Single Phase To Three Phase Rewind - Lathe Uprate

[Ken I]

I am going to rewind my single phase lathe motor from 220 single phase 50Hz. To 220V three phase 125Hz and run it off a single phase VFD (variable frequency drive) and in the process increase the motor's output nearly three times. I will cover this in a sequence of posts.

Those of you with only single phase supplies might find this interesting in that you can rewind to three phase and enjoy the benefits of using a VFD.

You don't have to change your motor.

Ken’s Lathe Uprate

The Lathe :-

A BV20BL Ø220mm (over bed) x 520mm Chinese Lathe supplied under a local brand name. I have no idea who the Chinese manufacturer is or what other brand names it goes under – if you recognize it under other brand names or suppliers please let me know.

Driven by a stupidly undersized 550W (3/4 H.P.) motor via a “V” belt drive.

Primary Drive layout & Motor – which is a dual capacitor start / run system.

The Problem(s)
As with all “cheap” Chinese machinery all the usual suspects.
No manual, no circuit diagrams, no spare parts list or diagrams – and virtually no service or spares support – if it breaks you fix it yourself.
The electrics on Chinese machines are also typically poor or not up to code – but apart from the stupidly small motor with appalling starting torque, it has served me well for a few years but I have become thoroughly fed up with its’ underpowered quirkiness.

The machine is so underpowered that it will not even start at its two highest speeds (1950 & 1345rpm rpm) and has to be warmed up by running it for about 20 minutes at the next lowest speed 736rpm – this also suggests that the gearbox is filled with some sort of oil that gets very viscous when cold – in winter it can struggle to start at the 736rpm speed.

I also dislike the stepwise nature and limited selection of speeds offered by the gearbox and the high isn’t high enough and the low isn’t low enough – my previous smaller lathe had two gears “high” and “low” and a variable DC drive which I liked.

Suspecting there might be an electrical problem (like the centrifugal start switch or capacitor not working) I removed the motor and switch gear to check that it was all working properly.

There are two capacitors – an 18 UF “run” capacitor and a 100 UF “start” capacitor – which gives the start winding an extra boost until the centrifugal switch cuts it out.

No problems – everything was working and connected as it should.

Since the nameplate states 1400rpm @ 50hz – then this is a 4 pole motor – note the start switch and capacitors are hard wired internally – would have been better brought out to the terminal block. Presumably because this allows a 4 terminal block instead of 6 – probably just to be cheaper ?

Synchronous speed at 50 Hz. Would be 1500 rpm for a 4 pole motor so 1400 seems a little slow – 1450-1470 would be more normal – if the 1400 is correct it represents an inordinately large “slip” figure.
A large slip figure can assist starting torque – which certainly is not the case here – nor can I see any obvious indicators like large rotor clearance or small squirrel cage bars – curious ?

I suspect the abysmal starting torque is more a symptom of using too small a motor rather than any fault of the motor itself.

Whilst the motor was apart, I counted the winding slots = 24 which is the sort of number you would expect because if you are a motor manufacturer you would want to use the same core laminations for two, four and six pole motors as well as single phase and three phase so you want a common denominator that will divide by 2,3,4 & 6. So you normally see numbers like 12, 24, 36 …..

This is important as I want to rewind this to a 3 phase 4 pole motor and run it off a VFD (variable frequency drive).

Since it was a mission to pull the motor and I’ve always wanted to uprate it I am now going to do so and “Hot it up” by spinning the rotor faster with the VFD by using a higher frequency.

The Modifications.

These are the changes I intend making :-

• Change the glue used by the manufacturers’ as lubricant in the headstock.
• Change the drive belt to a Synchroflex toothed belt drive.
• Drop the pulley ratios to improve toque delivered to the chuck.
• Rewind the motor to 3 phase + add VFD drive.
• Run the motor at max. 125 Hz – thus more than doubling its’ power.
• Add potentiometer to adjust motor speed from low to high.
• Add a tachometer to display actual spindle rpm.

Getting the single phase motor to run on a VFD requires rewinding the motor to three phase – and for a lower 50 Hz. rated voltage so I can still reach full torque at 125 Hz. at 220 volt (that’s the maximum output of the VFD I have to hand) and ±3750rpm (depending on slip) which at the same torque equates to ±1.4kW output - et voila.

See my attached article on getting more power out of a squirrel cage motor :-

Doubling the power output is accomplished by rotating the motor twice as fast (2800-3000rpm) whilst still outputting its usual torque – so you are not “overloading” the motor torsionally and 3000rpm is the speed this motor would run if it was a two pole – so I’m not doing anything “over the top” here.

I have done this numerous times to get way more power out - as much as 8 times – sounds crazy but eminently doable – see the article. I have done this whenever I needed much more power-to-weight ratio from an ordinary squirrel cage motor – in my line of work – robotics – where I need to keep down the weight of any motor driven tool on the end of a robot arm.

I am going to reduce the primary belt ratio from 1.74:1 to 2.14:1 which will give me 23% more torque across the board and when gearing down on the lower ratios will give me up to 125% more torque than the lathes (currently pathetic) output. To that add the increase in torque available from changing to three phase (see later) probably brings it up to 150%.

Yes I have examined the gearbox and I’m sure it can handle it with ease. I think the penny pinching manufacturers just used a stupidly underpowered motor.

Next: Calculations for rewinding

Regards, Ken

 

[Ken I]

Before going into the calculations - I would like to point out some technical issues.

This is all in the attachment to the previous post - so if you read that or aren't interested in the technicalities - you can skip this post.

Single Phase vs 3 Phase Motors.

Firstly – a single phase motor is a non-self starting push-pull system which does not generate equal torque over a 360° (electrical cycle) – to get it to start, requires the addition of a “start” winding via a capacitor which shifts the electrical phase angle of the start winding by 90° so we now have two push-pull systems at 90° (or thereabouts) which is much better – starting torque is still poor so in some cases an additional booster capacitor is added for starting and is switched out by a centrifugal switch once up to ±70-80% of rated speed. Note some single phase motors switch out the “start” winding completely – leaving you with a very inefficient push-pull system. (However this can still be built to deliver the required power – but up to twice as “coggy”).

A three phase motor generates 3 sets of push-pull motions at 120° to each other so delivers much smother and consistent torque and is obviously self starting. A three phase motor in fact generates a perfectly rotating magnetic field (long boring explanation omitted).

The top left graph shows the torque delivery from a single phase motor (after starting and switching out the start winding) – which typically is about 70% of what could be achieved if the available torque from the iron components was delivered smoothly all the time. (≈ the shaded area between the curve and zero – adjusted for the RMS figure quoted in the header.)

The top right shows a single phase – during startup - with the start winding (red curve) boosted by a larger capacitor to improve the starting torque.
For for this type of motor the windings are often the same for start and run (they don’t have to be – the start winding is more commonly a “lighter wind”).

During starting, the additional capacitor raises the “start” bumps in the graph from 50% to 95% (again, anything is possible).

The bottom left shows a single phase with the start winding (red curve) permanently running – typically the start capacitor matches the impedance of the windings thus delivering a field approximately half as powerful but shifted by 90° by the capacitor to induce a rotation to the field.

Many single phase motors are run this way without a centrifugal switch cut out for the start boost capacitor (or even the entire start circuit).

Some very old single phase motors use only induction – by using a “shaded pole” – this is a start winding shorted out by the centrifugal start switch – which induces an out of phase current in the start winding to get the motor turning – this is normally switched off after starting. I haven’t seen one of these in over 30 years except on small low powered fan motors and such but for some curious reason, Chinese manufacturers appear to be making large shaded pole motors.

For illustrative purposes, I have drawn the graphics “idealized” in practice these waveforms would have all sorts of “noise” in them.

I have also chosen to have the start winding at 50% power and shifted by 90° - it’s always somewhat less than 90° in practice and it doesn’t have to be 50% - the manufacture can do an uneven wind and capacitor match-up that will equal the run winding – makes it more expensive and I have seen it done which can get you to about 90% delivered torque.

However these days it is more common to see what I have described above or the start winding is more usually the “lighter” winding in the case of capacitor run (without a centrifugal switch cut out).

When I got to strip my motor I found that the start wind was in fact a lighter wind.

So a “capacitor run” single phase motor gives an improvement from 70% to 83% torque delivery, a ≈18% improvement which is also smoother with less noise (Hum) over a single phase run only wind.

The lower right diagram of a three phase motor shows that the torque delivery averages about 95% of the available.

Thus a three phase motor can deliver ±15% (depending on the original single phase configuration) and up to 36% more torque than a single phase run only wind – from the same frame size / ironmongery. Whilst simultaneously being smoother (less “coggy”) and with less noise (Hum).

So simply rewinding the motor to three phase will produce a significant improvement in torque and power – as much as 36% but more likely 15%.

Motors are normally designed / wound to handle ±10% overvolt threshold (to allow for fluctuation in the mains supply) – so you can push your output that much further when running off a VFD as the VFD’s output will not change upward because of upward variations in the mains – thus allowing you to sail closer to saturation than you would normally dare.

So I could rewind this for ±200V and get another 10% out of it at 220V – so I could probably get 25% to 30% more out of the motor just by rewinding and switching to a VFD drive.

I could stop there but I also want higher rpms – so I might as well go the whole hog.

At this point you are going to ask “why not just buy a bigger motor ?” good question and good advice – if it’s easier to do – in this case it is going to be very difficult to adapt to a larger frame motor and a rewind is going to be less costly than a new motor (or shoehorning in an available cheapo.) The VFD was going to cost in any case – so – in for a penny – in for a pound.

Running A Motor At A Higher Frequency.

The problem with running a motor faster off a VFD is the increasing frequency increases the impedance of the motor and torque starts to drop off as the speed increases and therefore power flatlines – as per diagram below for a 380V motor / VFD :-

In the diagram above, the voltage (red line) rises more or less linearly with frequency up to the normal rated speed and voltage. The torque (green line) remained more or less constant as the voltage was rising as the frequency induced impedance rose thereby maintaining current and therefore torque. But at the motors normal rating your VFD runs out of volts and the impedance keeps on increasing – so revs go up and torque goes down and power (magenta line) flatlines.

The blue line is speed which is a linear function of frequency.

In the above diagram “slip” is ignored and linearity is assumed for simplicity.

To get around the problem of “torque fade” (in order to get more power) you rewind the motor to achieve the desired current (torque) at the desired top end frequency (speed) – The resultant “wind” is a motor that runs on the natural frequency (50 Hz.) at a much lower voltage.

I'll get into that in the next post.

Regards, Ken

 

[retailer]

It certainly is do-able, I did a similar thing a few years back when I wanted to run my mill from a VFD, I had some basic info on the winding layout but not enough for a noob such as myself to confidently tackle the job, so I joined the Eng-Tips forum. The forum has a some very knowledgeable motor winders - while none of them actually provided a winding layout they guided me through the winding layout design process, better for me to understand the layout/coil connections etc than to just follow a set of instructions. My rewind may possibly have been a bit more complicated as I rewound it to be 2 speed, probably not really needed with a VFD - I do at times use the second speed just to keep the motor closer to it's designed operating frequency of 50Hz - it was a success here is the link to a post I made on another forum about the rewind.

https://metalworkforums.com/f65/t200978-arboga-mill-motor-rewind
Good luck with the rewind I'm sure it will go well.

 

[mu38&Bg#]

This will be interesting. I thought there was often a fundamental difference in the laminations of single phase motors preventing 1PH to 3PH rewind. I suspect those motors are just that inefficient. I have one from a combilathe doing work on my band saw. Seems gutless for a 3/4HP. I have a much larger ABB waiting in queue.

Winding voltage (V/Hz) of induction motors is something few understand, and I originally learned about from you here long ago. I did understand the concept in DC motors, but knew nothing about AC. My CNC lathe runs a 230V motor at 400V to 87Hz. I guess this is easy and common in Europe.

I had an ordeal with a Lenze 10kRPM motor from a mill I got last year. It needed bearings and of course I dropped the rotor. Local motor shops don't seem to understand balance grades and said it checked out and .001" runout on the shaft was OK, but it vibrated badly. I ended up balancing myself. It's not perfect, but these motors are hard to find and new rotors are not available. Lenze offers several windings for their motors to take advantage of high frequency operation.

 

[TSutrina]

I am going to rewind my single phase lathe motor from 220 single phase 50Hz. To 220V three phase 125Hz and run it off a single phase VFD (variable frequency drive) and in the process increase the motor's output nearly three times. I will cover this in a sequence of posts.

Those of you with only single phase supplies might find this interesting in that you can rewind to three phase and enjoy the benefits of using a VFD.

You don't have to change your motor.

Ken’s Lathe Uprate

The Lathe :-
View attachment 120035
A BV20BL Ø220mm (over bed) x 520mm Chinese Lathe supplied under a local brand name. I have no idea who the Chinese manufacturer is or what other brand names it goes under – if you recognize it under other brand names or suppliers please let me know.

Driven by a stupidly undersized 550W (3/4 H.P.) motor via a “V” belt drive.
View attachment 120036
Primary Drive layout & Motor – which is a dual capacitor start / run system.

The Problem(s)
As with all “cheap” Chinese machinery all the usual suspects.
No manual, no circuit diagrams, no spare parts list or diagrams – and virtually no service or spares support – if it breaks you fix it yourself.
The electrics on Chinese machines are also typically poor or not up to code – but apart from the stupidly small motor with appalling starting torque, it has served me well for a few years but I have become thoroughly fed up with its’ underpowered quirkiness.

The machine is so underpowered that it will not even start at its two highest speeds (1950 & 1345rpm rpm) and has to be warmed up by running it for about 20 minutes at the next lowest speed 736rpm – this also suggests that the gearbox is filled with some sort of oil that gets very viscous when cold – in winter it can struggle to start at the 736rpm speed.

I also dislike the stepwise nature and limited selection of speeds offered by the gearbox and the high isn’t high enough and the low isn’t low enough – my previous smaller lathe had two gears “high” and “low” and a variable DC drive which I liked.

Suspecting there might be an electrical problem (like the centrifugal start switch or capacitor not working) I removed the motor and switch gear to check that it was all working properly.

There are two capacitors – an 18 UF “run” capacitor and a 100 UF “start” capacitor – which gives the start winding an extra boost until the centrifugal switch cuts it out.
View attachment 120037
No problems – everything was working and connected as it should.

Since the nameplate states 1400rpm @ 50hz – then this is a 4 pole motor – note the start switch and capacitors are hard wired internally – would have been better brought out to the terminal block. Presumably because this allows a 4 terminal block instead of 6 – probably just to be cheaper ?

Synchronous speed at 50 Hz. Would be 1500 rpm for a 4 pole motor so 1400 seems a little slow – 1450-1470 would be more normal – if the 1400 is correct it represents an inordinately large “slip” figure.
A large slip figure can assist starting torque – which certainly is not the case here – nor can I see any obvious indicators like large rotor clearance or small squirrel cage bars – curious ?

I suspect the abysmal starting torque is more a symptom of using too small a motor rather than any fault of the motor itself.

Whilst the motor was apart, I counted the winding slots = 24 which is the sort of number you would expect because if you are a motor manufacturer you would want to use the same core laminations for two, four and six pole motors as well as single phase and three phase so you want a common denominator that will divide by 2,3,4 & 6. So you normally see numbers like 12, 24, 36 …..

This is important as I want to rewind this to a 3 phase 4 pole motor and run it off a VFD (variable frequency drive).

Since it was a mission to pull the motor and I’ve always wanted to uprate it I am now going to do so and “Hot it up” by spinning the rotor faster with the VFD by using a higher frequency.

The Modifications.

These are the changes I intend making :-

• Change the glue used by the manufacturers’ as lubricant in the headstock.
• Change the drive belt to a Synchroflex toothed belt drive.
• Drop the pulley ratios to improve toque delivered to the chuck.
• Rewind the motor to 3 phase + add VFD drive.
• Run the motor at max. 125 Hz – thus more than doubling its’ power.
• Add potentiometer to adjust motor speed from low to high.
• Add a tachometer to display actual spindle rpm.

Getting the single phase motor to run on a VFD requires rewinding the motor to three phase – and for a lower 50 Hz. rated voltage so I can still reach full torque at 125 Hz. at 220 volt (that’s the maximum output of the VFD I have to hand) and ±3750rpm (depending on slip) which at the same torque equates to ±1.4kW output - et voila.

See my attached article on getting more power out of a squirrel cage motor :-

Doubling the power output is accomplished by rotating the motor twice as fast (2800-3000rpm) whilst still outputting its usual torque – so you are not “overloading” the motor torsionally and 3000rpm is the speed this motor would run if it was a two pole – so I’m not doing anything “over the top” here.

I have done this numerous times to get way more power out - as much as 8 times – sounds crazy but eminently doable – see the article. I have done this whenever I needed much more power-to-weight ratio from an ordinary squirrel cage motor – in my line of work – robotics – where I need to keep down the weight of any motor driven tool on the end of a robot arm.

I am going to reduce the primary belt ratio from 1.74:1 to 2.14:1 which will give me 23% more torque across the board and when gearing down on the lower ratios will give me up to 125% more torque than the lathes (currently pathetic) output. To that add the increase in torque available from changing to three phase (see later) probably brings it up to 150%.

Yes I have examined the gearbox and I’m sure it can handle it with ease. I think the penny pinching manufacturers just used a stupidly underpowered motor.

Next: Calculations for rewinding

Regards, Ken

Great article. I worked at an aircraft generator system and motor manufacture. Further I packaged electronic speed drives so understand the whole system.
The typical VFD makes step changes to create 3 phase. The single phase input is converted to DC at a higher voltage above the peak for the output. Chopper and a high frequency transformer is typically used. This output is also converted to DC. Now the 3 phase VFD step width modulates, turn off and on connection to the DC buses to put in energy into the winding in typically maximum of 12 steps. Thus only a few steps to make a very ruff sine wave.
Why I am mentioning this is that the motor mass makes it seem that a smooth sine wave is sent to it. But the motor is producing significantly more heat smoothing out the electrical steps into smooth rotation. Eddy current losses in the iron and copper or aluminum increase significantly.
How to get a handle on the amount best approach is to find the motor horsepower rating of the same induction motor when driven by smooth 3 phase power and VFD driven. Note that in some industrial application the motor speed is used to replace valves to adjust flow but the pump being driven is the same. Thus the peak speed is 60 or 50 Hz and lowered to be more energy efficient then using a valve. So you can find the two different rating of a motor.
The losses are higher at all speeds. So if you use the VFD to run at lower speed but high power the motor is going to produce more heat, but the cooling will be less by a cube ration if I remember correct. Flow is a square relationship. I would still make the change but there is a cost.

 

[awake]

Ken, this is a fascinating project. I would not have thought it possible to re-wire a motor from single to three phase, or change the frequency, so I have much to learn from your posts (which I am still digesting).

The obvious question, of course, is why go the route of re-winding rather than simply swapping out the old motor for a three-phase motor. You mention cost and size as the issues. I could see a new 3-phase motor costing more than the materials for a rewind ... but what about a used motor? I would think a used 1 to 2 hp motor might be available pretty cheaply, and might still squeeze in to the size you need ... ??

Of course, there is also the fun factor - I have often done something that could have been done more cheaply by buying something ... but I wanted the fun of doing it myself. My philosophy is, why spend $5 on a part when you can spend 6 hours in the shop making it instead?

 

[Ken I]

Towards the end of my last post I said...

" At this point you are going to ask “why not just buy a bigger motor ?” good question and good advice – if it’s easier to do – in this case it is going to be very difficult to adapt to a larger frame motor and a rewind is going to be less costly than a new motor (or shoehorning in an available cheapo.) The VFD was going to cost in any case – so – in for a penny – in for a pound. "

In my case the rewind cost me ±U$40 - I got it back and it runs as planned - Will give you the info in the next post. Apart from the technicalities, the rewind is almost certainly the simplest approach.

That cr..ppy Chinese 3/4HP is now producing 2+HP without raising a sweat - it even runs cooler and more efficiently.

Regards,
Ken

 

[comstock-friend]

Thank God I'm retired from a company that built thousands and thousands of product with motors, and was able to bring home a few here and there that had a cosmetic ding or two. Life's too short to want to make something it's not into something you want, unless the effort floats your boat. Me, I'd rather be making chips...

 

[awake]

Ken, that is astonishing - I would have guessed a rewind would cost way more than that. Thanks again for introducing this provocative project!

 

[bluejets]

I would have guessed a rewind would cost way more than that.

In most places it does.

 

[Ghosty]

Ken, that is astonishing - I would have guessed a rewind would cost way more than that. Thanks again for introducing this provocative project!

It was cheaper to buy a new motor and VFD than get my old motor rewound, a little shoe horning and it all fitted in a AL250G

 

[Ken I]

Thanks for the comments and interest so far.

U$40 is cheap for a number of reasons - this is Cape Town, my rewinder does lots of commercial work for my business and I have an account with him, he is also quite used to my frequent strange requests.
Most rewinders see this sort of job as "trouble" and quote you a price they want you to refuse.

If rewinds are more costly than a new motor - why do motor rewinders even exist ? Other than for small motors (and especially small commutated motors) rewinding is cost effective.

Whilst clearly you would do whatever is most cost effective - I'm just pointing out that rewinding is possible - and while you are at it you may as well jack up its performance to utilise the extended frequency range offered by the VFD. Swopping to a three phase motor - and then rewinding that for a higher frequency would be pointless if you could do the same thing with your old single phase.

Rewinding The Motor.

On stripping the motor I found the start winding was a lighter wind vis:-

Slots = 24 Single Phase 4 Pole

Winds................4 x 3...........................4 x 3
Span..............1-4 3-6 5-8..........1-4 3-5 5-8
Turns...........35-61-35..................37-67-39
SWG.............22 SWG....................24 SWG
..................(Ø0.711mm).............(Ø0.559mm) (Sorry about all the dots - site's software suppresses spaces & tabs)

I’m not sure whether the 37-67-39 start wind was an accident (i.e. it should have been 39-67-39)
Or was intentional to introduce a small phase angle of ±5°? Because of the small resistive component, the 90° phase shift caused by the capacitor is always “off” by a few degrees – so this may well be intentional to “correct” this and improve starting toque – but it hardly seems worth the bother – I suspect it’s an error.

I need to change the impedance to such that its original impedance at 50 Hz. Occurs at 125Hz.

To solve this you can safely assume a straight line relationship of voltage vs frequency of the form
V = mf + c where c (constant) is the zero Hz voltage and m is the multiplier equating to Volts per Hertz and f is the frequency viz :-

There are various methods of estimating “c” but using 5% as a rule of thumb is good enough.

So c is 11 Volts and the voltage will rise to 220 Volts at 125Hz or (220-11) ÷ 125 = 1.672 Volts per Hertz.

So our new 50Hz voltage will be 11 + (50 x 1.672) = 94.6 Volts

So we must rewind to get the 50Hz impedance down by the ratio of 94.6 ÷ 220 = 0.43

However impedance is to the square of the number of turns and is linearly affected by the area of the iron core vis :-

O.K. so the run winding was 131 turns spread over 3 pole segments per half pole, so for our 94.6 Volt three phase wind wound over two pole segments we need :-

[ √({131^2} x 94.6 ÷ 220) ] x 3 ÷ 2 = 128.8 turns ≈128 Turns (Note: ^2 means "squared" the site's software doesn't support superscript.)

The area currently occupied in the slot is (131 x π x 0.711^2 ÷ 4) + (143 x π x 0.559^2 ÷ 4) = 86.73mm^2

This will be replaced by 2 x 64 turns (per slot) of thicker wire vis :-

(128 x π x Ø^2 ÷ 4) = 86.73mm^2

(The actual area of the slot will be bigger due to packing density which can be from 85% to 90.6% max. theoretical – for simplicity I will presume the before and after pack density will be the same – thicker wire in a “relatively” small slot will have slightly worse density – keep that in mind when rounding.)

Therefore Ø = √ (4 x 86.73 ÷ 128 ÷ π) = 0.929 ≈ 20 SWG = Ø0.914 (Always round down.)

So the rewind data for 90V 50Hz. 3 Phase 4-Pole will be :-

Slots = 24 Three Phase 4 Pole

Winds...........3.........4 x 2
Span.....................1-3 2-4
Turns.....................64-64
SWG..................20 SWG
.......................(Ø0.914mm)
Connection - Delta

None of this is terribly critical as you can program your VFD to suit – or use the “Autotune” feature if it has one.

Here’s the rewound motor :-

Top – All the single phase starting gear removed.
Bottom Left – The Yaskawa V1000 VFD I’m going to drive it with.
Bottom Right – My rewinder is an accommodating soul but point blank refused to reuse the original 4 terminal block and brought out the windings correctly to a six terminal block.
Photo taken during bench testing.

Actual unloaded performance during bench testing

The dots are actual data points - Pretty much exactly as planned.

Note how the amps started to fall away at 120Hz and then continued to do so after the voltage flatlined at max. 220V. Technically I would have wanted to turn that corner at 125Hz so maybe two turns of wire less would have been spot on - however its as close to perfect as it ever needs to be.

As you can see the output voltage is to all intents and purposes a straight line ending at 220.3Volts at 125Hz. If I extrapolate to zero Hertz I get my intersect at 11.74 Volts (remember my earlier estimate was 11 volts at 0Hz. And 220V at 125Hz – pretty damn close. I also expected to be 94.6V for 50Hz. – that came in at 95.2V for 50Hz.

Had this been a 3-Phase motor, I could have analysed it on the VFD prior to rewinding for much better “data” for c rather than an experienced based “5% best guess” but as you can see you don't need to be spot on.

The motor now puts out 1400W @ 3700rpm and can sustain that for 100% rating - it will push out 2.2 to 2.5kw for a low duty cycle - so I've got way more grunt than I used to have.




Next - installing and setting the VFD.

Regards, Ken

 

[TSutrina]

This will be interesting. I thought there was often a fundamental difference in the laminations of single phase motors preventing 1PH to 3PH rewind. I suspect those motors are just that inefficient. I have one from a combilathe doing work on my band saw. Seems gutless for a 3/4HP. I have a much larger ABB waiting in queue.

Winding voltage (V/Hz) of induction motors is something few understand, and I originally learned about from you here long ago. I did understand the concept in DC motors, but knew nothing about AC. My CNC lathe runs a 230V motor at 400V to 87Hz. I guess this is easy and common in Europe.

I had an ordeal with a Lenze 10kRPM motor from a mill I got last year. It needed bearings and of course I dropped the rotor. Local motor shops don't seem to understand balance grades and said it checked out and .001" runout on the shaft was OK, but it vibrated badly. I ended up balancing myself. It's not perfect, but these motors are hard to find and new rotors are not available. Lenze offers several windings for their motors to take advantage of high frequency operation.

I was part of a team that built a 3 phase induction motor for a proof of concept for an electric torpedo. It was a 300 Hp motor that I could pick up, about 100 lbs. and I also built a VFD drive that also I could pick up. A 300 Hp 60 Hz motor was in the background of a photo with two of us holding the two parts and we were inside the outline of that motor. The machine was provided 1000 Hz. Typically we made 400 Hz motors and generators. A special silicone steel was used to minimize iron losses. The drive your purchasing keep the frequency low so that the iron used is the same as that used in 60 and 50 Hz motors. For the 1000 Hz motor we also needed special wire for the winding. Every strand was insulated from the other strands.

 

[TSutrina]

Ken, this is a fascinating project. I would not have thought it possible to re-wire a motor from single to three phase, or change the frequency, so I have much to learn from your posts (which I am still digesting).

The obvious question, of course, is why go the route of re-winding rather than simply swapping out the old motor for a three-phase motor. You mention cost and size as the issues. I could see a new 3-phase motor costing more than the materials for a rewind ... but what about a used motor? I would think a used 1 to 2 hp motor might be available pretty cheaply, and might still squeeze in to the size you need ... ??

Of course, there is also the fun factor - I have often done something that could have been done more cheaply by buying something ... but I wanted the fun of doing it myself. My philosophy is, why spend $5 on a part when you can spend 6 hours in the shop making it instead?

You can find the winding pattern on the web for a stator with a particular number of slots for the typical two, four, and six pole motors. The stamping pattern of the lamination are standardized so you chances are high of finding the pattern. If not then at least you have examples to help find a layout. For example a six pole number of slot can be reduce for a 4 pole by the ratio # pole x 4 / 6 = # poles for a 4 pole motor.

 

[Ken I]

TSutrina, Correct - you have to know the winding pattern and what you are going to change it to - but with only 24 slots the options are limited as regards the single phase 4 pole and none existent for the three phase 4 pole wind (there is only one realistic wind option).
So in my calculations above I know I am going to be changing from having it wrapped around 3 pole pieces per winding to being wound around two pole pieces - but if the single phase had been (optionally but unlikely) wrapped around two I would not have to have made the 3/2 adjustment for the core area.
But as your number of slots goes up (Typically 12, 24, 36, 48 etc. etc.) the layout permutations become exponentially larger.
The size of motors we are dealing with here will almost always be 12 or 24 slot.

My motor could have been built with 16 slots - which would have made rewinding it to three phase impossible - but the manufacturer could then only use those 16 slot blanks (very expensive tooling and machinery) for single phase two or four pole and absolutely nothing else. By using 12 or 24 slots the manufactures could accommodate 2 & 4 pole motors in single or three phase plus 6 pole in three phase. I have never seen numbers other than 12, 24, 36 etc. - they may be out there - I just haven't seen one.

For a bit more clarity here's the before and after stator :-

Top left - Showing one of two pairs of "run" windings. See note:
Top Right - The single phase start and run 4-pole winding (Start is Blue, Run is Red.)
Bottom Left - Both pole pairs for one phase only (i.e. 4 pole).
Bottom Right - The full 3 phase 4 pole winding.

Note: In the top left, I did not show both pairs of windings as it then looks like a single continuous winding as you can see in the top right.

You can see the change in pole segments wound around that also has to be accommodated by your revised wind.

Regards, Ken

 

[TSutrina]

I said that you could find diagrams for winding a motor with a number of slots on the web. Since now I know the number of slots here is a 5 minute or less search.

 

[Ken I]

Installing The VFD.

I used a Yaskawa V1000 inverter (which I had to hand) – 220V single phase input and 20-400Hz (configurable) 3-Phase output at 220V max. This inverter can push out as much as 4kW.

I have set it for running from 30Hz to 125Hz off a 2kΩ external potentiometer.

You can get 1.5 kW out of a 0.55 kW motor – but you can’t get 1.5 kW out of a 0.55 kW VFD – the additional power has to come from somewhere. Given the improvement in torque delivery by the change to three phase, I may in fact be able to get as much as 2.2kW out of this motor – final trials and heat build up may place limits on the actual output.

I bolted the VFD above the splash guard on the wall behind my lathe so I could view the output frequency display.

I am going to add a tachometer display (which hasn't arrived yet) to show the actual spindle rpm – in an enclosure under the shelf above the VFD.

I used the original start, stop, forward and reverse switch unit to control the VFD and added the adjuster pot to the change-gear cover plate.

My previous speed range was 170 to 1950 in 6 geared steps – after modifications it will be 83 rpm to 3950 rpm infinitely adjustable – still with geared steps but changes need be much less frequent.

Example: My “favourite” gear = 736 rpm will (after these modifications) allow me to go from 360 rpm to 1495 rpm just by turning a dial – and still have more power and torque than I used to.

For heavier work, simply gear down and rev up – like driving a car.

Setting The VFD.

Depending on your VFD, you have to set the various parameters to get the desired voltage frequency output “curve” outlined earlier.

The simplest form for my rewind would be as outlined above 11 volts at zero Hz. Ramping up to 220V at 125Hz.

However my VFD has an Autotune function – this has two principal option methods :-

a) Rotating Autotune which spins the unloaded motor up at various frequencies and voltage profiles to “hunt down” the saturation limits and determine the best frequency / voltage “curve”

b) Does a static impedance hunt – but this requires the motor rating plate data for input – specifically it needs to know the Amperes rating.

To a lesser extent method a) requires some plate data as a starting point as well.

Because we have now totally screwed up the motor relative to its original rating plate we need to at least come up with some idea of what the rating plate should be.

This can get very complicated so I’m going to play fast and loose with the math here.

We need to invent a new “rating plate” series of values.

We already have 220V and 125Hz – but we do not know the current or power factor – for simplicity I’m going to assume the power factor will be much the same so we don’t really need that and we can calculate the current a couple of different ways. Call them “estimates” since we are not going to use all the formulas (as we don’t know all the motor data – and really we don’t need to either).

If your inverter needs to “know” the power factor, use the original value or 0.85 estimate.

The original motor’s run winding was 131 turns over 3 pole segments.

The current for the motor was given as 3.9A We need to torture that data a bit first :-

Firstly the 3.9A is for the start and run wind in parallel – but since these add up as vectors we have to use Pythagoras (as a close approximation) and I am going to assume the start winding draws half as much current, then (after solving – not shown) :- 1.8^2 + 3.5^2 = 3.9^2
(For those of you not familiar with AC formulae – here we have the apparently bizarre 1.8A plus 3.5A equals 3.9A – rather than the 5.3A that you would get by simple arithmetic. Because the currents are out of phase with each other.)

This is why :-


(We could also just arbitrarily select 85% [3.3A in this case] of the original rating as an estimate.)

So the run winding is ≈3.5A x 131 = 458.5 At (Ampere•Turns)

We are going to have to maintain that over 126 turns but only two pole segments – used to be three - so our new current will be :-

I x 126 x (2÷3) = 458.5 … Solving for I we get 5.45 Amps

Another way we might get this is by calculating it from our projected revised output power :-

The original motor was 550W @ 1500 rpm (synchronous speed) – we are going to maintain the torque and run it to 125Hz or 1375W (same torque throughout the rev range).

That derives from 220 x I x √3 = 1375 (we have ignored power factor and efficiency as being the same and therefore would cancel out on either side of the equation.)

Solving for I we get 3.6A (per phase – remember)

But as mentioned earlier we might get a further 25-30% out of the motor by the improvement in deliverable torque and running closer to saturation.
That would bring our Ampere estimation up to 4.5 to 4.7A

So let’s say somewhere from 4.5 to 5.5A.

If you dial in an inappropriately high number the Autotune will error out (because it will detect that it is wandering beyond saturation) – but it won’t hurt to underestimate – I dialed in 4.5A and it autotuned with no problem and ran beautifully at that with absolutely no further adjustments needing to be made – lucky or well guessed. See next graphic.

The following is a graph from the actual output of the VFD display for Volts, Amps & Hertz for an unloaded motor (drive belt removed = free to run) after running Autotune.

As shown in a prior post - you can see the output voltage is to all intents and purposes a straight line ending at 220.3Volts at 125Hz. If I extrapolate to zero Hertz I get my intersect at 11.74 Volts (remember my earlier estimate was 11 volts at 0Hz. And 220V at 125Hz – pretty damn close. I also expected to be 94.6V for 50Hz. – that came in at 95.2V for 50Hz.

To further test the motor, leave it running flat out and keep an eye on the temperature – if it heats up rapidly you have probably wandered beyond saturation limits – (where further voltage increases cause massive increases in current and heat for no gain in output power) – dial down your expectations.

Similarly you can keep on autotuning with increasing values – when you start to see a disproportionate upward increase in current and heat (or the Autotune errors out), then you have reached saturation. Dial back about 5% for safety margin.

I would recommend plotting a graph like the one above as it’s the easiest way to spot errors – I would also recommend running a little either side of your intended range – like the graph above.

Finished Project / Results.

As mentioned my previous top speed of 1950rpm could not be achieved from a cold start and even when “warmed up” took as much as 10 seconds to get up to speed (any longer resulted in a trip) even with my lightest chuck mounted.

The new set-up – still using the belt drive and crappy oil and my heaviest 4-Jaw chuck – did this in the 2 seconds ramp up set on the inverter and did so effortlessly to an insane 5000rpm – I’m unlikely to ever go there again (at least not with the big 4-jaw on).

However the dynamic braking could not stop it in the 2 seconds ramp down and errored out.

It could handle my usual smaller chucks and collet etc.

It can handle my heaviest chuck at 45Hz, in top gear (6th) = 1750rpm or 5th gear at 110Hz. = 3360rpm

This is because of the limit to braking energy that can be absorbed by the dynamic brake – causing a D.C. bus over-volt alarm – you can add an external braking resistor to sink the heat to solve this problem (or extend the deceleration). I’m not going to bother, for the unlikely/rare occasions that I might need to exceed 1840rpm in 4th with my big chuck on, I can just dial down the speed before stopping.

I repeated the test in the 4th gear – but now producing 1840rpm at 125Hz. And it ran it up and down effortlessly in 2 seconds either way.

It even tolerated being thrown directly from forward to reverse – see prior video.
Based on some mains power readings against the high speed / inertia of the system, I would calculate that this rig is now capable of pushing out 1.4kW. pretty much where I estimated it was going.

Taking readings from the VFD showed mostly 2-4 Amps at 220+V = 750W to 1.5kW.

Severe loading against the largest chuck in 6th gear produced a high spike of 6.9A which is approaching 2.6kW – I would again assume my earlier estimates of a sustained maximum possible output of 2.2kW is doable but I wouldn’t want to try and achieve that for continuous running.

I could probably get more but I don’t need it – so job done.

Some Cautions.

I would not recommend doing this to a well-designed and appropriately powered lathe (or any other machine tool) unless you have ascertained or believe that the machine is definitely underpowered and the rest of its construction can handle the increase in performance.

Do not run an overclocked motor at low revs for extended periods at high torque - it will overheat - rather gear down and rev up - I generally try to stay above 50 Hz.

Next - Changing the drive, lubricant and various other odds and ends improvements.

Regards - Ken

 

[Ken I]

Sorry for not finishing this post yet - but in an earlier post I said the lathe has served me well for a number of years - talk about hubris begetting nemesis - the tailstock threaded bush stripped while machining the bushes for the conversion to Synchroflex belt drive.

Guess what - no spares (shocker) - so now I have to machine a new bush with a Ø12 x 2.0 Left Hand Acme thread - can't really see me pulling that off down a 30mm long hole - so a tap it's going to be.

Yeah - right ! better chance of finding a left handed Japanese bicycle thread.

So I'm busy making a couple of taps to sort that out.

Regards, Ken

 

[Ken I]

In the meantime, the Tachometer arrived and I have installed it :-

Nice cheap and effective - comes with the Hall sensor mounting bracket (supplied flat), Hall sensor, magnet etc. but you need your own "wall wart" power supply - lots of those left lying around from long dead cordless devices.

I didn't use the magnet supplied but instead used a rod magnet I have to hand which I fitted into the KM locknut - Ø6 drilled hole - it held at 5000rpm just on the strength of the magnet - but I Locktited it in for good measure. Note: Hall effect sensors are magnetic polarity sensitive.

The photo above showing the speed control potentiometer mounted just above the normal start / stop / forward / reverse controls.
The hole above it was for a PITB bolt which held the door closed - really ? Been replaced by a NiB magnet which keeps the door closed and allows you to just pull it open.
I'll have something to add about that oil level sight glass in a future post.

Rear view - still running on the "V" belt drive for now.
I added additional mounting spindles to store all the changewheels on.
My rewiring still uses the safety switch - but I can now jog (but not run) the motor with the door open.

Regards, Ken

 

[TSutrina]

Installing The VFD.

I used a Yaskawa V1000 inverter (which I had to hand) – 220V single phase input and 20-400Hz (configurable) 3-Phase output at 220V max. This inverter can push out as much as 4kW.

I have set it for running from 30Hz to 125Hz off a 2kΩ external potentiometer.

You can get 1.5 kW out of a 0.55 kW motor – but you can’t get 1.5 kW out of a 0.55 kW VFD – the additional power has to come from somewhere. Given the improvement in torque delivery by the change to three phase, I may in fact be able to get as much as 2.2kW out of this motor – final trials and heat build up may place limits on the actual output.

I bolted the VFD above the splash guard on the wall behind my lathe so I could view the output frequency display.

I am going to add a tachometer display (which hasn't arrived yet) to show the actual spindle rpm – in an enclosure under the shelf above the VFD.

I used the original start, stop, forward and reverse switch unit to control the VFD and added the adjuster pot to the change-gear cover plate.

My previous speed range was 170 to 1950 in 6 geared steps – after modifications it will be 83 rpm to 3950 rpm infinitely adjustable – still with geared steps but changes need be much less frequent.

Example: My “favourite” gear = 736 rpm will (after these modifications) allow me to go from 360 rpm to 1495 rpm just by turning a dial – and still have more power and torque than I used to.

For heavier work, simply gear down and rev up – like driving a car.

Setting The VFD.

Depending on your VFD, you have to set the various parameters to get the desired voltage frequency output “curve” outlined earlier.

The simplest form for my rewind would be as outlined above 11 volts at zero Hz. Ramping up to 220V at 125Hz.

However my VFD has an Autotune function – this has two principal option methods :-

a) Rotating Autotune which spins the unloaded motor up at various frequencies and voltage profiles to “hunt down” the saturation limits and determine the best frequency / voltage “curve”

b) Does a static impedance hunt – but this requires the motor rating plate data for input – specifically it needs to know the Amperes rating.

To a lesser extent method a) requires some plate data as a starting point as well.

Because we have now totally screwed up the motor relative to its original rating plate we need to at least come up with some idea of what the rating plate should be.

This can get very complicated so I’m going to play fast and loose with the math here.

We need to invent a new “rating plate” series of values.

We already have 220V and 125Hz – but we do not know the current or power factor – for simplicity I’m going to assume the power factor will be much the same so we don’t really need that and we can calculate the current a couple of different ways. Call them “estimates” since we are not going to use all the formulas (as we don’t know all the motor data – and really we don’t need to either).

If your inverter needs to “know” the power factor, use the original value or 0.85 estimate.

The original motor’s run winding was 131 turns over 3 pole segments.

The current for the motor was given as 3.9A We need to torture that data a bit first :-

Firstly the 3.9A is for the start and run wind in parallel – but since these add up as vectors we have to use Pythagoras (as a close approximation) and I am going to assume the start winding draws half as much current, then (after solving – not shown) :- 1.8^2 + 3.5^2 = 3.9^2
(For those of you not familiar with AC formulae – here we have the apparently bizarre 1.8A plus 3.5A equals 3.9A – rather than the 5.3A that you would get by simple arithmetic. Because the currents are out of phase with each other.)

This is why :-
View attachment 120119

(We could also just arbitrarily select 85% [3.3A in this case] of the original rating as an estimate.)

So the run winding is ≈3.5A x 131 = 458.5 At (Ampere•Turns)

We are going to have to maintain that over 126 turns but only two pole segments – used to be three - so our new current will be :-

I x 126 x (2÷3) = 458.5 … Solving for I we get 5.45 Amps

Another way we might get this is by calculating it from our projected revised output power :-

The original motor was 550W @ 1500 rpm (synchronous speed) – we are going to maintain the torque and run it to 125Hz or 1375W (same torque throughout the rev range).

That derives from 220 x I x √3 = 1375 (we have ignored power factor and efficiency as being the same and therefore would cancel out on either side of the equation.)

Solving for I we get 3.6A (per phase – remember)

But as mentioned earlier we might get a further 25-30% out of the motor by the improvement in deliverable torque and running closer to saturation.
That would bring our Ampere estimation up to 4.5 to 4.7A

So let’s say somewhere from 4.5 to 5.5A.

If you dial in an inappropriately high number the Autotune will error out (because it will detect that it is wandering beyond saturation) – but it won’t hurt to underestimate – I dialed in 4.5A and it autotuned with no problem and ran beautifully at that with absolutely no further adjustments needing to be made – lucky or well guessed. See next graphic.

The following is a graph from the actual output of the VFD display for Volts, Amps & Hertz for an unloaded motor (drive belt removed = free to run) after running Autotune.
View attachment 120120
As shown in a prior post - you can see the output voltage is to all intents and purposes a straight line ending at 220.3Volts at 125Hz. If I extrapolate to zero Hertz I get my intersect at 11.74 Volts (remember my earlier estimate was 11 volts at 0Hz. And 220V at 125Hz – pretty damn close. I also expected to be 94.6V for 50Hz. – that came in at 95.2V for 50Hz.

To further test the motor, leave it running flat out and keep an eye on the temperature – if it heats up rapidly you have probably wandered beyond saturation limits – (where further voltage increases cause massive increases in current and heat for no gain in output power) – dial down your expectations.

Similarly you can keep on autotuning with increasing values – when you start to see a disproportionate upward increase in current and heat (or the Autotune errors out), then you have reached saturation. Dial back about 5% for safety margin.

I would recommend plotting a graph like the one above as it’s the easiest way to spot errors – I would also recommend running a little either side of your intended range – like the graph above.

Finished Project / Results.

As mentioned my previous top speed of 1950rpm could not be achieved from a cold start and even when “warmed up” took as much as 10 seconds to get up to speed (any longer resulted in a trip) even with my lightest chuck mounted.

The new set-up – still using the belt drive and crappy oil and my heaviest 4-Jaw chuck – did this in the 2 seconds ramp up set on the inverter and did so effortlessly to an insane 5000rpm – I’m unlikely to ever go there again (at least not with the big 4-jaw on).

However the dynamic braking could not stop it in the 2 seconds ramp down and errored out.

It could handle my usual smaller chucks and collet etc.

It can handle my heaviest chuck at 45Hz, in top gear (6th) = 1750rpm or 5th gear at 110Hz. = 3360rpm

This is because of the limit to braking energy that can be absorbed by the dynamic brake – causing a D.C. bus over-volt alarm – you can add an external braking resistor to sink the heat to solve this problem (or extend the deceleration). I’m not going to bother, for the unlikely/rare occasions that I might need to exceed 1840rpm in 4th with my big chuck on, I can just dial down the speed before stopping.

I repeated the test in the 4th gear – but now producing 1840rpm at 125Hz. And it ran it up and down effortlessly in 2 seconds either way.

It even tolerated being thrown directly from forward to reverse – see prior video.
Based on some mains power readings against the high speed / inertia of the system, I would calculate that this rig is now capable of pushing out 1.4kW. pretty much where I estimated it was going.

Taking readings from the VFD showed mostly 2-4 Amps at 220+V = 750W to 1.5kW.

Severe loading against the largest chuck in 6th gear produced a high spike of 6.9A which is approaching 2.6kW – I would again assume my earlier estimates of a sustained maximum possible output of 2.2kW is doable but I wouldn’t want to try and achieve that for continuous running.

I could probably get more but I don’t need it – so job done.

Some Cautions.

I would not recommend doing this to a well-designed and appropriately powered lathe (or any other machine tool) unless you have ascertained or believe that the machine is definitely underpowered and the rest of its construction can handle the increase in performance.

Do not run an overclocked motor at low revs for extended periods at high torque - it will overheat - rather gear down and rev up - I generally try to stay above 50 Hz.

Next - Changing the drive, lubricant and various other odds and ends improvements.

Regards - Ken