ABB ACS550 Braking Torque

We are starting up a machine in our shop using ABB ACS550 drives. This is the first time we have used ABB drives on a project. We are communicating to them from an AB CompactLogix plc via Ethernet/IP. We have multiple drives controlling multiple s-wrap pull sections on a web line as well as turret and spindle control on a turret rewind. All the s-wrap drives act the same. All the s-wrap motors are 2HP motors with identical mechanical gearing. The drives are set up to operating in vector control (as far as we can tell) and the motors have 2048PPR encoders mounted to them. We have dynamic braking resistors connected to each of the drives.

When the drive is enabled with a zero speed command and I try to move the s-wrap roll the drive will produce a small pulse of torque and will then basically give up. The torque feedback bounces around near zero torque. The drive DC bus voltage feedback stays at about 675 volts, which is what I would expect to get from a rectified 480VAC supply. We have disabled the function that will try to regulate the DC bus by decreasing the torque reference on decel. The positive and negative torque limits are set to +/-300% respectively. When it drives forward the motor produces torque over 100%. When I try to overdrive the roll while the motor is running the drive seems to fight me a little more but it still seems to pulse the torque. And the bus level never twitches.

We suspect we are missing something in configuration but we can't find it. We can't get an ABB guy in here until Monday. Any ideas from anyone would be appreciated.

Keith

Is this a proven, functional design or a prototype?
It reminds me of the DC vs. AC in torque applications debate/conundrum...

I would try slightly adjusting P2603 and 2604 (IR Comp Volt, IR Comp Freq)
I am looking at the ACS550 U2 User Manual http://www05.abb.com/global/scot/sc...file/EN_ACS550_02_U2_UM_C_A4_LowScreenRes.pdf (your VFD model might be slightly different, however the IR Compensation rules apply to all inverter applications).

The VFD has to create an "artificial electric slip" in order to compensate for the lack of physical rotation of the motor...That's where DC applications still rule : Maximum Torque at "zero" motor speed.

From a conceptual standpoint we have done a slew of this type of machine, all of them either flux vector AC or servo. This is the first time we have done it with ABB, however.

I think ABB still produces the most capable torque applications VFDs.
We've used ACS600s for years and recently upgraded to ACS800s.
Maybe all you need is a little more "boost" at start-up from stand-still; P2603 and P2604 adjustments will allow for higher "breakaway" torque.

Originally posted by dmargineau:

Maybe all you need is a little more "boost" at start-up from stand-still; P2603 and P2604 adjustments will allow for higher "breakaway" torque.

Click to expand...

But why? If I were running V/Hz operation I could see why I would need to mess with stuff like that. But if I am in closed loop operation and have completed an autotune (as I am and I have) why would I need to add in a boost value?

Very limited experience with the drives and this comes from the ACS355 range but I think they keep things constant.

Did you enter all the motor parameters (see the chapter on start up)?
Have you set parameter 9902 to 8, Torque control?
Which operator panel are you using on the drive?

I see...and P9904 is set to 1 or 2?...definitely not 3...
What is the exact model/firmware of your VFD?

Well, 1 would be no good either as that is Vector Speed. But using the standard Macro in 9902 puts in all the settings for Vector Torque. Unless of course you were taking the p*ss, in which case 'have a nice day'.

I guess I need to clarify what I am doing.

I want to operate in closed loop speed control. However, I need to have some amount of braking torque capability. I have turned the velocity loop gains up so high that the motor started to vibrate but I still had very little braking torque. The way the drive is responding it is like it is anticipating an overvoltage event and reacting with some protection before that can occur.

I will check the parameter references you indicated and let you know what I find.

Thanks.

All the motor paramters were correctly entered before the autotune was performed.
Parameter 9902 is set to 1 (ABB Standard)
We are using the ACS-CP-EXT operator panel.
Parameter 9904 is set to 1 (Vector:Speed).
The drive part number is ACS550-U1-03A3-4 with firmware 313Dhex as read from the drive operator panel.

So you are actually trying to use the ACS as a "brake" by running it in Vector:Speed control with "zero" speed reference...
Have you consulted with ABB prior to deciding this configuration approach?
This might require a Vector:Torque control mode (with negative torque reference) when the "braking" function is needed; it should be pretty easy to "write" to P9904 if you are running the VFDs through EtherNet/IP-RETA-01 communications integration.
However, until completely changing your approach, I would try to use the Flux Braking feature by setting P2602 to 1.
Looking within the User Manual

(http://www05.abb.com/global/scot/sc...3107/$file/EN_ACS550_01_UM_G_A4_ScreenRes.pdf)

at page 141, the Breaking Torque vs. Speed Reference graphs are quite different depending of the motor's nominal KW.
It looks like (even if not completely plotted), for motors smaller than 2.2KW, the Braking Torque when Flux Braking is enabled and Speed Reference -> 0 has a more acute slope, probably intercepting the y-axis at a higher value.
And one more suggestion: You could try using the DC Hold function-> P2104 (DC HOLD CTL) = 1 when P2105 (DC HOLD SPEED)= 0, however this approach will reduce the motor's lifetime expectancy especially if it is not vector duty...

I am really not explaining myself well today.

I'm not trying to use the motor as a brake. I am trying to have the motor resist and overhauling load. consider a web application where the tension if higher on the downstream side of the pull roll than on the upstream side. In that case the pull roll is running in a forward direction but the torque from the motor is in the reverse direction. If I have my speed/torque graph right in my head that would mean quadrant 2 operation.

The zero speed thing is just the easiest place for me to test motor response. If I command zero speed to the drive I can grab the roll, try and twist it around and feel how the drive/motor fight my movements. Every other closed loop flux vector control drive I have ever use (AB, Danfoss, Bosch, Control Techniques, etc) will try and fight a motor shaft deviation at zero speed. The ABB ACS550 doesn't seem to do that. It also doesn't seem to consistently try to resist an overhauling load when running. This would infer youm could never use the drive in a hoist application.

Well...Probably I am not properly explaining myself too...TGIF...
ABB drives are quite "different" than "mainstream" inverters and, again, IMHO they are the best within torque applications.
We are using them while processing 10*3 lbs steel coils in winder applications, however, switching between Vector:Speed Control Mode and Vector:Torque Control Mode when needed via communications.
From longly investigated past issues, I have "discovered" that their PWM is not constant but variable, depending of the "loading" (ABB could not "reveal" the corresponding dependency function!); so, compared to an Allen-Bradley PF which always outputs at 4KHz PWM (default value, it could be user increased/decreased with the corresponding derating of the inverter)an ABB VFD will not return a commonly expected result.
It seems to me that you will need to run the ABB VFD in Torque Control Mode, at least at the moment you are losing the optimal web tension; if you are fully EtherNet/IP integrated it is a matter of RSL5K code.

You might have to wait until Monday in order to get ABB's opinion about the issue; they should be able to properly configure their own products and maybe you could let us know about the outcome.

I found my issue. As part of some testing that was being done to find the source of a recurring encoder fault we are seeing (that's another issue) the encoder feedback on the s-wraps had been disabled. So, at near zero speed the drive really didn't know that the motor shaft was turning. Once we re-enabled the encoders the motors fight my attempt to rotate the roll almost immediately and the drives go to torque limit very quickly.

I have done some reading about Direct Torque Control and it's implications with carrier frequency (or lack thereof). I've always kind of liked the idea, althought I'm not 100% sure I'm convinced of any significant technical superiority of that control method over standard PWM. In practice, though, the requirements of motor torque generation lead to a relatively consistent pulse pattern being generated from the drive. "The literature" would have you believe the motors will sound like a radio tuned between stations (classic white noise). In my experience there is a definite predomonent frequency that is produced.

Thanks for all the ideas.

Boy, isn't that a doozy...
Glad you are (hopefully) going to enjoy the weekend.
On the carrier frequency issue...I think that "variable depending of the load" makes sense from a manufacturer's point of view -Don't work it harder than it needs to be working!- and also from an end user one: the motor control will be more accurate over the
entire speed/torque ranges since a higher PWM means "tighter" control provisions towards the maximum operating limits.
I am aware that Allen-Bradley has been working on this for quite some time now, to no avail so far...