I think you need to explain a little more on what you are tring to do, what you are trying to hold and what torque is needed.

Not only that, but I would think that would be a question to ask of A-B. In general a Closed Loop Flux Vector drive on an Inverter Rated AC induction motor can produce up to maximum locked rotor torque of the motor at zero speed. There are a lot of "ifs" in that statement however, both for the VFD and the motor, and only the manufacturer of the drive (or someone deeply versed in it if you're lucky), can really answer that definitively.

Eng-Tips: Help for your job, not for your homework  Read FAQ731-376

I am not sure of the required holding torque at this time since it is early on in the design.  I was mainly curiuos about using VFD's and 3ph induction motors in a point to point motion application.  I have read alot of correspondance regarding the similarities/differences of servos systems and vfd's, none of which touched on maintaining position with holding torque.  I guess I can always use a brake, but they are prone to mehanical wear with extensive use.

Thanks gents

By holdingtorque, do you mean resistance to an overhauling load, resistance to turning, or holding the shaft motionless as with a mechanical brake?
yours

If you are looking at the difference between a VFD and a servo for point-point motion control then it comes down to your required accuracy and repeatability. Of course a servo control is going to be better but you pay more for it. Look at your application requirements first before your try and fit a product into it otherwise the 'perceived' savings will not happen. The investment in the system right for the job will always pay dividends.

I ment holding the shaft motionless, acting as a mechanical brake.

Thanks

You are still way short on useful info 450x.. Notably waross's questions.

You are possibly describing a vector drive as compared to a scalar drive the "standard" VFD.  They can hold loads all day long. With a bunch of conditions which we cannot guess about with your meager info.

Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

Sensorless vector systems are not good at developing torque at zero speed for more than a second or two.  If you add an encoder to the motor shaft and shift to Flux Vector control, nameplate rated torque at zero speed is possible for extended periods of time.  That's assuming you have provided cooling for the motor under these difficult conditions.

Generally, AC drives are not good at positioning in the sense that they have no way of keeping track of where they are or where you are telling the load to move.  Servos do that routinely.  Having said that, it often possible to get acceptable servo-like action with a precision AC drive following a positioning controller in a PLC or stand-alone.  Some would call that a single axis controller.

Hope that helps

An important issue to also consider is that even if you did design a system with a Closed Loop Flux Vector drive and a properly cooled motor to hold a load motionless, it's never a good idea from a safety standpoint to rely upon that if the nature of your application involves risk of injury or damage should the power fail. The "perfect" combination is to use the VFD/motor for normal operational braking, and then have a mechanical safety brake anyway in case of a power failure. Crane hoist designs require that of course, but an often overlooked aspect is to consider the mass in movement at the exact moment of a power failure. Is it going to coast and crash into something or someone? Is gravity involved? Is inertia of a dangerously spinning load such as a blade involved?

This is why we have been asking you to give more details. As I said early on, lots of "ifs".

Eng-Tips: Help for your job, not for your homework  Read FAQ731-376

The Yaskawa G7 drive with an encoder has a feature called "zero servo".  I have had a chance to play with this drive a bit, not the zero sevor, but Yaskawa says it can hold the motor like it had a machanical brake.  The G7 after i have played with it shows that it has some good control in the slow speed area(open vector 2)

Quote:

...but Yaskawa says it can hold the motor like it had a machanical brake.

But not without power it can't!

That's a very good and important point!

If you tell a machine designer that this drive can hold the load like a mechanical brake - and he takes your words for it - you are both in for a very unpleasant surprise the day (or night) you get a power outage.

Gunnar Englund
www.gke.org

Yeah one of those..
"I wish I'd
t a  k    e     n
T    H          E
S              T             A            I                     R                      S  !!!!!!!!!!!!!!!!!!!!!!!"

Nites.


Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

Yes I understand that.  It will depend on application.  He asked I had answer.  That is for him to take a look at not me.  Im just giving options.

I had done that successfully with ABB ACS600 series inverter.

I had a motor driving a cryogenic(liquid nitrogen) pump at 4600rpm and we need in inverter to run the motor at 77Hz (This had saved a mechanical gearbox).

The pump is required to operate at -150deg.C and it is only required to be started intermittently (once in a few days).  Before the pump can be started, we need to cool down the pump from atmospheric temperature to -120deg.C.  Before the pump is cooled down to -120deg.C, we don't want the pump rotor to rotate or else the pump seals will be damaged.  Therefore we applied brake on the motor through DC current from the inverter.  This braking function is available in the inverter control menu.

Our VFD driven motor has a separate motor driven cooling fan.  Therefore, when the motor is under DC current braking condition, it is still being cooled by the fan.

However, the DC holding torque (current) level is adjustable as you wish.  Base on our experience, the DC holding current level we set is not high and I hardly feel very much warmth on the motor body during DC hold, I don't think the winding insulation will fail in the event of cooling fan breaking down.

I don't have a Powerflex 700 drive but I believe electrical braking should be a standard feature in that drive.

If you download Powerflex 700 User Manual from Allen Bradley website:
http://literature.rockwellautomation.com/idc/groups/literature/documents/um/20b-um002_-en-p.pdf
I am sure you can find more details about Stop/Brake Stop/Brake mode.

From the last couple of posts, it seems that the subject has shifted to DC Injection Braking.  That type of braking will tend to hold an induction motor at zero speed but not with very much torque since it takes rotor slip to magnetize the rotor.  So, for DC injection braking, you will get a little drift speed as holding torque rises.

It is also true, as the poster above mentions, that the ABB ACS600 in DTC mode, can hold a load at zero speed.  But you must be careful since this is only true for a couple of seconds.  After that, the drive looses control of the motor torque and will release the load.  Addition of an encoder will fix that but, as has been stated many times already, this is not a suitable brake technique for safety purposes.  For that, trust only a spring-set mechanical brake.

Thanks DickDV.
That's the first honest explanation of the holding torque issue with DTC drives I have heard from either ABB or competitors, and now I understand. A number of ABB marketing people leave out the "couple of seconds" issue, and customers have experienced some disastrous failures as a result. On the flip side, a lot of their competitors only mention the part about it losing the model, allowing people to believe that it can't do it at all. Your explanation puts the two incongruous stories together and makes much more sense.

Eng-Tips: Help for your job, not for your homework  Read FAQ731-376

A very easy "demonstration" to do at trade shows is to grip the shaft (there used to be a hand-wheel) of a machine that is being kept still by a DTC system. Make a mark on shaft and end-bell. Then hold the torque for a while and you will see how the shaft creeps away from the mark.

I did that when the DTC was introduced many years ago. The "stand soldiers" immediatley switched the drive off. They told me that it was too risky - the motor could start spinning when someone was holding on to the wheel and someone could get hurt.

I refrained from asking them if they didn't trust their system - running away just like that doesn't sound very good. I think that it would have been too tough on them.

Gunnar Englund
www.gke.org

DickDV;  Can you explain to me why the drive would be able to stop the load for a few seconds then lose control?  I can see no logic to this at all!  I believe you, but cannot even vaguely picture the why.

Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

Smoked,

See my contribution above. It has limited control from the first millisecond. But you can only see it after a few seconds of creeping.

Gunnar Englund
www.gke.org

I see that skogs but I want the why that was the how.


Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

450x
As you can see, there are numerous methods of 'holding' a motor shaft, whether it be for a few seconds or until smoke appears due to too high an injected DC current. The main point is that if your application requires a brake that, if power failed and gravity took over, then what would you require to happen? As already mentioned, no matter how good the supposed performance, it means nothing when the main source of power has gone.
Mechanical wear or not, most brakes are designed for failsafe and this is considerably more failsafe than the VFD.

The WHY is simple. You need current in the rotor to get any torque. And if the rotor is at standstill, the only way of getting current in the rotot is to produce a slow three-phase systems that rotates against the torque. If this system rotates too slow, the shaft will move. If it rotates too fast, the shaft will rotate against the torque. Both are bad, but rotating with the torque is less bad.

Now, if there was an encoder that could tell the drive where the shaft is, than it would be easy - and works very well. That's why you always need an encoder in hoisting applications.

But the DTC people (sometimes) claim that it does not need an encoder to keep a hanging load. So, there you are. A motor without encoder. A torque. A motor model that is valid at some temperature - but not for a wide range.

Producing torque means rotor current. Current means heat and heat means increased rotor resistance, which means that the motor model isn't valid any more. In bad cases, the creeping changes to a fast rotation and that's where the drive loses control and the whole thing goes berserk.

Gunnar Englund
www.gke.org

Producing torque also needs movement. A true flux vector drive is zero torque at zero speed.
You may recall Gunnar, before the advent of the digital signal processors, Siemens actually used the bang-bang technology now know as Direct Torque Control but moved on from that about 15 years ago or so.

Yes, that's right. The DTC modulator, in effect, keeps the flux vector between two tolerance circles. Using a two-point controller technique, aka Bang-Bang.

Gunnar Englund
www.gke.org

Hi Gunnar
A couple of questions,
With an encoder used as you describe, do I understand that as the model becomes imprecise, the shaft will turn a small amount. The encoder responds to this small movement with the appropriate adjustment to hold the shaft almost motionless.
Also, what is the effect of changing the load on a stopped hoist motor. That is, people getting on or off an elevator or a concrete bucket being loaded or dumped?
Is a position encoder mandatory for these applications and is it good practice to add a mechanical brake?
Thanks
Respectfully

Thanks guys, now I see.

Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

Yes waross. Never build any hoisting device with an open loop VFD. And don't forget the mechanical brake. It is mandatory. Not only good practice.

Gunnar Englund
www.gke.org

Thanks skogsgurra
Respectfully.

itsmoked, probably better to go with skogs explanation that to try to contradict him.  I've tried that before and found the reaction a bit extreme.

While there seems to be some delight here is bashing DTC, I have yet to find any sensorless vector system to compare to it.  As long as you use it within it's limitations, very good performance results and I have built my reputation on it for the last eleven years.  Of course, the marketing people tend to get carried away with their claims and that does no-one any good, in my opinion.

Just for an example of the capabilities that I see in DTC, I have about two dozen hydraulic pump (aircraft) test cells operating from 8000 to 20000 rpm and from 50 to 500hp that routinely must go from no load to full load with a 3ms torque loop update and with a total speed error of not more than 10rpm +/-.  Properly commissioned and tuned, this is acheivable without an encoder on the motor.

I'm satisfied that this is remarkable performance and have no trouble selling it day after day.

I believe DickDV has misunderstood about DC HOLD and DC braking.  These are the texts copied from the ABB manual:

Quote:
2104  DC CURR CTL
      Selects whether DC current is used for braking or DC Hold.
      0 = NOT SEL
          . Disables the DC current operation.
      1 = DC HOLD
          . Enables the DC Hold function. See diagram.
          . Requires parameter 9904 MOTOR CTRL MODE = 1 (VECTOR SPEED)
          . Stops generating sinusoidal current and injects DC into the motor when
            both the reference and the motor speed drop below the value of parameter
            2105.
          . When the reference rises above the level of parameter 2105 the drive
            resumes normal operation.
      2 = DC BRAKING
          . Enables the DC Injection Braking after modulation has stopped.
          . If parameter 2102 STOP FUNCTION is 1 (COAST), braking is applied after start is removed.
          . If parameter 2102 STOP FUNCTION is 2 (RAMP), braking is applied after ramp.

2105  DC HOLD SPEED
      Sets the speed for DC Hold. Requires that parameter 2104 DC CURR CTL = 1 (DC HOLD).

2106  DC CURR REF
      Defines the DC current control reference as a percentage of parameter 9906 (MOTOR NOM CURR).

2107  DC BRAKE TIME
      Defines the DC brake time after modulation has stopped, if parameter 2104 is 2 (DC BRAKING).

Unquote.

I have seen the DC HOLD working at my pump as long as the speed reference is set to below 100RPM (this is my setting), and is not limited to short duration. I can meaure the DC current being injected into the motor winding as long as the motor is under "DC Hold" condition.

And ABB gave a warning notes in the Manual: "Injecting DC current into the motor causes the motor to heat up.  In applications where long DC Hold times are required, externally ventilated motors should be used."

I believe I understand the issues here, digitrex.  DC braking can go on for as long as the motor remains under its thermal limit and has no relationship to DTC at all.

Our discussion on the capabilities of DTC, on the other hand, involve the drive controlling motor torque at zero speed as a brake.  That's where the short term limit is encountered unless an encoder is involved.

I suppose that you could say that DC injection is a non-synchronous method of forcing zero speed where the use of DTC or any other sensorless vector attempt at zero speed is a synchronous or near-synchronous forcing of zero speed.

The important point in all this is that neither is acceptable for safety stopping and holding purposes.

sed2developer:

"Producing torque also needs movement. A true flux vector drive is zero torque at zero speed."

I can't agree with this at all. True flux vector drives can produce full torque at zero speed. For flux vector drives, there is nothing special about zero speed. They compute a slip frequency proportional to the desired torque, and produce an electrical frequency equal to this slip frequency plus the mechanical frequency (rotor speed).

However, as several have pointed out, a shaft sensor is almost certainly required to detect zero speed well enough to really hold position. Most "sensorless vector" drives use back EMF to estimate velocity, and this goes away near zero speed. (There is a lot of interesting research into injecting high frequencies into the motor and observing the response to get true position control without a shaft sensor, but I haven't seen good commercial implementations yet.)

I will also agree with the assessments of ABB DTC here. Great technology for many purposes, but occasionally oversold by their marketing people.

Curt Wilson
Delta Tau Data Systems

Yes, those are the important words: "occasionally oversold by their marketing people" and that's why DTC gets some beating from technicians.

Gunnar Englund
www.gke.org

Curt
rather than disagreeing, I think you are agreeing. Is there any point of full torque at zero speed? There is a point to full torque at any change from zero speed.
The accuracy of vector performance at low speeds is the question and the ability for drives to perform the vector calculations at frequencies <4 or 5 Hz is questionable. Most have to revert to speed control at frequencies sub 4Hz.

digitrex, I just now see your point about DC Hold and DC Braking.  I tend to call both "DC Braking".  ABB's intention, I believe, is to refer to DC Braking as dynamic braking as in slowing a load down.  DC Hold refers to holding a motor at standstill without any intentional rotation.  Both are independent of DTC as I mentioned above and use DC injection which can go on for as long as the motor remains within its temperature limits.

sed2developer:

"Is there any point of full torque at zero speed?"

Yes there is! It is vital for crane/hoist applications, as some have mentioned above. The cranes that lift the Space Shuttle on and off the special 747 to take it from a California landing back to Florida use flux-vector controlled induction motors, and you can be darn sure that NASA cares about full torque at zero speed!

There is a vital distinction to be made here between flux vector control with a shaft sensor and flux vector without a shaft sensor (so-called "sensorless vector"). Nothing in the flux-vector control algorithms using a shaft sensor "falls apart" near zero speed, so they are fully capable of maintaining full torque at or near zero speed. And if a position shaft sensor (encoder or resolver) is used, a position loop can be closed to hold true zero speed with zero steady state error in the speed (because the position loop provides integral velocity control).

However, the flux vector control algorithms without shaft sensors run into two problems near zero speed. First, the back EMF they use to detect speed gets very small, so the signal-to-noise ratio gets horrible. Second, they have no position sensing capability to provide the automatic zero steady-state error in speed that a position sensor provides.

Curt Wilson
Delta Tau Data Systems

This is one of the more fascinating discussions here I must say. It's what makes this forum so wonderful.

Quote (digitrex):

I have seen the DC HOLD working at my pump as long as the speed reference is set to below 100RPM (this is my setting), and is not limited to short duration. I can measure the DC current being injected into the motor winding as long as the motor is under "DC Hold" condition.

While I don't doubt your experience, you must understand that we are talking about very disparate applications. Holding a pump impeller still is VERY different from suspending a Boeing 747 from 4 hoists in mid air and releasing the mechanical safety brakes! I have done that with Flux Vector Drives using encoder feedback and a torque proving feature found (for a long time) only in one particular version of Yaskawa drives (Electromotive) and a few other brands modified with Delta Tau front ends (CSWilson's company). Since then, many other companies have perfected torque proving (by a variety of names) and it is widely available and very proven in the field. Whatever the hype, DTC or DC Hold will NEVER be safe enough to accomplish that. ABB came in with their ACS600 with DTC at the time we were doing that project and begged Boeing for a chance at it. We lifted a 40 ton test weight, set the safety brake, released the brake, and the DTC drive dropped the load. End of story, they never were given a 2nd chance. The Electromotive drives held it locked in place, you could not even detect movement when the mechanical brake released.
Note to sed2developer. I can (from the above experience) attest to the validity of "full torque at zero speed" with an encoder feedback.

Eng-Tips: Help for your job, not for your homework  Read FAQ731-376

I've learned a lot from just lurking on this thread - thanks, everyone.  

Dave

450x

You want to use an AC drive for a simple positioning application.  As everyone has pointed out you'll need to have encodder feedback to the drive for it to be able to hold the load at a position accurately.  I have recently been playing arround with an ABB ACS800 drive with the position software installed.  The motor is 11.5kW with forced cooling and a 1024pulse encoder (two chanels and zero pulse).  The positioning accuaracy is fantastic, and in that particular application we are loading up the positioning drive/motor with an identical motor on a regen drive in torque control.  It was very entertaining to have the position drive hold a position and then dial up a  torque on the regen drive.  There is no observable movement of the motor shaft, yet the drive displays were showing lots of Amps and matching torque values, as you'd expect.

If your application does not require the accuracy of a servo drive system then I would urge you yo consider the postioning drive.

Many thanks to everyone for your input, as always it was very imformative.


Martin

Hey, thanks for responding. So many times we go skiping along down our tangential paths around here, we lose sight of an OP's original question. Glad to see you were paying attention and appreciate it.

Eng-Tips: Help for your job, not for your homework  Read FAQ731-376

450x, as jraef say, thanks for responding.
now quickly down my tangential path...
Curt, I obviously bow to your far greater knowledge on this subject and was by no means trying to counter your description. It was my description in this thread that was wrong(quite possible) or I'm fundamentally wrong (also quite possible). I was simply trying to say (maybe if I put it the other way round) that 'with' movement in a closed-loop vector drive then the function of generating the active current, and therefore 'holding' the shaft stationary (if your setpoint was zero) will occur. Using your hoist with the 747's as an impressive example, the 747's are trying to create the error in the rotor angle and it is there you find out how well your drive performs.My point earier about zero speed and zero torque was simply (and maybe wrong) that if there was zero setpoint and no error, then certain drives will overflux due to the controller oscillating, effectively looking for an error. So the torque generated is not actually necessary but an indication of the performace of the control algorithm. This is the point where I think, "do I hit submit post and make a complete d*&k of myself", but as you can see I did and therefore the need to learn is a stronger urge than the other.
   

sed:

Not to worry. The whole point of this forum is for everyone to learn (except me, of course...)

Let me say at the outset that I consider a flux-vector controlled AC induction motor using a shaft position sensor to be a real positioning servo drive. I know lots of people who use them as such. The only thing they really give up to what most people consider to be servo drive/motor systems is that they have substantially lower torque-to-inertia ratio, because the rotor moment of inertia is typically 4 to 5 times higher than for a permanent-magnet brushless servo motor of the same power rating.

In our own positioning controllers, only two setup variables need to be set differently for controlling induction motors as opposed to permanent-magnet brushless servo motors. First, induction motors require a non-zero "slip gain" (slip-to-torque ratio), whereas for PM servo motors, this parameter must be zero. Second, induction motors require a non-zero "magnetization current" command (direct current) in order to induce current, and hence, a magnetic field in the rotor. This is not required when the rotor has permanent magnets creating its field.

Now, a positioning servo drive, with either a PM or an AC induction motor, can sit all day at zero position error, whether or not there is an external load. It does not need to hunt. The key is integral gain in the position loop. The integrator can "charge up" so that a torque command is output even in the absence of an error at the moment.

It is a common misconception that servo drives need to "hunt" dynamically to hold position. A good servo system can sit fat, dumb, and happy all day long at zero error if there is a constant load (zero or non-zero). Hunting is usually a sign either of response to changing disturbances or bad setup.

Curt Wilson
Delta Tau Data Systems

thanks for taking the time out for the explanation!

Very interesting Curt.  Thanks.  Usually I hear "charge up" as "wind up".  I don't understand how DC current in the stator can induce something in the rotor or is this like a circular magnet with a few gaps in it and only one section has a winding(the stator)?

Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com

It isn't DC current in the stator. It is a slowly rotating flux. Three-phase system at low-low frequency. Just enough to produce the torque needed to hold the rotor.

Gunnar Englund
www.gke.org

itsmoked:

I wasn't trying to speak terribly precisely, and I used "charge up" as a colloquialism. In an analog system, the voltage on a capacitor typically does charge up to perform the integral function. In a digital system, the number in a register gets bigger.

But "wind up" usually refers specifically to the "charge up" of an integrator past where it can do any good -- that is, when the servo output is already saturated at its maximum magnitude. This can be very problematic, because the very large value in the integrator can cause the servo to overreact as the servo comes out of saturation. Most decent servo algorithms now have some kind of "anti-windup" feature that prevents the integrator from charging up further if the output is saturated.

Skogs clarified the nature of what is happening to create holding torque nicely. At 0 rpm (0 Hz) rotor speed, the stator frequency may be 2 Hz. As far as the rotor electromagnetic dynamics are concerned, this is no different than a 58 Hz rotor mechanical frequency (1740 rpm for a 4-pole motor) with a 60 Hz stator frequency. The rotor "sees" a 2 Hz slip in either case.

Curt Wilson
Delta Tau Data Systems

skogs and Curt,
Okay so the DC isn't actual DC.  Thanks for the clarification. Makes sense to me now.

I see your "charge up" as it describes analog/capacitor PIDs actual functioning methodology (analowhat?).

Windup/anti-windup I understand as half my work is coding PIDs into embedded systems.  I have fought that war many times/ways.

Keith Cress
Flamin Systems, Inc.- http://www.flaminsystems.com