VFD speed ranges with an induction motor?

UKworkshop.co.uk

Help Support UKworkshop.co.uk:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

sploo

Somewhat extinguished member
Joined
8 Nov 2014
Messages
3,981
Reaction score
1,327
Location
West Yorkshire
I've recently picked up a metalworking lathe that's fitted with a 3 phase, 2 pole (2800 rpm at 50Hz) induction motor, and I'd like to achieve something around a 9:1 speed ratio with a VFD; with the highest motor speed translating to the spindle running at 3000rpm with an appropriate pulley ratio.

From what I've been able to discover, a 2 pole motor can generally be run at up to 125% of base frequency (i.e. 62.5Hz on a 50Hz supply). Indeed the motor I have is actually rated to run at 2800rpm at 50Hz or 3400rpm at 60Hz, so that should be fine.

I understand the minimum allowed speed is driven by the motor's Turndown Ratio (e.g. 10:1 meaning it can be run at 1/10th of base speed). I don't have any data for this motor, but independently I've seen warnings that a Totally Enclosed, Fan-Cooled (TEFC) motor can have cooling problems if run too slowly (as the cooling fan is on the motor spindle). I've read that a "standard" motor is probably good for 4:1, and "inverter rated" motors good for 10:1.

If I assume 4:1, then that's a minimum frequency of 50/4=12.5Hz. 12.5Hz to 62.5Hz gives me a 5:1 overall ratio; so still some way short of the target 9:1.

I'm told that 4 pole motors (approx 1400rpm) can be run up to 200% of base frequency. Seems high to me; though if the moving parts of the motor are no worse in terms of balance than a 2 pole motor then ~3000rpm should be mechanically safe (I don't know about the electrical side).

So - what's the best option:
  1. Existing 2 pole motor run between 12.5 and 62.5Hz for a 5:1 ratio (25 to 125% of base frequency)
  2. Existing 2 pole motor run between 7 and 62.5Hz for a 9:1 ratio, but with external cooling (14 to 125% of base frequency)
  3. New 4 pole motor (might as well be inverter rated if I have to buy a new one) run between 6 and 50Hz for a ~9:1 ratio (12 to 100% of base frequency)
  4. New 4 pole motor run between 11 and 100Hz for a ~9:1 ratio (22 to 200% of base frequency)
I understand that a motor run at less than base frequency is constant torque (and therefore runs at less than the max rated power), and that a motor run over base frequency (without an increase in voltage) will be constant power (so decreasing torque as rpm increases). I'm just not sure which is better for a lathe.

Gut feel tells me that constant power might be better - as higher spindle rpms are usually used for smaller work (and tend to be less torque demanding)?
 
Think I've worked out the route (based on finding some information on the motor I have); but I'll add to my waffling above in the hope it might be useful for someone. The motor is a 2.2kW Electrodrives DA90LA, and if I read the chart I found (page E19 in http://www.farnell.com/datasheets/98858.pdf) correctly, it's rated to provide constant torque between 16.7-50Hz with a nominal power of 2.1kW (i.e. basically the max available torque down to 16.7Hz).

Including the above, the nominal power figures whilst providing constant torque for the following frequency ranges are:

16.7-50Hz -> 2.1kW
10-50Hz -> 1.8kW
5-50Hz -> 1.6kW
2.5-50Hz -> 1.4kW

I assume the above to mean that the max torque drops below 16.7Hz and steadily falls.

The spec lists the motor as having a max torque of 7.5Nm, and given Power (W) = torque (Nm) x rpm / 9.5488, that means...

2200W = 7.5Nm x 2800rpm / 9.5488

...which looks to make sense.

I assume this can be used to calculate the max torque at lower frequencies. I.e. if nominal power at max rpm is only 1.6kW if a constant level of torque is required from 5Hz, and Torque (Nm) = Power (W) / (rpm / 9.5488), then I assume it means that max torque at 5Hz would be:

5.5Nm = 1600W / (2800rpm / 9.5488)

As the minimum listed frequency is 2.5Hz the motor should have a Turndown Ratio of 20:1 (2.5Hz - 50Hz); which from other articles I've read seems realistic for a decent inverter rated motor.

The table also states the motor is constant power (2.2kW) from 50-100Hz. If I understand that correctly it means the motor is rated to be able to run at 100Hz (~5600rpm); with the torque falling above 50Hz, down to half its 7.5Nm max at twice the base frequency.

Putting all that together (with a bit of interpolation between the calculated torque figures at 2.5, 5, 10 and 16.7Hz), I get a torque vs frequency graph from 2.5Hz to 100Hz of:

motor torque.PNG



As my desired max spindle speed is ~3000rpm, and at the 50Hz base frequency the motor is running at 2800Hz, I'd have (roughly) a 1:1 ratio between the motor and the spindle. This would see the motor run between ~6 and 50Hz.

If I instead run the motor up to 100Hz then reduce the motor-to-spindle ratio by 2:1 I'd have the same rpm and torque at max speed (at 100Hz the motor is at 2x the speed, but 1/2x the torque; so a 2:1 reduction gives 1x the speed and 1x the torque). However, at lower frequencies (the motor would now run between ~11 and 100Hz) the 2:1 motor-to-spindle ratio would give torque benefits.

The following graph shows the spindle torque (red) between 6-50Hz with 1:1 ratio to the spindle (i.e. same as the motor torque). The green line shows the 11-100Hz range with a 2:1 reduction ratio between the motor and spindle (hence twice the torque across equivalent frequencies).

spindle torque.PNG


That's a very long post. Hopefully anyone who's made it to the end has either found it useful; or will tell me where I've made some really dumb errors.

.
 
I skipped through all your calcs.
You are right about the pros and cons you posted up front.
Slowing a motor down significantly using a VFD is problematic because at slow speed you often want the added torque provided by a gearbox or belt drive ratios. There is a real risk of stalling.
At high speed, you are often working on a slim part or taking a light finishing cut. Loss of torque is less important here.
As a 4 pole motor is physically almost identical to a 2 pole, it's mechanically safe to overspeed it to 2800 odd rpm.
At 100Hz, there may be somewhat greater losses and heating than at 50Hz. So electrically it's not entirely a given that a 4 pole will be OK at 100Hz, but will probably be OK. Check mfr data if you can.
Overall, I think it's much better to speed up a 4 pole than slow down a 2 pole in lathe applications.
I use an ABB 2.2kW 4 pole motor to spin my vintage Harrison L5 metalworking lathe. I have tested it at 100Hz without issue but finally programmed it to go to 75Hz or +50% overspeed.
This gives me valuable extra speed needed for clean cuts on slim parts but is within the range that the lathe is designed for. It's a 1949, older than I am, so I ought to be kind to it.
I don't dial it down low. I use the gears.
For cutting, 10:1 range feels much too big to me. I think better to settle for a much smaller range like 4 or 5:1 it will still save you some belt changes and you will quickly be able to dial away from any resonances.
Though on a wood lathe, if you just want 120rpm to rotate a bowl in front of a hairdryer, that won't do any harm.
If you are going to be loading it at reduced speed, I suggest getting a higher power motor and vfd than the original fixed speed one. You'll be limited by what frame size fits, but may have the chance to step up a bit.
Last observation, I use both so can compare older generation VFD's with the newer ones using vector control. If you are spinning a motor really slowly, you can feel that the vector field control is smoother. But this is so slow you would never be cutting at those speeds.
Once you speed up a bit, you can't tell the difference.
 
Last edited:
First thing I would consider is whether the lathe itself is designed to run at 3000rpm, and why you want such a high top speed. If you brought that requirement down to say 2000 it would make life easier, and still be plenty fast enough for most work. How is the machine driven at present, does it have a gearbox or is it just a variable speed motor. I also run a Harrison, a 1961 5a. I have a 3hp 2800 rpm motor, and have just configured the vfd to give two pre set speeds, 1400 and 2800. With the machines gearbox this gives me 16 gears and spindle speeds between 30 and 1500 rpm. The spindle is actually rated for 2000 rpm, but I don't need that fast a speed.
 
I’ve been using inverters for years. My default setup is from 10 -100 hz and if good torque is needed at low speeds the use a vector mode inverter. Virtually all inverters will measure stator resistance and use that to estimate temperature rise in the motor and shut it down if it deems it to be too hot. If this keeps happening fit a fan on the end of the motor.
Before buying the inverter download the manual and check it has the features you need.
 
Gents - thanks for the thoughts.

The lathe is a Colchester Chipmaster 5x20; rated up to 3000rpm. Originally they used a mechanical CVT gearbox (a Kopp Variator) to achieve a 9:1 speed ratio with a single speed motor. In theory at 9:1 mechanical reduction you'd get a 9:1 torque amplification (unlike reducing the motor speed with a VFD); however, what info I've been able to find hints that you might get some slipping in the variator, so it wouldn't be quite that good.

The lathe also has two distinct gear ranges after the variator; giving 35-300rpm and 350-3000rpm (that latter gearing is still present on the lathe I've bought).

It's quite common for the variators to lunch themselves, and get replaced with a VFD; which I assume is what's happened to this lathe.

From what I've been able to work out, Electrodrives (the maker of the motor) became part of Brook Hansen; later Brook Crompton. I've just spoken to a chap there and whilst he wasn't able to be 100% certain he agreed that (based on the catalogue I linked above) the motor should be good for 2.5Hz to 100Hz (i.e. 5% to 200% of the 50Hz base frequency). He did that say these days they'd advise a safe top speed of 5200rpm (this 2 pole 2800rpm motor would be running at 5600rpm at 100Hz).

My little mini lathe has a top speed of 2500rpm; which I rarely use, so Fergie's right in that 3000rpm is probably unnecessary. I suppose I could gear the motor to provide 350-3000rpm in the high ratio (100Hz at the top end); knowing that I'll rarely have the spindle up that high. However, in the low ratio, the motor would have to be running at top speed to achieve 300rpm at the spindle; which may be unwise? In any event; I assume that with the 10:1 reduction in the low ratio, a 2.2kW motor will be giving plenty of spindle torque for the slow/large diameter cutting jobs?

Myfordman - I have a spare Huanyang VFD (I have several on different machines). That's obviously a cheap model, so I assume it's not vector mode (maybe scalar?). Does that mean it's unlikely to work well down at the 10Hz region?
 
Last edited:
Interesting, especially top RPM. I often wish I could rotate stuff faster than 3000RPM with carbide inserts as the SFM is too low.
 
Myfordman - I have a spare Huanyang VFD (I have several on different machines). That's obviously a cheap model, so I assume it's not vector mode (maybe scalar?). Does that mean it's unlikely to work well down at the 10Hz region?

I suspect you are correct. I cant see anything in the HY manual to suggest it has vector mode control. key words to look for are Vector or Sensorless rather than the basic V/F control.
I use V/F on my engineering lathe and mill. Keeping the belts in mid range and dropping to 5hz is great for tapping under power. I can stall the machine by hand and gently let it slip to control the thread cutting rate. When I need good torque, I drop the speed on the belts and run the motor at 30hz or more..
On the bandsaw a vector mode inverter plus a single belt change can get me 100 FPM for steel and 3000 FPM for wood.
Vector mode inverters are not necessarily always the dearest ones but the very cheapest ones are not likely to support vector control or stator resistance measurement for temperature estimation
 
So you bought the chippie then. Good man, lovely machine. I know it's not really what you want to do but can't help wondering if it might not be an idea to just fit a pair of multiple pulleys to give some change of ratios, and keep your motor in a narrower rev range. I suppose it all depends on how often you would need to move the belt. A mate of mine was going to look into replacing the variator on his with the cvt components from a twist and go scooter, don't think he ever got around to actually doing it, but an intriguing idea. The main reason they normally fail is because they haven't been run on the correct, and very expensive oil.
 
So you bought the chippie then. Good man, lovely machine.

Yea, I did go for it in the end. Fortunately despite the one I was after being cosmetically rough it turns out it was mostly just covered in a layers of grime. The paint has definitely seen better days, but so far I've found no nasty mechanical surprises (at least, nothing from visual inspection and moving parts around).

I know it's not really what you want to do but can't help wondering if it might not be an idea to just fit a pair of multiple pulleys to give some change of ratios, and keep your motor in a narrower rev range. I suppose it all depends on how often you would need to move the belt.

I've spent a bit more time investigating the various ratios of the pulleys and timing belts, and to achieve 3000rpm at the spindle requires spinning the clutch shaft at approx 2300rpm (there's a 35T to 27T ratio with a toothed between between the clutch and spindle).

The clutch shaft has an SPA V pulley with a 112mm PCD, and the motor has the smallest available 63mm PCD pulley. If the clutch has to spin at a max of 2300rpm, then the motor needs to spin at 2300 * 112/63 ~= 4100rpm.

It's not quite the 2:1 ratio I was considering, but it's within what I understand the motor can do (4100rpm is 73.5Hz, or 147% of the base frequency).

The lowest speed required is actually 8.57x slower rather than 9 (3000/350rpm), so gives a low frequency of 8.6Hz, or 17% of base; which is again well within what the motor will do.

If that turns out to give me insufficient torque in the low rpm region of the high range (i.e. >=350rpm) then I might have to consider some sort of two stage pulley system. I assume that once switched into low ratio (35-300rpm) there will be a 10:1 torque amplification at the spindle, and that it will be "enough".

A mate of mine was going to look into replacing the variator on his with the cvt components from a twist and go scooter, don't think he ever got around to actually doing it, but an intriguing idea. The main reason they normally fail is because they haven't been run on the correct, and very expensive oil.
I've seen such a CVT fitted (DIY) onto a milling machine. A fascinating thing; with pulleys that are split into pairs of conical discs; the opening and closing of the conical discs changing the effective pitch diameter of the pulleys.

Judging by what I found in the motor/variator cabinet on my Chippie, I suspect the expensive oil from the variator had been sprayed in a catastrophic but decorative arc over the inside faces of the lathe :oops:
 
The bike ones work by the halves on the wheel being held together by a powerful spring, so giving the largest diameter of the pulley at rest. On the engine end you have a fixed half pulley attached to the crank, and a moveable one controlled by rollers than run up ramps inside a fixed case. At rest the plates are apart, so giving a low ratio. As the speed of the engine increases the rollers are thrown out by centrifugal force and roll up the ramps, forcing the plates at that end together and increasing the effective size of the pulley. The wheel end plates are then forced apart against the spring by the belt so increasing the effective ratio. The system can be tuned by varying the weight of the rollers, lighter rollers will give better acceleration, but a lower top speed. Heavier rollers will give more sluggish acceleration, but ultimately a higher top speed. Very similar in principle to the variomatic transmission on Daf cars, if you remember them. Can't see any reason why it shouldn't work on a lathe. If I remember his idea was to control the separation of the driven pulleys by fitting a threaded collar with a thrust bearing acting on the movable pulley at the driven end. A small worm gearbox would enable you to screw the collar in and out via a handwheel to set and maintain the speed at your chosen figure.
 

Latest posts

Back
Top