VFDs, v/hz, sensorless, full vector for machine tool

[jimmy3456]

Basically I have a machine that has an open loop or v/hz type spindle drive. Has been installed for many years but just not happy with the performance. The drive just cannot respond to all the varying loads and bogs easily.

I have decided to install a difference drive with a feedback circuit. For this application, around 1% or so of rpm variance would be tolerable but load can vary quickly and motor needs to respond quick to keep spindle at speed.

Would todays sensorless systems be a decent solution or do I need true vectoring with an encoder? Will true vector really shine well above? My target is to get bad fast accel/decel rates as well as reasonable torque response for cutting loads for a machine tool.

Due to the high costs and this old machine, I would prefer a used drive for this. Because I do not need tight rpm control, a simple tach would probably do for the sensor but would a true vector system work with this or do they all need a true encoder?

 

[DickDV]

Perfect application for an ABB ACS800 drive with DTC. No encoder but plenty of stiff speed regulation and torque down to zero speed.

 

[Skogsgurra]

There are better, but less well-known, devices. One was tested here: thread237-220699

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[BrianE22]

I don't know what type of motors they use but you might want to see what Fanuc offers for spindle motors on machine tools. There's loads of used ones (and drives for them) on Ebay.

 

[cswilson]

Unless you want to achieve true positioning capability with the spindle motor (as for hard tapping), I don't think you need a full vector drive with a shaft sensor. A "sensorless" vector drive should definitely provide you with the type of dynamic response you want.

Curt Wilson
Delta Tau Data Systems

 

[jraef]

Certainly a SVC drive will be a quantum leap better than what you already have, which as you point out is just less than optimal.

I would select a drive that has SVC but allows for an inexpensive upgrade to Full Vector by simply adding the encoder feedback from the motor. Many if not most are now that way.

Re:used drives. I am not a fan of this, you seldom can tell how a VFD was used and how long it has been out of service. Often times a VFD is sold off because the machine it was used on was a failed application, and that may mean the drive was abused before they gave up.

But if you do it, be careful. If left sitting on a shelf for a year or more, the capacitors usually need to be "reformed" in order to avoid damaging them when first energized.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[ozmosis]

It really does depend on the application required for the machine tool. If it is high speed machining and the motor is specifically designed for high speeds/high frequencies, then you have to be a little careful on drive setup and selection. Make sure you are operating in the design range of the spindle or the cost of a DIY solution could involve rewinding a spindle motor if it burns out. Cooling of spindle motors is usually an optimum design and running outside these limits will typically overheat the motor.

 

[jraef]

Patrick, remember he said it has been running on a V/Hz inverter for years already. Too late to worry about that now...

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[jimmy3456]

In talking with some mfgrs of drive systems, all apps engineers are recommending true encoder feedback direct from the motor. The actual acceleration of a load might be better with an encoder simple because of the instant full torque at 0 hz in which a sensorless would not develop full torque until 1-3hz or 30-90 rpm on a 4 pole motor.

I still am unable to quantify the actual difference in an application such as this. I am not too interested to spend an extra grand just to shave off .25s of spindle accel but might consider if we are talking 2 seconds.

I am also curious about decel rate. Does the encoder system work the same way to rapidly watch the decel rate to inject power at the right time and actually improve decel, or is this just a simple situation of adding resistor capacity to slow down faster?

It seems as though the actual at-speed load response time might be pretty close in which there might be little difference here but again, I am still trying to compare the two.

 

[sreid]

The one thing you can do with Full Vector Control is Rigid Tapping.

 

[jimmy3456]

The machine will do rigid tapping regardless because it already has a 1000 line encoder on the spindle for feedback to the control. It basically commands a spindle speed, then follows along precisely with axis motion to match. Works pretty well even with the old drive.

 

[jimmy3456]

Guys, after some research and links in this thread, it looks like the main difference in sensorless and vector other than actual speed regulation which is not as important for my application, is the low rpm torque. It looks like VERY few drives can really achieve full torque at near 0 speed. I think ABB said on the ACS800 that it needed 1-2hz which will be something to think about.

I do not run a machine that low very often but with a large saw, I have run in the 50-100rpm range. I also have to think about tap reversal.

So, are there any other drives than the one I saw here (Optidrive) than has proven to provide really good response at low rpm?

 

[jraef]

You don't mention a gearbox, so assuming you want 50RPM directly out of a 4 pole motor, that's a 30 or 36:1 speed reduction. I don't know of any reputable Sensorless Vector drive that can't handle that.
My list (in order of people I used to work for or with for lack of a more random plan)
Siemens
ABB
Teco
Bardac (US version of the Optidrive)
Schneider / Toshiba
Yaskawa
Allen Bradley

My preferences in terms of ease of use:
ABB
Yaskawa
Bardac
Teco



"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[jimmy3456]

The machine does have does have a gear box. I just recently worked some numbers to figure out what is going on.

hi range is .8:1 over driven so max speed at motor is 202hz, and max spindle is 7500.

Low range is 3.36:1 so 202hz is 1800 max at the spindle.

That puts us at 100rpm at the motor for our lowest application at 30rpm at the spindle. That pulls us back to 18:1. What do you think?

I had a drive retailer end up calling ABB to talk with a specific engineer for machine tools and he was not sweating the ACS800 sensorless system for this. They advertise near full torque at 0 rpm on the drive though another engineer said 1hz. The response time to load condition was 2-3ms and I am not sure if this is in line with an encoder system.

I really wish I could find a comparison of drives to verify this level of performance but the linky above was testing the ACS350 and it was unable to do much at 0 rpm, even with an encoder but the optidrive pulled the load great. I would just like confirmation of this. It would certainly save some time and hassle if we can go sensorless but I respect JRaef's opinion here due to the vast experience in drives.

Again, if I can achieve 99% of the performance with sensorless, I will not chase vector, but if that gap grows to say 90&%, I probably would.

 

[DickDV]

I routinely use the ACS800 on lab jobs requiring full motor torque at zero speed. The only thing that pushes me into an encoder is if, at zero speed, the shaft must be motionless. With DTC (sensorless vector), there is a small amount of twitching back and forth at zero. To get rid of that, encoder is needed. Above about 5rpm, the rotation is as smooth as can be.

The ACS350 is not going to give you zero speed performance but you didn't ask for that in your original posting. At 30rpm, ABB, Siemens, Mitsi, and Yaskawa will all give you good torque and speed regulation using sensorless vector control.

 

[jimmy3456]

Do you think due to the lack of full torque on some other drives, we would see a reduction in spindle acceleration? We commonly accel from 0-7500 50x/hour.

It is kind of looking like sensorless might work. I also have to consider tapping performance where a tap will dive into a work piece and reverse. Breaking the tap free for reversal is my only concern. Right now, the machine really struggles with 1/2x13 taps.

 

[DickDV]

With a good sensorless vector system, I would expect an increase in acceleration rate due to the increased torque available. If you really want to maximize the torque capabilities, size the drive so the short-term 1 minute current rating is twice the motor FLA. That will give you about all the torque the motor is capable of. To get more, you would have to change the motor.

 

[ScottyUK]

Hi jraef,

I'm surprised Control Techniques didn't make your list. I haven't had any dealings with them since they were assimilated by Emerson but as an independent company they were right up there in terms of performance and reliability. The Unidrive is at least worth a look.

Their technical folks at Telford were good too, although I don't know if that survived being Emersonised either. I hope it did.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jpolson]

Is upgrading or changing the motor an option? I agree with DickDV that a larger AC motor may really help.

I know it is a dirty word in a lot of circles, but have you looked at DC motor/drive combo's? We used DC for years on belt feeders that run slow and require a lot of torque.

Joel Olson

 

[jraef]

ScottyUK,
Nobody around me is selling or supporting the Control Techniques drives as far as I can tell. They used to have a local integrator (wholly owned), but they shut them down about 6 years ago. Nobody picked up the ball as far as I can tell.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376