[pvogtcharlotte]

Folks,
I have three brand-new Allen-Bradley servo drives and a Bob Campbell Breakout board. My plan is to configure the system and operate it with MACH3. My question is about interface the step and direction outputs from the breakout board to the servo drives. All the AB ultra drives can accept step and direction input, but each has two inputs (Step + and Step-) and (Dir+ and Dir-). I can't figure out how to connect this to the breakout board, and I am really afraid of burning up either the controller or the servo drives.
Can anyone help?
Thanks for any advice,
Paul Vogt,
Charlotte, NC

Two of the drives are Ultra 100 and one is an Ultra 3000

[wyobmf]

Hello,
Just wondering if you got your AB drives to work.
I am going to attempt the exact setup soon.
Thanks

[pvogtcharlotte]

WYOBMF,
Yes! In fact this works very well. There is a secret, and I will have to look through my notes in order to find out what it was. It was a particular jumper setting I think. Anyway it works very well, and no additional hardware was required. Send me an email if you like and I'll dig up more info.

I have bought two of the breakout boards, built two operational CNCs, and have had great fun, made money, and had zero problems.
Best Regards
Paul Vogt
Charlotte NC
[email protected]

[coolfox]

WYOBMF,
Yes! In fact this works very well. There is a secret, and I will have to look through my notes in order to find out what it was. It was a particular jumper setting I think. Anyway it works very well, and no additional hardware was required. Send me an email if you like and I'll dig up more info.

I have bought two of the breakout boards, built two operational CNCs, and have had great fun, made money, and had zero problems.
Best Regards
Paul Vogt
Charlotte NC
[email protected]

Paul,
I have Allen-Bradley servo drives and a Bob Campbell Breakout board. You mention on the cnczone about some clewer way to get differential step direction on the Ultra3000 to work from open collector output of the breakout board. Could you please share this invaluble infor to the forum members as there may be some doesn wonder arownd of the same dilemma.
Kind regards

[wyobmf]

If I remember right the Chambel board has active low outputs setup for Gekos and the ultra 3000 uses active high. I think you need an opto-issolator with the Chambel board, I used a PDMX that can be active low or high.

There is no "Secret" Just wire the Common, to the common on your BOB and then Just wire the step and direction single ended.
You can acomplish step and direction control with three wires.

[pvogtcharlotte]

Folks,
I worked this out some time ago with Bob Cambell's help. Here are some notes:

Dear Robert,
I found that once I moved the jumper from M1 to N1, my problem went
away. Also, the step and direction LEDs no longer turn on all at once when the reset button in Mach3 is pressed. Also, the random occurances of the
same fault have stopped.

I'm not an EE so I don't understand what is going on, but it is working
now.
Any final thoughts on the subject?
Thanks for your time,
Paul Vogt,
Charlotte, NC

Then this:
Paul,
What you have done is to turn off the charge pump safety signal.
I will have Jim Cullins the designer of the board to look at it.

Robert Colin Campbell
Bob Campbell Designs

Then this....

Paul,

The charge pump signal is being sent out from Mach. It is a continuous
10 or 12 KHZ signal that is detected on the breakout board. Once the
signal is detected it turns on out-4. You should see the led come on when you press the reset button. The charge pump signal is basically a safety device. If the connection to Mach is broken or the EPO switch is pushed, then the charge pump signalwill stop.

The charge pump signal then enable all outputs and enables the step and
direction signals to go to the drivers.

Robert Colin Campbell

Then this summary:
Robert,
Thanks for the information. Here are some details that someone might
find useful.
1) My application has two Alen-Bradley Ultra100 servo drives (Y & Z) and
one Allen-Bradley Ultra3000 Servo drive for the X axis.
2) The Allen-Bradley Ultra 3000 drive seemed to be immune to the
problem.
3) Both of the Allen-Bradley Ultra 100 drives (100 is an older model)
would quickly fault when the reset was pressed in Mach3
It may take 5-6 presses to trigger the fault, but it would always
happen. The fault also happened sporadicly, for no known
reason in the middle of running a part file.
4) The fault reported by the drives is "Master Encoder Illegal State
Fault"
5) Allen Bradley tech support was only able to tell me "use shielded
cable, it's a noise problem" (gosh, thanks!) (all my
cables are properly shielded)
6) My impression is that moving the jumper to N1 position immediately
solved the problem.
7) The (J24)jumper for the step/direction output is set for GND. This
was necessary to interface to the AB drives.
I may have to double check this, but I'm thinking that the jumper
sets the reference to 0V, meaning the steps are
+5 volt steps, referenced to common, which is at ground. This detail
is from memory, and I will confirm tomorrow.

I hope this is useful to someone,
Robert, I love your product, Thanks for making it.

Best Regards,
Paul

And this suggestion, which I have not done:

Paul,
When the jumper is in M1 and a reset is given the output buffers go into high impedance mode and are floating. With a opto coupled motor drive this is no problem how ever as with you drive having the inputs float makes it very sensitive to noise. In the N1 mode the outputs are always driven so you have no problems. I would try to put a 1k resistor from the step out terminal to the com on the screw terminal on the breakout board.
Sorry for your problem but I am glad you explained it well as I have not had that problem mentioned before.
Thanks
Jim Cullins
Sound Logic

[glentner]

To anybody still following this thread:

I'm trying the same thing, using an ultra 3000 with mach 3 to run my spindle. I've successfully wired it up and mach 3 can turn the spindle at the appropriate RPM, but I'm having a different problem.
It appears that my step increments on my servo are too high, causing very bad vibrations. Has anyone changed any setting in the ultra to make the step increments lesser or greater?

thanks,
Greg

[pvogtcharlotte]

Glenter,
I had to set the scaling in the 3000 to get the proper motion out of the drive. This was easy to do using the AB software. I can help if you're struggling. On my system, the 3000 drives the X axis, and is not nearly as fast as your spindle. I don't know what kind of issues this might create. Let me know if you think I can help.
Paul

[glentner]

Paul,
I think you helped me out at the beginning of this project over a year ago!

My question to you is if you ever set a specific value in ultraware to tell the driver how "far" is one step (one step = 1.8 degrees)? I know how to set the follower ratio, but i'm not sure that will solve my problem.

What is happening is at a slow RPM i can almost count the steps as the servo rotates. The constant start-stop-start-stop of the stepped input seems too "coarse" . This causes bad vibrations at a higher RPM and faults out with a thermal overload.

I'm guessing that if I can set the driver to a much smaller increment per step, that the motion would be smoother. (i could be barking up the wrong tree as well).

Thanks again for the help!
-greg

[pvogtcharlotte]

Greg,
there is a menu called "follower" under "mode configuration"
There you will find a setting called Gear Ratio, Master:Follower.
I have this set to 1:4 for my application.

Also, under Encoders, there is a setting for the Auxiliary Encoder, lines per revolution.
I believe this is the step input you will use for the master. It might give you the
adjustment you are looking for.
Mine is set for:

Encoder ratio, Load:Motor 1:1
Type: Rotary
Lines per Revolution: 2000

I also wonder if the behavior would go a way if you tuned it a little less hot.
Good luck,
Paul

[glentner]

Paul,
Some success. I found a follower ratio that worked, but it's 1:540! Quite a bit different from your 1:4. Changing the Lines per Revolution didn't seem to affect anything that I could notice.

I have another question on your setup. Do you know what the max RPM is of your servos? I believe you said they are axis motors, so I'm guessing it's not too high. I'm getting violent vibrations above 400 RPM (below 400 is fairly smooth).

When I run the motor and set the velocity in Ultraware everything is great. I can run my spindle 2000 RPM smooth as silk. So I suspect it has to do with the stepped input signal. Someone had suggested that i need 1MHz minimum pulsing to get a smooth servo motion, but Mach 3 is limited to 45KHz. That makes me wonder if ANYBODY is running a servo spindle controlled directly from step/dir out of mach 3??????

Any thoughts?

Thanks again for your help.

[keebler303]

Did you do any tuning?

You MUST!

An AB3000 will run smoothly with a step rate of only 1 Hz or less.

[glentner]

Auto-tuning fails for some reason, so no, I haven't tuned it with (or without) my spindle connected.

How could the motor run smoothly when it's only being pulsed once a second (1 Hz)? Wouldn't that be a start-stop every second?

[keebler303]

Are you using an AB motor?
If so, did you select the proper motor?

If not an AB motor, did you add your motor to the database?
Do you know the motor parameters for your motor?

Any error here will make the control loop perform terribly and autotuning has no chance. One thing I like about Baldor drives is they can measure all the motor parameters so you don't need to know it already. I have used a baldor drive to measure a motor and then put those parameters into an Ultra3000.

You can back down the autotuning parameters to prevent some problems. Lower max current is one thing you can do. If I recall correctly, the inner current loop parameters are set by a calculation using the motor parameters. You can get in there and start fiddling with the parameters but for a spindle you shouldn't need a perfect tune.

Since you are operating in position mode (with step/dir) you move one encoder count per step pulse. 1Hz would give you 1 encoder count per second and it would be a "start-stop", but the motion is so small you would likely never even see it. So it takes 1 "step", waits until the next second, then takes another "step". The problem with electronic gearing is it creates extra pulses so in your case you are moving 540 encoder counts, then stopping until the next second, then repeat. That is why you have the oscillation. You want the electronic gearing ratio to stay relatively low. As mentioned earlier, 1:4 may be ok. You also want to setup mach so you are using the full range of step frequencies.

45,000 Hz = full speed on the motor

Use the minimum gear ratio you can to get to your desired speed. I'm sure it is much less than 1:540.

I don't know specifics without knowing your encoder resolution and motor max speed.

Matt

[glentner]

First off, between Matt and Paul, you guys are invaluable!

Yes, I'm using an AB motor, a very common MPL model. Ultraware auto-recognizes and sets the parameters.

The odd thing with the auto-tuning is that it used to work fine. I had 4 of these MPL servo motors and used the auto-tuning to make sure they worked ok when they came in.
I ended up bricking two of the servos when I removed the integral brake so I stopped doing that. But I still used to be able to run auto-tune on the last two motors, never had a problem with that (until recently).

I think i should reset the drive to factory defaults to see if it will then auto-tune.

I see what you are saying with the 1 Hz stepping. That really clears things up with the gearing. Tell me if this makes sense to you:
In mach 3 I have the counts per rotation at 8000. This is why i have a 1:540 electronic gearing (figured out by trial and error). If reduce the electronic gearing and increase the 8000 counts/rotation, arn't I going to hit the limit that mach 3 can output for even a slow RPM?

8,000 counts/rotation * 1,000 rpm = 8,000,000 counts per minute
8,000,000 counts per minute / 60 seconds = 133,333 counts per second or 133 KHz.
133 Khz > 45 Khz (max mach 3 can run) therefore I can't go 1,000 RPM

I'm i think about this all wrong???

[keebler303]

What is max speed on your motor?

Your calculation does not take into account the electronic gearing.

A 1:2 ratio would mean 4000 counts from mach/rev of the motor
A 1:4 ratio would mean 2000 counts from mach/rev of the motor

You get the idea. You need nowhere near 1:540 to get 6000 RPM or whatever your max speed is.

As you change the gearing ratio in the drive, you need to change it in mach to match. The electronic gearing is basically reducing the resolution of the encoder.

That reminds me, you might be able to set the drive to something other than 4x decoding, I am not sure if you can on these or not. That might give you the ability to drop to 2000 counts per rev without any gearing.

Matt

[keebler303]

Also I'll assume that you are powering the brake so that it is off, correct?

Matt

[glentner]

I've finally had some time to fiddle around more with my motor a little more.
I think I have a slightly better understanding of what's going on and I'm back to the original question I proposed:

How does the Ultra 3000 know how far one step pulse is when it step/direction follower mode?

I tried adjusting the aux. encoder counts but that doesn't change anything. The aux. encoder settings must only be used if you put the ultra in aux. encoder follower mode (makes sense). So when i put it in step/dir follower mode, is it moving one step per built-in encoder count? The encoder is 260k counts per rev, so i feel like that's pretty high.

As far as I can tell, the only settings that effect the RPM are the counts out of Mach 3 and the gearing in U-W. When I max out the pulses that Mach 3 can achieve on my system (something like 35khz), I have to gear it up 1:540 in U-W to achieve the desired RPM. This is causing my vibrations due to the course stepping of the motor.

I feel like if there was a way to tell U-W that one step pulse = 1.8deg (for example) that I could then lower my electronic gearing WAY down and achieve a smoother rotation.

As an alternative test, I set the motor mode to preset velocity mode, added a few switches, and can run the motor at 8 different preset RPMS using the digital inputs of the switches. Everything runs smooth like that so my motor tuning seems fine.

!frustrating!

[keebler303]

OK, if your motor runs smoothly in Velocity mode (closed loop velocity) that means your velocity loop and your inner current loop are probably reasonably well tuned. That does not mean your position loop is well tuned. I have a feeling it may be tuned well and the problem is your large electronic gearing.

One step from mach moves the motor one encoder count with 1:1 electronic gearing.
One step from mach moves the motor two encoder counts with 1:2 electronic gearing.
One step from mach moves the motor twenty encoder counts with 1:20 electronic gearing.
And so on.......

Its seems the major problem here is that you have an incredibly large counts/rev number and a slow pulsing engine (mach3). I think you may be up against a brick wall with the step/dir method.

A more logical choice may involve a bit of electronics work, or a couple bills.

Here is the first option: US Digital | Products » ETACH2 High Speed Encoder to Analog Tachometer

I am currently using one of these, although not in the same way you would use it. You can set this box up to accept a 0-35KHz signal from mach (the step pin) and output a 0-10V analog voltage proportional to the step frequency. Then you would use the Ultra3000 in velocity mode with an analog command. This would work for sure.

The cheaper option would be building something like this yourself. It could probably be done for $20 or so but would require some work. You would want to build a frequency to voltage converter, plans can be had from Google.

Let me know if you need more help
Matt

[keebler303]

Another option similar to the US Digital offering, DRG-SC-FR from here: DIN Rail Mount Signal Conditioners

This is more expensive but maybe easier to setup, I have used these on past projects as well. The max input frequency is 10KHz but that would give you plenty of resolution and you can just set the max appropriately in mach.

Matt