14 replies to this topic

[GGOSS]

Many of the newer soft starters offered by various manufactureres/suppliers these days provide a 'torque control' start mode. How does this work? How does it compare to voltage ramp or current ramp? What are the benefits of torque control? Is it a gimmick?

Regards,
GGOSS

[marke]

But current ramp and constant current start modes are also closed loop systems that dynamically control the starters output voltage. What differentiates these closed loop systems from torque control??



[Edited on 17/5/02 by marke]

[theDOG]

little buddy,

I think you are overreacting a bit!!
I think the question was how is "torque control" acheived? As marke suggests, current ramp and constant current are closed loop systems, yet they are different (by description) to torque control. What is a "more linear starting characteristic"? Are you refereing to motor speed, motor current or motor torque?
I beleive the term torque control may be a "marketing term" used by the manufacturer to baffle the uneducated!

[marke]

If at any speed, you vary the start voltage, then you are altering the torque by the voltage reduction squared, so you are controlling the torque at that speed relative to it's full voltage torque. The implication is that you are able to define a new torque speed curve that is independant of motor characteristics!!

[Harry Dampers]

The benefits of torque control is that you can accurately ramp up the torque of the motor ......ramping of the current IS NOT the same.

You cannot operate in torque control unless you can actually MEASURE the torque at the motor shaft .......that means you must be able to factor in all of the motor losses , including magnetising current etc.

A couple of suppliers are now offering this ....apparently it works perfectly and motors get away with less overall current than with traditional methods .......works great on Conveyor belts as well.

By varying current or voltage you are off course controlling the torque ......but you are not neccesarily APPLYING torque in a smooth manner

[marke]

Yes, but, I think the point here is that it is still a play appon words. The SS does not directly control the torque, it controls the voltage which then influences the torque. If the actual shaft torque was measured, then you could make that claim as the torque is then the controlling component of the feedback loop. The parameters that are measured are current and voltage, and they do not give you torque. At any one speed, increasing the current or the voltage will increase the torque. Likewise, reducing the voltage and current will reduce the torque, so all are controlling the torque. In the early days of SS, there were a number of manufacturers that had a start voltage control and they labelled that as start torque (well it did influence the start torque) and they calibrated it as 0 - 100% so if you set it on 40%, that meant that you had 40% torque. Not so!! Motors vary dramatically in their starting characteristics and to be absolute, you need to program in the curves for the particular motor in question.
What some manufacturers are doing, is to approximate the kw drawn by the motor during start and from that they are approximating the shaft KW and using this as the feedback stimulae and it works well, but torque is a function of shaft KW and shaft speed. Keep your shafyt KW constant and vary the speed and the torque varies also.

[Harry Dampers]

Mark , that is the whole point , the motor shaft Kw is measured very accurately as is the motor speed .......and therefore torque can be calculated and used as feedback.

At least that is the claim, it makes sense to me .....and it works well in practice ......at least one of the manufacturers uses technology that has been used and proven to protect sensitive and expensive pumps ..typically those of the Magnetic or submersible type , where accurate measurement of torque/ kW is a must .

Apparently this method of torque/kW sensing is so accurate that it can replace "tilt switches"and other sensors in sump pumping applications .........ie. by meausuring the load at the pump shaft , it can determine when the pump is running dry (and switch off).

[GGOSS]

Hello Harry and Mark,

I must say when I posted my question re Torque Control I did expect that it would generate some friendly argument. It also appears to be a topic of interest to many recording some 78 hits to date.

My understanding is that products offering this feature and operating on a 'sensorless' basis are imbedded with typical motor data and equivalent circuit, thus taking into consideration all aspects of motor operation. However I do stress the word 'typical'.

I can advise without hesitation that a major and well-respected supplier of pumps to Australian industry has conducted its own independant research and has found that Torque Control and Current Ramp provide almost identical performance in 'real world' pumping applications.

The tests (which included chart recordings of flow, pressure and current) were conducted on a 150kW submersible pump with approximately 150 meters of head. Their comments were, "when the chart recording were compared (through an overlay technique), it was near impossible to pick the difference between the two technologies.

On the subject of pumps however, we must be aware that many submersible motors are fitted with 'thrust' bearings in lieu of roller bearings. Where thrust bearings are used, there is a compromise between soft start performance (minimising electrical and mechanical stress) and bearing service life.

As the thrust bearing relies heavily on the fluid being pumps as being its lubricating medium it is essential that the motor is accelerated very quickly. This acts to lift the rotor section of the motor thus allowing the fluid being pumped to pass over an lubricate the bearing before the rotor returns to its normal operating position as the motor approaches rated speed.

Needless to say faster acceleration and deceleration times result in increased pressure surging and possibly higher starting currents, hence the above-mentioned 'compromise' in starter performance.

Going back to the original question, if there is anybody out there that can provide a detailed description of torque control functionality (how it works) and/or put me onto a related technical paper, it would be very much appreciated.

Regards,
GGOSS
PS: Harry, welcome to the forum. I look forward to seeing more posts from you in future.

[jraef]

GGOSS,
I realize that this thread is now old, but I know you check in here often and I do have something to add on the subject.

As a purveyor of such a feature, I am unwilling to share intimate details of design but I will say that we use the microprocessor to perform PID algorithms from closed loop feedback on the voltage control and current control, with one pushing the other to maintain a linear ramp of kW as was previously mentioned. Our normal current ramp is a little less sophisticated, using hardware components to maintain maximum voltage without exceeding the current profile. Does it sound like the same thing? You bet it does.

That said, I am willing to divulge that we too found little or no appreciable difference in field tests. One of our competitors makes a big deal about how this feature prevents the pump from surging at the very end of ramp as the motor pulls in. We have found that a) it doesn't happen all that much, and current ramp and even voltage ramp with current limit and a good anti-oscillation circuit perform equally as well. There is nothing wrong with torque ramp, but it isn't the end-all solution to every problem as marketing gurus would have you think.

My theory is that once some manufacturers went to an all-digital firing circuit they were forced to do this in order to accomplish what analog systems have done for years. Then marketing put a spin on it as a "must have" feature.

The only application where there seemed to be a slight difference was on long overland conveyors, where belt stretching is a problem. Torque ramping provides a more linear acceleration profile which helps reduce the "rubber band" effect that can happen on these belts. Better yet of course is a tach-feedback ramp because it is truly linear with regards to speed.

[GGOSS]

Hello jraef,

Many thanks for your valued input...and for confirming my suspicions.

By the way, I don't believe it's ever too late to respond to any posted question, particularly if the answers or opinions will assist in expanding the knowledge base of the this forums members and visitors.

Thanks again and regards,
GGOSS

[Igneous]

Current ramp and constant current start modes has a current feedback... To have a torque control it's nessesary to have a torque feedback and therefore torque measurements (direct or indirect) are also needed.

[water]

I've had cause to look into torque control over the last year or two although I've only explored it with one manufacturer. Their patented system is to monitor the phase angle of the motor current during starting (and stopping) and apply that to a motor model to derive the torque at the shaft. I read a technical description they sent me and it all made sense. My interest is not in starting pumps but stopping them to reduce waterhammer, and my experience so far says that the performance is chalk and cheese. Straight voltage ramping, regardless of the set up options on the starter, cannot effectively ramp down a centrifugal pump. The interaction between the pump torque curve and the motor torque curve means that you cannot get any kind of useful performance in terms of water hammmer. The motor will quickly fall from near full torque to a fraction of full torque (stall condition) such that little benefit is achieved (in my experience). This could be different for speciallist motors with a non-standard torque speed curve. However, the torque control adds an extra feedback path that stabilises the ramp down with a conventional motor. The feedback essentially achieves a controlled stall of the motor holding it in that region of speed that is otherwise unstable. Practical trials showed good results. The tradeoff is the prolonged high motor currents because the motor is in being held in a bad operating point. Expect currents nearly as high as your peak starting current when the motor was getting away. The difference is that these currents are held for perhaps 60 to 70 percent of the ramp, so if you have a 15 second ramp you could expect to see maybe 2 to 3 times full load current for around 10 seconds. In my trial, a 20 second ramp produced I2R effects 4 times as high as the start cycle during the stop. Hence the number of starts and stops becomes super critical if thinking of using a soft starter to do a real soft stop. I say this because no mater how you achieve it, if you are not controlling frequency as well as voltage, then your cannot achieve a meaningful soft stop without these necessarily high currents.

Cheers

[marke]

Hello water

Welcome to the forum.

There are a few manufacturers who offer "torque control". Some of them are just using marketing license to get the torque control, others are using a motor model to give the necessary feedback and are able to give a much more predictable performance.

I wonder who the manufacturer was that supplied you with the technical paper, it could be an interesting read.

Best regards,

[jraef]

Hello water, welcome to the forum.

While I can appreciate your skepticism, I must disagree with you on one point: your statement that decel control to reduce water hammer is "chalk and cheese". It works, I have seen it work, lots of people are benefiting from it. Is it a panacea? Absolutely not, but neither is it chalk and cheese. It does need to be set up correctly and there is no magic formula that can predict every nuance of pump control, but I have personally set up and adjusted many many soft starters to reduce water hammer to the point of being undetectable (to the casual observer). The trick is, however, that as much as possible must be adjustable. Some manufacturers pay "lip service" to the concept by just putting in a decel ramp time adjustment, but many others offer configurable versions that perform very well.


Side note on the Torque Control issue:
I now work for a manufacturer who provides this and I have learned a great deal more about the inner workings (although no intimate details). I used to think it was a form of kW control based loosely on power factor, but as we know, pf is non linear and has a relatively steep drop off point at around 50% loading. If you think about it, current has a similar issue albeit almost mirrored in that a current/load curve looks flat at the low load end, increasing as load increases. What the "torque control" algorithm does (ours anyway) is to calculate what is referred to as "active current" by using both curves to attain a more linear relationship to torque output. It's not perfect, but it's very close and repeatable, providing a very smooth acceleration ramp. How much more useful that is compared to a simpler voltage ramp or current ramp is, IMHO, still open to discussion.