OT: PowerFlex 70 Torque Variation

[OkiePC]

Allen Bradley makes great PLCs, we end users love them so much, we think, "All their other stuff must pretty good...", but in some cases, we'd be wrong.

The speed regulation is marketed as a "percent at base speed*", but it is technically an RPM value that should be considered the max error at all listed speed ranges..., right?

So, if you can measure the fact that they can't meet those published specs...that's heavy ammunition for a settlement, full refund, something big to compensate you, the CUSTOMER, for your expen$e$.

Keith, thanks for the FOC info, I will try to follow that when I get to work.

Dick, thanks for bringing the motor math to the discussion, what about gearing and load inertia? Do those factors need to be mathematized?

 

[Manglemender]

This is an interesting thread.

I remember building (not personally) some rewinding equipment for a certain large photographic supplier a few years back and they insisted on using AB drives and motors for the rewinds against our advice. Guess what, they didn't work due to poor torque regulation at low speed. The whole lot was shipped back to the UK from site (Mexico I think?), refitted with Siemens motors / drives then shipped back to site.

Very Expensive.

 

[JRW]

In defense of AB, isn't the pf70 their low end drive?
I know I wouldn't put MM4's on a film casting line. They would be S120

 

[DickDV]

JRW, AB has the PF4, PF40, PF70 and PF700 AC drive products in ascending order of performance. So, you would expect that the PF70 would have fairly good performance specs.

And it does! Except they are false. There is no excuse for that whether high, midium or low performance.

OkiePC, you are correct about the max error at all speeds. There is a bigger problem with stating speed error the way AB does. The motor slip speed is by far the biggest component of speed error so the choice of motor is a key part of the error calculation. For example, if a four pole motor has 1% slip, then the drive has 18rpm of error to manage. But, if the motor is 4% slip, then it has 72rpm to manage. A more responsible way of stating speed regulation in my opinion, is % of motor slip. V/Hz drives with slip compensation can typically get down to one-half of motor slip, sensorless vector down to one-quarter of motor slip, special high-performance sensorless vector like DTC or Mitsi's system can get down to one-tenth of motor slip, and flux vector (with an encoder) would be one pulse per revolution plus/minus one pulse. My point is that the motor has to be included in any estimate of speed error. I recently started up a drive/motor system with a motor nameplated at 1787rpm. That's only 13 rpm slip from no-load to full-load and, before adding any electronics at all, the motor is a 13/1800 = .72% speed regulator! That's really tight and would make almost any drive look good!

As to gearing and load inertia, the figures I gave were at the motor shaft. Passing thru a speed reducer would cut the speed error by the reducer ratio but would also cut the operating speed by the same ratio so the % error would stay the same. Inertia isn't a consideration as long as you are operating at constant speed. But, during speed changes, the torque needed to accel or decel the system inertia (that includes the gears and motor rotor) adds to the load torque and the motor has to produce the total of the two to acheive a specific speed change profile.

And I can't agree more with OkiePC's comment about AB PLC's and the marketing leverage they use to sell everything else. While I'm not even a beginner on PLC's, it is clear to me that the AB models are top of the line. Would that that were true of the rest of their products!

 

[JRW]

Dick- your key word is "fairly"
This is a high end application

So my point is- its the wrong drive/motor for the application.
You get what you pay for.
It may have worked in the past for other material- but not this.
I've worked on some Bruckner machines in the past. This is top knotch equipment. I promise there wasnt PowerFlex drives and low end induction motors on this stuff.
(Im guessing Keith is running film)


Remember- these are opinions on a web site- only Keith knows his application

Good luck

 

[DickDV]

Well, JRW, its pretty hard to select the correct drive for an application when the manufacturer's specs do not accurately reflect the actual performance.

Also, note that the OP tried other brands of drives with similar levels of performance specs and the results were good. I suppose the prices were also similar.

You are certainly correct in saying that underspec'ing the components will lead to disappointment. That's true for any brand and any type of equipment. But, the issue here, I believe, is performance being misrepresented by the manufacturer. That's inexcusable, in my opinion.

 

[JRW]

Well, JRW, its pretty hard to select the correct drive for an application when the manufacturer's specs do not accurately reflect the actual performance.

Also, note that the OP tried other brands of drives with similar levels of performance specs and the results were good. I suppose the prices were also similar.

You are certainly correct in saying that underspec'ing the components will lead to disappointment. That's true for any brand and any type of equipment. But, the issue here, I believe, is performance being misrepresented by the manufacturer. That's inexcusable, in my opinion.

I agree. It should meet the spec.
So what type motor is he using?
You think with a high performance ac motor it would meet it?

 

[DickDV]

In this case it doesn't sound like the choice of motor is an issue. The irregular speed behavior seems to be an internal issue with the drive processor.

I say that because the problem is there even with a pulse encoder. The use of an encoder greatly reduces the burden on the microprocessor (it doesn't have to do the vector math) but the problem is still there. An encoder also reduces the system sensitivity to motor slip since the speed is measured directly rather than being calculated from the vectors and the motor model.

No, I'd say that this is a drive limitation.

 

[JRW]

It will be interesting to see the outcome of this one.

 

[Peter Nachtwey]

Modern DSPs should be able to do the commutation easily

If find it interesting that there are four poles and two spikes per revolution.

 

[kamenges]

Originally posted by Peter Nachtwey:

If find it interesting that there are four poles and two spikes per revolution.

You too, eh? That's what I've been trying to tell Rockwell for a couple of weeks now. But who am I (see about 10 posts back)?

We are using a Marathon Black Max motor. While this isn't absolutely universal I have been told by some who do both high performance induction and permanent magnet brushless servos that for steady state regulation it is easier to control an induction motor since you don't get the reluctance torque pulses. But I'm not 100% convinced of that.

Keith

 

[PLC]

Bump


What's the status?

 

[kamenges]

It's dead. AB said they were going to try a PF700 but that hasn't gone anywhere. We have decided to go with a different manufacturer (Rexroth Indradrive C, in case anyone is scoring at home). It will be a week yet before we get parts. I suspect our guy will follow the stuff out after Easter.

Keith

 

[kamenges]

Just an update on this one:

We had our Rockwell rep come in today with a PowerFlex 700 drive. In testing this drive didn't exhibit any of the symptoms that the PowerFlex 70 did. The PF7000 displayed no velocity variation at low speed even with very low velocity loop gains. In addition the velocity loop gains could be set high enough that at 10 RPM the motor could not be perceptibly stalled by random loading even though the torque output from the drive rose to near 100%.

Rockwell's current stance on the PF70 fluctuations is that there is some type of interaction between the velocity and torque loops that is causing the issue. I don't see this as reasonable but that's OK.

Additionally the velocity regulation spec is based on an IEC (or IEEE) spec that defines how the speed variation is measured. Apparently they look at average velocity over a 2 minute period and base the regulation off of that. So even with the significant variations we were seeing the drive would have probably have measured with very tight speed regulation under this method. I have to believe that most if not all drive manufacturers use this test specification. So beware of numbers.

Keith

Keith