So who is running what inverters?

[peter.blais]

Thinking about upgrading my machine to an inverter, mostly to get rid of the rotary converter (noisey). I was thinking just the odvious hitachi or whatever SJ100 thats recommended on the ajax site. Anybody have any experiance / pointers / whatever?

[Michael M]

Make sure your motor will work with the inverter first. I bought the 5HP SJ1100 for my Centroid conversion on my 325 Tree Journeyman, and it is really struggling with the 3 hp (sustained/5 hp for 15 minutes) 8000 RPM Yaskawa spindle motor.

If you've got a fairly conventional motor then the standard VFDs are probably good enough. But I've had to stretch the accel/decel times on the SJ1100 up into the 30 second range to stop it from faulting when going to or from full speed. I'm not sure it's making much power at either end of the speed range.

I'm going to be trying a borrowed 7.5hp VFD (and I do have three-phase into the VFD - it isn't doing any conversion) to see what that does, but I may have to go with a VFD that is designed to be a spindle drive, instead of a "universal" type of VFD. That won't be at all inexpensive if I have to do that.

cheers,
Michael

[Al_The_Man]

Originally Posted by Michael M Make sure your motor will work with the inverter first. I bought the 5HP SJ1100 for my Centroid conversion on my 325 Tree Journeyman, and it is really struggling with the 3 hp (sustained/5 hp for 15 minutes) 8000 RPM Yaskawa spindle motor.

It pays to buy the Flux vector type for speed control and the higher end ones like Mitsubishi have a self tuning feature and will accomodate standard 3ph motors as well as vector/spindle rated types, WEG also has an ecoonomic line that will do the same.
Al.

[Michael M]

Al, the Hitachi is a sensorless vector - I don't can't find anything about "flux" in the documentation and I'm not any kind of an electronics tech so I really don't understand many of the arcane details.

It does have two methods of self-tuning for the sensorless vector mode, but one of them faults when it tries to run the tuning sequence, and the other one says it was successful but the VFD faults when you try to run it. So s-v mode isn't useable, or at least not with any combination of parameters I can come up with. It works, but not very well.

The encoder is still in the motor and since I saw it run with the OEM Yaskawa spindle drive before it was taken off I don't see why the encoder shouldn't still be functioning. It appears that having sensor feedback is the best way to go, and at this point I'm not holding out much hope for this bigger standard VFD resolving the issues though I'll certainly hook it up and give it a try.

It seems that the Hitachi works fine with fairly "run of the mill" 3 phase motors, but it sure doesn't play well with this high speed Yaskawa motor.

cheers,
Michael

[Al_The_Man]

One main problem seen with some VFD's is overvoltage on decel, this is due to inadequate braking circuit, most have external braking units that can be addded, and is often necessary if fast braking is required.
Al.

[Michael M]

I'll probably end up adding a braking resistor. But I had to slow the accel ramp down a lot too, as otherwise it would give an over-current fault.

Perhaps the 3hp spindle motor has delusions of grandeur and thinks it is really a 7.5HP motor because it sure wants all the power the 5HP Hitachi VFD can supply, and then some.

cheers,
Michael

[HillBilly]

Originally Posted by Michael M The encoder is still in the motor and since I saw it run with the OEM Yaskawa spindle drive before it was taken off I don't see why the encoder shouldn't still be functioning. It appears that having sensor feedback is the best way to go, and at this point I'm not holding out much hope for this bigger standard VFD resolving the issues though I'll certainly hook it up and give it a try. It seems that the Hitachi works fine with fairly "run of the mill" 3 phase motors, but it sure doesn't play well with this high speed Yaskawa motor. cheers, Michael

Is there any name plate data on the motor. Its sounds more like a servo than a standard three phase induction motor.

Darek

[Michael M]

It isn't a servo, it is a spindle motor, and everyone so far seems pretty much in agreement on that. It runs on the Hitachi SJ1100, it just doesn't get the power it needs at high and low RPM. It runs to 8000 RPM which seems to be a lot faster than the motors that VFDs are generally used to drive. I can set the hz on the VFD and get it to run up, but the encoder feedback is looking like something that was definitely on that motor for a purpose.

cheers,
Michael

[Al_The_Man]

CNC spindles predominatly have a pulse generator for feedback rather than use flux vector control, this ensure tighter control, alot depends on the quality and type of the VFD when used without the PG feedback.
Al.

[Shizzlemah]

It would be folly to use a spindle drive that does not have encoder feedback. You will get an appreciable increase in available torque.

Flux or Sensorless vector is typically fine for gear reduction spindle drives, like on an old engine lathe or bridgeport. Places that have lots of reduction from the motor....

I have an SJ100. It stinks for CNC use. 5HP is grossly underpowered for a 5HP CNC lathe. Cannot accel fast enough for any usable CSS. Wont decel fast enough when retracting (like pecking a groove) and will go overvoltage and fault out.

[Al_The_Man]

I have had mixed results with sensor and sensorless VFD's, The ones I found good were Schneider Altivar series and Mitsubishi.
I have a couple of older style Mitsubishi VFD's on a couple of mills with original Pancake style induction motors, the flux vector sensorless will show excellent torque down to low rpm, and will accelerate faster than on straight 60hz and will stop on a dime without faulting, the encoder feedback version would O.V. on decel, so I had to put an external braking unit, now they match. They both use 120hz tops.
I believe the later ones do not need the external braking.
This is a site showing a few manuf. opinions on sensorless.
http://www.manufacturing.net/ctl/article/CA182473.html
Al.