am at
AMC Drive pots
-Loop Gain - set to off as per the documentation when using Current/Torque mode
-Current Limit - Now adjusted all the way clockwise and back one turn
-Ref In Gain - Now adjusted all the way clockwise and back one turn
-Test Offset - Adjusted CCW till there was movement. Adjusted 6 turn CW until there was movement then CCW 3 turns

Pin Functions
-Ref In to JP11 0
+Ref In to JP11 GND
Signal GND, Inhibit In, + Inhibit in, -Inhibit in all to GND

Encoder
A+ and A- to A0+ and A0- on JP1
B+ and B- to B0+ and B0- on JP1
Index channels to A1+ and A1-

Loaded the file you suggested (KanalogInitialPID.mot) to channel 0 and 1
Entered DAC 0. Servo motor can be turned by hand one direction or the other but returns to position
When a move is commanded from the Step Response screen the following happens:
Channel 1 now has no response
Channel 0 now moves off position and then returns to it.

I am unsure of what to do next.

[Myk Bates]

From what I have read it seems that tuning the servo on the bench is pointless because that will all change once it is connected to a load. I have played around with all the numbers in the step response screen and am starting to get a feel for how they are interconnected with the motor on the bench. I've got smooth slow motion and violent thrashing about. Clamping it down proved to be a good idea. I have also read that each system is so different that there is no rule of thumb to follow or baseline of numbers to use for any particular system.

Based on this it seems that my next step is to connect up limit and e-stop switches and put this all in a box on the mill. Once I verify limit and e-stops are working correctly I can start tuning the servo motors.

I do have one question. The Kanalog sends out a +-10V to the drive. The drive interprets that into motion for the servo. Theoretically I cant send too much voltage to the drive and damage it or the motor no matter what the numbers in the step response screen are. Is that correct? I ask because I am not sure how all the numbers in the Step Response screen correlate to send information to the drive. Can setting Max Limits for output or integrator too high cause damage? As I go into this somewhat blind I want to make sure I am waiving a stick and not a sword while trying to find my way.

Thanks

[TomKerekes]

Hi Myk,

Pin Functions
-Ref In to JP11 0
+Ref In to JP11 GND

Normally the Kanalog DAC output would be connected to +REF. And Kanalog GND connected to -REF. But if the Amplifer input is truly differential then I suppose this should work ok. I would connect them the normall manner and if necessary to reverse the motor drive direction then reverse something else (Change the sign of OutputGain).

Loaded the file you suggested (KanalogInitialPID.mot) to channel 0 and 1
Entered DAC 0.

I don't understand why you are configuring two axis channels to drive one motor. If you configure two axis channels to drive the same DAC to the same Amplifier there will be all manner of problems. This is like having two people telling you to do two different things at the same time. Make sure you only have one axis enabled to drive one motor.

Servo motor can be turned by hand one direction or the other but returns to position

That is a good thing. That is how a servo works. If you move it away from where it is supposed to be it generates a position error that drives it back to where it should be.

When a move is commanded from the Step Response screen the following happens:
Channel 1 now has no response

Not sure why your are using channel 1. Hopefully you don't have two channels enabled.

Channel 0 now moves off position and then returns to it.

That is what should happen. The "Move" test makes a move of the specified size, delays, then moves back to where it was.

I do have one question. The Kanalog sends out a +-10V to the drive. The drive interprets that into motion for the servo. Theoretically I cant send too much voltage to the drive and damage it or the motor no matter what the numbers in the step response screen are. Is that correct? I ask because I am not sure how all the numbers in the Step Response screen correlate to send information to the drive. Can setting Max Limits for output or integrator too high cause damage? As I go into this somewhat blind I want to make sure I am waiving a stick and not a sword while trying to find my way.

Those are great questions. You don't want to damage anything. The "Max Output" parameter can be used to limit how much output (motor current in your case) is ever generated. It is up to you to understand your amplifier and motor and how much current it can handle on a continuous basis. But normally when doing initial testing you want to set the Max Output very low and start off doing slow moves with slow acceleration. Then if working well, increase your speeds and accelerations of your tests. As you increase your accelerations you may see that your Max Output is hit and the axis can't accelerate and there are large errors. At that point you may wish to increase the Max Output. Watch the "Output" (green plot) on the Step response screen to see how much Output current is being used as you make test moves.

Another protection mechanism is the Max Following error. If you set this to a small number then if the servo goes unstable, or you try to move too fast, the axis will fault and become disabled rather than doing something like violent oscillation or a large overshoot.

HTH
Regards

[Myk Bates]

Answers and Questions

I used channel 1 because of what I described on my earlier post. I was getting no response from channel 0 but then when I switched to channel 1 the motor took off. I was trying to rule out channel 1 in this next go around. My wire goes from the AMC drive to JP11 Dac 0. Does this correspond to channel 0 on the step response screen?

I read Al's Servo Tuning Principles and downloaded the pdf. He describes using mainly Kp and Kd to tune with and adding in Ki later to help. I'm assuming that those correspond to P, I, and D on the Step Response screen. I do not understand how to use the other settings on the Step Response screen to tune the servo. There is "Max Limits", "Motion Profile", "Feed Forward", and "Dead Band". Max limits has been described as a limit to protect drive and motor. My drive cannot put out enough current to hurt the motor so I'm not worried about that but I would not want to damage the drive. Tom, can you explain what the numbers mean and how I can use them to calculate output? Or direct me to documentation that does so.

Thanks

[TomKerekes]

Hi Myk,

I believe what happened was that you did a Move on axis channel 1 which was configured as a Step/Dir drive to send out step pulses on JP7. This was probably the worst thing that you could do, it corrupted the communication through JP7 to Kanalog and somehow generated a DAC voltage. But ironically because the motor moved I think you interpreted this as success.

Yes Kp Ki and Kd in Al's article correspond to P I D gains. There are a number of threads in our forum on tuning. Please read through them.

On the Step Response Screen you can click "Help" and then click on the Parameter and a description of the parameter will be displayed.

At this point you should not use any FeedForward or DeadBand. Your current step is to tune the feedback to get good performance without any feed forward or dead band.

The amplifier can accept a full +/-10V input so there is no possibility of damaging the amplifier input.

The Motion Profile settings determine the motion that you are testing. Basically how fast the motion from start to finish occurs. Velocity and acceleration are fairly simple to understand. Jerk is less commonly understood. It is the rate acceleration (torque) is allowed to be applied. It limits the shock that the system feels. If you don't understand Jerk set it to 100 times the value of the acceleration.

The basic idea is to adjust PID gains to get the best performance for various moves of different sizes and at various velocities, accelerations and Jerks.

The IIR Filter screen is a more advanced form of tuning using the Frequency Domain. If you don't understand those at least make sure they are clear.

HTH
Regards

[Myk Bates]

Latest Update, What did I do now?
I spent some time cleaning out the control cabinet of my mill and putting the new Drives and Kflop+Kanalog in. I will be using as much of the old wiring as I can. Everything is set up temporary for now. I will finalize placement once I've got everything working.
I got the Z axis wired and did some small moves to make sure it worked. I then hooked up the limit switches and an E-stop to the opto inputs per my last drawing that was posted. I used the Digital I/O screen and verified they worked. Then I used the Config screen to enable them and that is where everything quit working. Other than the enabling the limits I made no changes to the Config screen. It has the default values. I also made no changes to the Step Response screen which was working fine. Both are set to channel zero. I assume I need to make changes to the Config screen and Download those changes but I dont know what to change.

Can anybody clue me in on what I did wrong?

[TomKerekes]

Hi Myk,

If you enable Limits you need to tell KFLOP which Inputs they are connected to. Do you realize this?

Is there an error message on the Console Screen?

If you disable the Limits do things still work?
Regards

[Myk Bates]

There is no message on the console screen and when I disable the limits there is not change. I enabled limits on the console screen chose "watch limit" and "stop when low" I also left the stop bit as "0" but dont know what that is or how it functions. I have wires connected to opto inputs 0, 1, and 2. They show a check mark under the normal state and are blank when the switch is tripped.

How do I tell KFLOP what inputs I am using?
What is the stop bit and how does it function?

[TomKerekes]

Hi Myke,

I assume you meat to say "Config/Flash Screen" rather than Console Screen?

By "stop bit" do you mean "Limit Switch Options bit no." ? If so the "bit no." is how you tell KFLOP which input bit number the limit switch is connected to. The first Kanalog Opto Input Bit is bit 136.

But it is hard to tell what you did differently such that the Axis worked before and now it doesn't. A common problem is to enable an axis before it is configured properly. This can alter the IO on KFLOP JP7 (for example to drive in the default Step/Dir mode) which will kill all communication to Kanalog. In this case the Kanalog Charge Pump Enable LED will go out.

Otherwise you will need to describe what steps you are doing and what happens.

Regards

[Myk Bates]

Tom your terminology is correct, mine was not. I was going off memory which has its hazards. If Kanalogs first opto input bit is 136 I assume that opto one would be 137 and opto 2 would be 138? If I left zero's in the Limit Switch options bit no. then could this be causing the problem?

I assumed opot 0 was bit no. 0 but it sounds like I need to enter a number. What number do I enter? Under limit switches there is a Negative and a Positive. I am assuming that negative is the limit switch that would be triggered if the axis moved to far in the negative direction and vise versa in the positive? If that is the case, bit no. 136 could be entered in for the bit # under negative and 137 for the bit # under positive.

I would also need to enter a bit # for the E-stop. How do I do that?

By clicking "Download" I entered info into KFLOP memory for bit no 0 so I need to clear that. If I uncheck the boxes under Limits and click download will that set things back the way that they were before?

If I have more than one limit on an axis (ie Z has negative positive and home) how do I enter the third axis? If it were positive would I enter my first two #'s (bit no. 136, bit no. 137) for negative and positive and then click download followed by entering another #, say 138, in the positive and clicking download again?

Let me know if I'm on the right track here.
Thanks

[Myk Bates]

Thought of another question. The max following error is set very high which is fine for testing. Where should that end up once everything is working?

[Myk Bates]

Got home and checked a few things. Here are a couple (cell phone quality) screen shots of the step response and config/flash screens. I also checked and the green light on Kanalog is on. Changing the bit no. from 0 to 136 and 138 had no impact.

[TomKerekes]

Hi Myk,

It looks like you are configured to output to the wrong DAC. If your amplifier is connected to Kanalog DAC 0 then it should be 0 not 8. A logical troubleshooting method might be to look at the Move Plot and see that the Output going to the DAC (green plot) is going to a big value, but you see no movement, so look at how the DAC is configured, wired, etc...

Under limit switches there is a Negative and a Positive. I am assuming that negative is the limit switch that would be triggered if the axis moved to far in the negative direction and vise versa in the positive? If that is the case, bit no. 136 could be entered in for the bit # under negative and 137 for the bit # under positive.

Yes

I would also need to enter a bit # for the E-stop. How do I do that?

There isn't any directly configurable support for an EStop input. If EStop is handled in hardware then it usually isn't necessary. EStop kills power which causes a following error which disables the axes. But you can always monitor an EStop input and do whatever you wish.

By clicking "Download" I entered info into KFLOP memory for bit no 0 so I need to clear that. If I uncheck the boxes under Limits and click download will that set things back the way that they were before?

No. Each time you download to an axis channel you overwrite the parameters. They don't accumulate. This might help.

If I have more than one limit on an axis (ie Z has negative positive and home) how do I enter the third axis? If it were positive would I enter my first two #'s (bit no. 136, bit no. 137) for negative and positive and then click download followed by entering another #, say 138, in the positive and clicking download again?

No. Home switches are handled by a Homing Program.

The max following error is set very high which is fine for testing. Where should that end up once everything is working?

After everything is working check what the normal max following error is. Then set the Max Following Error to a slightly larger value.

HTH
Regards

[Myk Bates]

I'm not sure how that 8 got there. Some where in my trying of different things I guess but that is not the problem. I tried changing back t 0 and downloading that to KFLOP but nothing changed. Any other ideas?

(I did verify that the servo amp is connected to DAC 0)

[Myk Bates]

I clicked move on the step response screen just prior to enabling the limits on the config screen. The servo moved the axis just fine. I had the same parameters in the step response screen then as I do now. The only thing that I changed was adding values into the limits on the config screen.

I've tried different things like disabling the limits, putting in the right bit no 136 and 138 but no change in response.

Prior to enabling the limits I got lots of output waves on the graph and the command an position loops were very close in appearance. Once I enabled the limits I get what you saw in the pictures I posted.

The green light on my servo amp is green and I have changed nothing there. I checked for loose wires and found none. I'm at a total loss. I have no idea what to check.