OT Motor Slip

I am interested in measuring actual motor power. I think that motor slip is directly proportional to torque, and given torque and rpm hp is readily determined.

The fundamental equation,as I understand it, is simple: actual T = actual rpm/nominal rpm x nominal T

Measuring rpm is simple. Full load rpm and hp (hence torque) is on the motor nameplate. I have some questions about implementation:

1) how accurate is the nameplate rpm?

2) is the full load rpm going to vary across individual motors of the same manufacture, model, and hp or is it pretty consistent?

3) what is the probable accuracy of this technique, assuming my rpm measurement is accurate?

Hi Tom.

Has this got to do with the compressors in your WWT plants ?
When I worked at HV-Turbo it was a test that some customers insisted on.
As you say the RPM is the easy part. It is the Torque that is tricky. We used a force transducer on the shaft between the motor and the compressor. The force transducer emitted the torque value by some kind of radio signal.
I also attended a few motor tests at some of the motor manufacturers. They use exactly the same technique.

I cannot answer your questions because guessing the torque from the slip seems to be not accepted as a test procedure between vendor and customer.
I am away from my lecture books, but I can see that your formula must be wrong. The torque increases as the rpm decreases (the slip goes up). In your formula it is the other way round.

Tom, I don't have any specific figures but, in my experience, nameplate slip is a pretty arbitrary number. I wouldn't be surprised if there was more than 10% error in many nameplates and, that much variation between motors with identical nameplates.

That's mostly speculation tho. I don't have any real data.

I will say tho that when I have an application where the error needs to be minimized, and I have the benefit of independent measurement of torque (several test labs have this capability), I will so a load test of the motor with the drive running V/Hz with no slip compensation. From this test I basically generate my own nameplate data specific to that motor. The results when running in DTC mode are slightly better.

JesperMP said:

Has this got to do with the compressors in your WWT plants ?
When I worked at HV-Turbo it was a test that some customers insisted on.

Click to expand...

You are correct. For large turbo blowers we can justify a lot of instrumentation to monitor performance. On smaller blowers I'd like to get good results but a little more economically.

By the way, are you aware that Siemens has purchased KKK and HV-Turbo?

JesperMP said:

I am away from my lecture books, but I can see that your formula must be wrong. The torque increases as the rpm decreases (the slip goes up). In your formula it is the other way round.

Click to expand...

I stand corrected - I'm sure I transposed the terms in the formula. Fortunately, this is the easy part!

DickDV said:

I wouldn't be surprised if there was more than 10% error in many nameplates and, that much variation between motors with identical nameplates.

Click to expand...

Yeah, that's what I'm afraid of too!

This question is close to two issues where I work.

In the first case we have a clutch that disconnects when a set torque value is exceeded. It also has a switch that is supposed to open when the clutch "disconnects". Since we bypassed the switch we have no problem everything runs fine. I don't like the situation but in this case am willing to go along with boss'es desires.
I checked motor current for several nites motor is running at about 50% nameplate amps and does not feel warm.

Second case we have a large mixer which has two screw type mixing paddles. Driven with motor gearbox and chain drive for each screw. Keep breaking the coupling between final chain sprocket and "screw shaft" have done so now for the fifth time. The cure is to keep increasing size of coupling AND blame the operators for not putting ingredients in mixer and overloading. What I find interesting is we never trip the overloads on the drive motor. Again this is a case of "patch it together we need the production".



What I would like to do in both cases is measure the actual torque and find out what the heck is going on.
The big mixer has a calculated (estimated torque of 660 ft lb). The shaft size is 2-7/16 inch.
The other unit is about a 1 to 2 HP motor shaft size is about 1".

What would cost be for the torque meter and would it fairly easily connect (with adapters etc) to both shafts?

In both cases I am sure glad it is not my machinery and yes I do know full well this is abuse.

So far I have not had to disconnect a safety switch that has employee safety issues. That will eventually happen. Hopefully I will be brave enough to refuse.

Dan Bentler

Dan and Tom. At the risk of being a bit brand-specific, ABB's DTC system is ideal for the issues you guys are raising. As for variations in actual vs. nameplate data, when DTC runs the motor ID run (the standard run, not the reduced or ID Mag run), most of the nameplate error is identified and corrected in the motor model. The result without an encoder is very little speed error (about .1% of motor slip from no load to full load) and good estimation of shaft torque (about 1% of full rated torque).

With that level of performance, you can depend on the drive to limit torque accurately and either reduce speed on overtorque or fault as in "electronic shear pin".

Or, if you want independent measurement of torque, set one of the analog outputs to torque (4-20ma) and use your PLC or other instrumentation for level detection.

Another option would be to set one of the output relays to torque level detection. It will pick up at your programmed torque level and, from that point, what you do with it is up to you.

The key is to choose a drive that can identify shaft speed and torque accurately. Being able to do that without an encoder leaves a lot of extra cost out of the system.

DickDV said:

Dan and Tom. At the risk of being a bit brand-specific, ABB's DTC system is ideal for the issues you guys are raising.

Click to expand...

Oh, if only I had VFDs on all these applications, Dick, your solution would be ideal. I'm aware of the ability of a VFD to give accurate torque output, and that is actually what inspired the original question. My problem is that the applications are rero-fit to existing equipment and for reasons both technical and economic replacing constant speed throttling control with VFDs isn't always an option. Adding a pick-up to sense speed is a low cost alternative.

Tom and possibly Dan also,

If this is a retro fit electronic shearpin application . I think these people can help.
http://www.loadcontrols.com/application_notes/application_notes.html

DickDV said:

Dan and Tom. At the risk of being a bit brand-specific, ABB's DTC system is ideal for the issues you guys are raising. As for variations in actual vs. nameplate data, when DTC runs the motor ID run (the standard run, not the reduced or ID Mag run), most of the nameplate error is identified and corrected in the motor model. The result without an encoder is very little speed error (about .1% of motor slip from no load to full load) and good estimation of shaft torque (about 1% of full rated torque).

With that level of performance, you can depend on the drive to limit torque accurately and either reduce speed on overtorque or fault as in "electronic shear pin".

Or, if you want independent measurement of torque, set one of the analog outputs to torque (4-20ma) and use your PLC or other instrumentation for level detection.

Another option would be to set one of the output relays to torque level detection. It will pick up at your programmed torque level and, from that point, what you do with it is up to you.

The key is to choose a drive that can identify shaft speed and torque accurately. Being able to do that without an encoder leaves a lot of extra cost out of the system.

Click to expand...

Dick DV

Sounds good. Boss is asking for money to put VFD in large mixer (7.5 HP motor 3 ph, 450VAC). This sounds ideal. Tell me more PLEASE.

I don't give a DxxN about brand specificity, nor if you are in conflict of interest, make a profit on em, what I want to hear is
1. Reliability how many have you installed?
2. Uhhhh simplicity of design(??) I guess is the term - can I have a boob install it and get it right? Little overkill here I know but you get the idea.
3. Related to #2 how many other peripherals (?) do I have to install to make this thing work? Yes I know you said no encoder.
4. How well laid out is the manual do I only need to stay on one page to completely enter a parameter or do I need 19 stickys to figure it out?
5. Can I have a visual display of real time torque so operators and maintenance know what is really going on?
6. Can I tie the overtorque trip in VFD to an overtorque alarm lite?
7. Can I program so that ONLY maintenance can reset the VFD after overtorque? I wanted to put in CT's with integral switches for this function but,,,,, With the ABB VFD this would not be needed.
8. Cost and lead time of course.

Many thanks
Dan Bentler

Two questions.

1. Have you ever programmed a VFD before
2. Do you want to control the torque or just limit it? I would assume just limit torque in which case any old drive will do.

Neither option is really that hard to do.

Tom, how about that for reading more into a post than is really there! Sorry about that.

So, now assuming across the line operation, the first issue at hand is to verify the nameplate full load speed which ultimately defines the torque/slip relationship. I know of no other way to do that than to test the speed at a known torque output. This would take a test rack of some sort---a dyno rack actually.

Here is one problem tho. The slip varies somewhat with line voltage. The higher the voltage, the less the slip. So, to hold calibration, you would have to know the operating voltage. And, since this is liable to drift all over the place, the calibration would also.

That pretty much leaves a direct-reading torque transducer as the method of choice. Unless, of course, you can tolerate some error.

Dan, with respect to your questions:

1. Reliability and repeatability are not a problem. Both are very good. I've probably started up +300 of these drives in the last 18 years. This can be said about many brands, not just the one I represent.

2. No, a boob cannot install it. And, a boob can't properly program it either! That's why there are certified startup technicians to do this work. A go-anywhere, do-anything gizmo will never be simple or plug-and-play and I suspect that you already know this.

3. The torque detection and shearpin operation are all part of the drive's standard furnish. I haven't added any options.

4. The manual is good but I can't guarantee one page performance. This is a full-featured drive with lots and lots of choices. I went to school for three days to get certified and, after 12 years of history with this type of drive, I'm still learning new things every week.

5. The ACS800 has three lines of display that can be customized to the application. Display of motor torque in % of full load is common and easy.

6. You can set up one of the three output relays to operate at the torque limit level. That would give you the control you need for an Overload Lite.

7. The drive can be configured so that the Fault reset comes from an external switch and you could make this switch a key-operated unit. The problem with this is that cycling the power aways resets the fault. Also, any other drive faults would also be reset the same way. There is a fault log 64 events deep that keeps a record even after resetting or cycling power. So, the record is there for history purposes. It is not easy or intuitive to clear out the fault log to destroy the historical record. This also is quite typical of most drive brands.

8. As for cost and delivery, I'm probably already close to the edge on promoting specific products on the BBS so, for further discussion on this, why not email me. My address is rbdv01 and then the and symbol and then gmail with a dot and then com. I hope that format confounds the spammers.

Dick
Many thanks. All were what I want to hear except for being boob compatible. Well I can dream but I cannot have everthing.

Will fly by the boss man see what he says.

Many thanks for taking the time to help.

Dan Bentler

DickDV said:

Here is one problem tho. The slip varies somewhat with line voltage. The higher the voltage, the less the slip. So, to hold calibration, you would have to know the operating voltage. And, since this is liable to drift all over the place, the calibration would also.

Click to expand...

I am glad you added that bit of info. In WWTPs voltage varies even more than normal - they are always at the end of the distribution system.

It sounds like "cheap and dirty" isn't going to fly - no surprise there, but like the man says, "I can dream, can't I?"

I will probably just go with amp monitoring on the little guys, and actual three-phase kW monitoring on the big ones. Still not perfect, but easy to implement and it sounds like the accuracy is going to be as good as I can get without a VFD.

And, for those interested, ABB is a top quality drive. ABB, Cutler Hammer, A-B, and Yasakawa (formerly Magnetek), and Danfoss are all made in the Milwaukee area, and ABB can go toe to toe with any of them and whip most.