Junkyard motors parade

AC, DC, amps, volts and kilowatt. It's all discussed in here
User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Mon, 22 Jul 2013, 13:11

jonescg,
I think these windings work a bit different from usual. My guess: there must be 12 poles, 4 for each phase.
If you accept slots 1-4 as one pole, and slots 13-16 as other pole (even when all wires members of the same coil group) I think all makes sense. In this conf., each coil group actually works as 2 poles.

Turn per coil are currently 56/4=14. In the setting I suggested, it would be just 56/28=2. For all the coil group (4coils) it would 8.

I am still learning the jargon, so if anybody thinks I am misusing any term please feel free to correct me.

T2
Groupie
Posts: 121
Joined: Sat, 08 May 2010, 04:51
Location: ON,CANADA

Junkyard motors parade

Post by T2 » Mon, 22 Jul 2013, 16:02

I am glad you have taken up the idea of downsizing to a 5.5Kw.

You expressed anxiety over the reduced efficiency expectations with smaller motors, but aren't those specifications based on 50/60Hz operation and outside of the higher rpms of automotive usage ?

Tests reported here three years ago (AEVA- 104v motor) indicated that operation at higher rpms improves efficiency. I would say that Motor efficiency is not a problem right now.

Your book drawing concentrico por polos consecuentes with K=48
is exactly where I am. There are minor differences from my previous post which described a K=36 stator layout.

Your book reference shows a LAP winding. I proposed a WAVE winding.
MMF-wise the results seem to be entirely the same.
On introspection the Total Path Length of copper is also identical.

Let's try the numbers

I have to admit until Siwajasta came along I didn't have a handle on a rewind quantitatively but now with that info you can do a lot with very little.
In my 36 slot design for 5.5Kw I multiplied his 102 20awg wires per slot by 5.5/7.5 yielding 75 wires/slot for the smaller 36 slot stator

This is what I got, if anybody see something suspicious in my numbers please let me know.
There are 56 wires in each slot, 4 in-hand, single layered.


Does this new machine have 48 slots same as your book page ?

If so, your 5.5Kw with 4 slots/pole = 224 wires total/pole

SIWAJASTA's 7.5Kw with 3 slots/pole = 306 wires total/pole
My estimated 5.5Kw with 3 slots/pole = 225 wires total/pole

The 224/225 results show remarkable correlation between an actual 5.5Kw and a 5.5Kw intrapolated from the 7.5Kw frame size.

So, a "simple" way to lower the voltage would be using the same scheme, but 24 wires in-hand. That would result in 32vac. Right?

You will have to tell us the name plate voltage for a WYE configuration in order to verify that statement.

To avoid confusion surrounding the use of 50/60Hz and for the benefit of the North American audience, later on I would suggest we start the custom of adopting the use of V/Hz units.

In the eighties I remember a casual reference to a 68Vac 240Hz winding which at 0.28V/Hz is not too distant from a 100Vac aircraft motor running from 400Hz. Surprisingly there was never any suggestion at the time for tapping aviation sources for AC traction motors for EVs. There were no inverters available when they first appeared on aircraft and I have it on good authority that these motors are mostly for intermittent duty tasks, for example, like raising/lowering landing gear. And they got away with DOL starting on 400Hz apparently. That sounds awful painful to me. Perhaps they had deliberately bad impedance rotors.   
T2

User avatar
Bluefang
Groupie
Posts: 90
Joined: Wed, 13 May 2009, 22:37
Real Name: Derek Hohmann
Location: Gold Coast, Queensland

Junkyard motors parade

Post by Bluefang » Mon, 12 Aug 2013, 02:51

Hello, Any chance you can do me a huge favor and count how many laminations are in 1cm? I am trying to figure out the avg thickness of lamination in this size motor as i might be able to get a pair from work for nothing to rewind. Though i will be attempting to rewind as BLDC with multiple controllers wound separately in the motor similar to my current e-motorbike projects hub motor.

Its been discussed on the endless-sphere forum http://www.endless-sphere.com/forums/vi ... 30&t=51469 Turn the 50kw peak motor into a 150kw BLDC motor

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Fri, 16 Aug 2013, 11:13

T2,
The 5.5KW motor rewind is something I am still pondering, the info I posted referred to the 9KW one. I am rewinding that one first, if it works without much vibration (I don't really know if the squirrel cage is broken) I will test it on the dyno.

This motor is 230V/400V, 50Hz. It had 56 wires in each slot, 4 wires in-hand.
So I went the easiest way (according to my books), and calculated:

56 / 4 = 14 turns
Divisors of 14: 2,7,14. I choose 7, which gets good results
14 / 7 = 2 turns
If I use 2 turns instead of the original 14, I get 230/7=32.8V & 500/7=57.1V, two nice voltages for playing with the Curtis.
56 wires / 2 turns = 28 wires in hand.

I guess that makes 0.65V/Hz & 1.14V/Hz, right?

Bluefang,
sorry, I read your post too late. I just sprayed my motor with the red spray!
I can count the laminations of the rotor if you are interested; but I am not sure they are the same as the stator.
BTW, I have been reading that thread. Really cool work, it's a pity nobody is trying something similar with a pure reluctant motor, which is much cheaper to produce (and much more ecological imho).

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Fri, 16 Aug 2013, 11:44

Time for some progress report.
My Curtis 1238 came from customs and we checked it with our "testing mule". It ran perfectly; 2 advices when setting a Curtis:
- Wire the interlock thingie, it will not work without it. Disabling the controller option is not enough.
- If it works, but slow and vibrating, maybe you need change the encoder direction/phases.

I bought some material for rewinding motors. As we are in august and everybody takes holidays here in Spain, the supply store was nearly abandoned. The only person remaining there was not an expert at all, so I had to buy a bit blindly.
Image
That's what I got:
- 15KG of copper wire, 1mm. 200E (290AUS)
- 20 square meters of insulating paper, 0.15mm (probably too thin) "Tercott CF DE". 56EU (82AUD)
- 2L of varnish, Royalac 30, air dry, class B (130C). 18EU (26AUD)
- Anti-flash spray (red spray). 8EU (11AUD)
- Polyester tie tape roll. 11EU (16AUD)
- 50 meters of insulation tubing (waste of money…). 41EU (59AUD)

That's material for 2-3 motors.

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Fri, 16 Aug 2013, 11:55

Yesterday we finished cleaning the motor. First we started with the top & bottom of the stator:
Image
Image

Then we went for the slots; about 5/6 slots contained remainings of the old insulating paper (which was pure plastic to my eyes). These remainings were a bit toasted and firmly glued to the walls of the slots.
In order to clean this plastic we tried a thin threaded rod and a gas torch.
Image
Image
Image
That didn't remove all the plastic, but I think it was clean enough for rewinding.
After that, we blowed all the dust using compressed air and procceded to paint the stator with insulating spray.
Image
Image
And that's all for now. Hopefully this weekend we will start the actual winding.
Image

User avatar
Johny
Senior Member
Posts: 3666
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Junkyard motors parade

Post by Johny » Fri, 16 Aug 2013, 15:36

Nice work. Was the stator coated (for insulation) originally?

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Sat, 17 Aug 2013, 06:05

Nope, it had 2 thick plastic discs. One of them was fragmented.

I saw this insulating spray on the shop and it was cheap. As Ivan Bennet and others are using it, I decided to buy one. I think it's a novelty in the world of rewinders.

T2
Groupie
Posts: 121
Joined: Sat, 08 May 2010, 04:51
Location: ON,CANADA

Junkyard motors parade

Post by T2 » Sun, 18 Aug 2013, 23:42

Peskanov, you wrote
If I use 2 turns instead of the original 14, I get 230/7=32.8V & 400/7=57.1V, two nice voltages for playing with the Curtis.
56 wires / 2 turns = 28 wires in hand.

I guess that makes 0.65V/Hz & 1.14V/Hz, right?

Let's take a look.

56 wires 4-in-hand = 14 turns = 400Vac in WYE connection
56 wires 28-in-hand = 2 turns = 400Vac/7 = 57.1 Vac in WYE connection
57.1Vac/50Hz = 1.14 V/Hz

Sure I agree.

OTOH, since an inverter running from a 100Vdc bus is only capable of supplying 70Vac line-to-line voltage.

It follows, therefore, that the maximum allowable frequency before the inverter voltage headroom is lost = 70/1.14 = 61 Hz = 1830 rpm[/b] using the motor connected in the WYE configuration.

Even if you tried 56 wires wound with all 56 wires in hand = 1 turn = 400Vac/14 = 28.5Vac in WYE connection gives only 3660 rpm.

Original line current with 4-in-hand was 16.4 Amps
So 56-in-hand will allow 230 amps continuous which is compatible with the 155 amps continuous current rating of the Curtis controller.
T2

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Mon, 19 Aug 2013, 11:47

T2, I see you always write about Star configuration instead of usual Delta one. What you propose (1 turn in Star, 28VAC) is pretty similar to what I am trying (2 turn in Delta, 32VAC).
Using your config (1 turn) in delta would provide even lower voltage, 16VAC, moving the max. power point to the 6000 rpm point. Too high revs imo.

The problem I see using lower voltages is that I loss the maximum torque of the motor. The specs of this motor (well, the closest ones I found) say Tm/Tn=3.1
This a minimum of 3x nominal amps, surely bigger because of heavy losses at high slip. If nominal is 230A (you example), 3x is 690, much higher than the modest 550A provided by the Curtis.

Wiring for /7 delta, I get a nominal current of 200A. Still much higher than the 550A of the Curtis, but I think it's a bit better.

In the other hand, going for frequency is much more efficient and will provide same torque using the correct transmission ratio. And I plan to keep gears, so I can do it.
Getting the full torque would be more interesting if a had a more powerful controller, then I could maximize both frequency and torque...

Jesus, what a headache! Image
Last edited by peskanov on Mon, 19 Aug 2013, 04:41, edited 1 time in total.

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Mon, 19 Aug 2013, 12:16

Bluefang, in case you are still reading this thread: we counted the layers despite the red spray, it's 16 in 1cm.

Report from our 1st day of rewind.
As the conductor is very thick (28 wires of 1mm), we thought about minimizing internal connections. The best way is to prepare as many coils as possible outside of the motor and then trying to fit them in the motor as one piece. The maximum you can do is 1 phase at a time, although it's pretty tricky as you have little space to work. Looking at the scheme, this means preparing a coil like the bold one (U-X).
Image

Our attempt failed, basically because we suck at fitting the wires in the slots. Next day (tomorrow) we wil try another method to accomplish the same thing.

Anyway, here is what we did.
First, we cut and inserted the insulating paper. I used a cheap paper cutter, it helps a lot.
Image
Image
Next, we built the coil pattern on a wood panel using nails.
Image
We also used this to get the length of the wire (13 meters).
Cutting 28 wires (13 meters long) was a bit tedious, but at least you exercise a bit.
Image
We inserted insulating tubes in order to pre-insulate the coil heads. That also helps keeping the wires packed.
That's possible & useful when you are doing only 2 turn per slot.
Image
Forming the coils. That was pretty fast, once the nails pattern is ready.
Image
Image
Image
That's it, a whole phase ready for insertion.
As I said, we abandoned this path as inserting the wires in the slots proved to be a pain in the ass. Maybe with more practice or the correct tools it's easy, but we found a more promising method and abandoned this one.
That's a pity, because that means 2 KG of copper lost Image. Well I still have 13KG and lately copper returned to it's old price of $3/KG so I should not worry Image.

User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Junkyard motors parade

Post by BigMouse » Mon, 19 Aug 2013, 15:37

Is having insulating tube on the wire heads a good idea? I would think the insulating tubing would stop the varnish from fully impregnating the wire bundles and locking them together (against vibration) as intended. I suppose all the wires inside a given tube are the same potential, but you still don't want the wires vibrating away and fatiguing.

I'm looking forward to seeing how you go with this. Thanks for documenting it. I hope your squirrel cage is good and this motor ends up being usable when you're finished.

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Mon, 19 Aug 2013, 23:59

Yep, that problem is in my mind. I have seen coil head insulated with plastic and tied strong before, so I guess air bubbles are not much of a problem.
However, I was thinking about making some cuts in the tubes (after winding) to allow varnishing.
Another way would be to insert 2 tubes instead of one. I can keep them slightly separated while the varnish floods the motor.
I still have to check the viscosity of the varnish...

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Wed, 21 Aug 2013, 05:26

Another day working on the motor...

We have discarded the insulation of heads using tubes. Once you get a group of coils done, you must mold them in order to make some space for the next ones. Space is scarce, specially if you a noobie (like me), so I discarded that idea.

Ok, this time we tried threading the wires longitudinally. That worked, but at the end our coil heads were too uneven. Also, some slots required too much work. Probably because some wires crossing inside the slot.

So we are again at the starting point. This thing of rewinding motors is an art! Image
I am not going to document the fails anymore, as I am pretty sure I have a lot of them in front of me. When I get a good coil group done I will try to document it.
Last edited by peskanov on Tue, 20 Aug 2013, 19:28, edited 1 time in total.

User avatar
Johny
Senior Member
Posts: 3666
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Junkyard motors parade

Post by Johny » Wed, 21 Aug 2013, 15:37

Thanks peskanov. The quick search I did on google showed lots of rewind before and after shots plus some in between but they appear to leave out that step where they get the copper into the slots. One video on youtube showed the guy with a wooden ruler and a hammer smashing the coil group into the slot - doesn't look to reliable to me....

Anyway - let us know what works and maybe a single sentence for each attempt (or idea) that failed when you do post the final one.
Thanks again for sharing.

User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Junkyard motors parade

Post by BigMouse » Wed, 21 Aug 2013, 19:58

There is a series of really good rewinding videos on Ivan Bennet's forums.

User avatar
Johny
Senior Member
Posts: 3666
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Junkyard motors parade

Post by Johny » Wed, 21 Aug 2013, 20:19


User avatar
BigMouse
Senior Member
Posts: 600
Joined: Thu, 28 Oct 2010, 02:39
Real Name: Vincent Tannahill
Location: Silicon Valley
Contact:

Junkyard motors parade

Post by BigMouse » Thu, 22 Aug 2013, 01:50

These are the videos I was referring to. Sorry I didn't link them earlier. Was posting from uni and didn't have time to go searching for them.

Click here for the thread with videos in context.

Specifically, it was very handy to see the winding technique.

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Thu, 22 Aug 2013, 05:40

Thanks for the links, Ive been checking Ivan's work, but somehow I didn't read the one Johnny posted.
There is something I don't understand about Ivan rewinds, he seems to work with half the wires I am dealing with...and I am using same gauge (18, 1mm) and have to fill more slots (48 vs 36). Maybe he is doing two layers?
My slots are packed to death, while he seems to have much more space. In the other hand, he is winding frames 160.
Puzzling...

User avatar
Johny
Senior Member
Posts: 3666
Joined: Mon, 23 Jun 2008, 16:26
Real Name: John Wright
Location: Melbourne
Contact:

Junkyard motors parade

Post by Johny » Thu, 22 Aug 2013, 15:24

Well he agrees that the more wire you can squeeze in the better.
Would it be easier to do say 2, 3 or 4 bundles so that they can fit through the narrow part of the slot then link them all up? Seems easier than single wires.
Then again - I'm just in the gallery watching so advice is easy....

User avatar
woody
Senior Member
Posts: 1701
Joined: Sat, 21 Jun 2008, 02:03
Real Name: Anthony Wood
Location: Mt Colah

Junkyard motors parade

Post by woody » Thu, 22 Aug 2013, 16:15

Looks therapeutic, like building a bicycle wheel...

I watched the first video - it shows him putting 3 passes of 18 strands in each slot, he is going to end up with double that in each slot - 2 * 3 * 18 = 108 strands.

Maybe work backwards using your scrap copper - see how many strands you can cram into one slot and work backwards (if you are winding like him, divide it by 6).
Planned EV: '63 Cortina using AC and LiFePO4 Battery Pack

User avatar
peskanov
Groupie
Posts: 153
Joined: Sat, 15 Oct 2011, 00:55

Junkyard motors parade

Post by peskanov » Fri, 23 Aug 2013, 08:03

Johnny,
I was trying to keep the method of professional rewinders I know, keeping exactly the same total copper section. But this is not so important when you use a Curtis, as it does not care about the "nominal voltage" of the motor at all.
But I have seen Ivan puts as many copper as it fits in the slot, without packing it too much.
My motor was packed to the max! Image
He also uses an easier to do layout. Coil heads in this layout are much easier to do! I am considering using it.

Woody,
I don't think he is putting 108 strands in a slot (more like 3*18 only), but I am going to ask him, I am registering in his forum.
Yes, I am considering reducing the number of strands. I have been able to fit 56 strands (as the original layout), but it's too hard, so I am thinking about removing 8 or 10.


T2
Groupie
Posts: 121
Joined: Sat, 08 May 2010, 04:51
Location: ON,CANADA

Junkyard motors parade

Post by T2 » Fri, 23 Aug 2013, 15:51

Yes, I prefer the WYE connection for two reasons.

First, the phase windings and their magnetic circuits must be almost identical for all three phases otherwise with the mesh (delta) connection there could be circulating currents around the mesh that could cause unnecessary heating of the stator winding.

Second, the WYE allows the possibility of the lead out wires of each phase to be made to the exact length that would reach directly to the controller output terminals thus avoiding extra terminations at the motor. I have seen the motors of the Prius similarly attached.

Then you wrote :
moving the max. power point to the 6000 rpm point. Too high revs imo

For best performance from any motor you want the base speed point to be as close to 9000rpm as convenient with the road speed about 100km/hr. Adopting the old idea of placing base speed at a low value as possible is no longer a good modern strategy IMO. It is the availability of high current controllers that have made this possible.

I see no problem with extra low V/Hz machines assisted by high current controllers PROVIDING overall gear ratios are 8.5 or higher.
T2

T2
Groupie
Posts: 121
Joined: Sat, 08 May 2010, 04:51
Location: ON,CANADA

Junkyard motors parade

Post by T2 » Fri, 23 Aug 2013, 23:04

Peskanov you wrote :

The problem I see using lower voltages is that I lose the maximum torque (capability) of the motor. The specs of this motor (well, the closest ones I found) say Tm/Tn=3.1
This a minimum of 3x nominal amps, surely bigger because of heavy losses at high slip. If nominal is 230A (you example), 3x is 690, much higher than the modest 550A provided by the Curtis.


Well, perhaps the motor is trying to tell you something. That this new winding has made the motor too powerful for the controller.

If you go back a few posts you will find Richo and myself, initially recommending that you start with something smaller, like perhaps 5.5Kw ?

You, however, insisted on going ahead with a 9Kw. Image Of course, having a slightly heavier motor is not a serious problem to have. You always have the option to switch to a 650A controller later on if things turn out to be successful.

In the meantime a delta connection giving 0.65 V/Hz is much too high for 100Vdc. That's just my opinion of course.
I want you to read the following which I lifted from the DIYELECTRIC website.

Find :       > EV Conversions and Builds > All EV Conversions and Builds > grant electric mini

"From tomofreno Post #34
The HPEVS transaxle has a ratio of 8.5:1 - the transaxle is single-speed. The AC50 has a max motor speed of 8500. The limitation is actually the controller, as the motor and the transmission have a top end of 10K RPM.

Thank you. I have a Swift (http://www.evalbum.com/3060) with an AC50 and nominal 115V pack. With 8.5:1 the estimated zero to 40 mph time would be about 7 sec and zero to 60 mph would be about 23 sec. It really dies at the end due to low torque at high rpm. The zero to 60 mph time with the original transmission is about 16 sec (measured). I don't need 5 speeds, so kind of watching for something like 2 speeds, and thought maybe 1, but the AC50 doesn't have enough power for that to work well at highway speeds. The 8.5:1 should work well and be very simple and convenient for lower speed city/secondary road driving though - all depends on what you want.
From PThompson509 Post #37
For those following, the transaxle is from Graziano, and is the UVT 900 gearbox. My friend had it modified to use 8.5:1 instead of the more standard 10.5:1."

Notice that bit about It really dies at the end due to low torque at high rpm. That sounds like what 0.65V/Hz will do for you also. Just sayin'. Image
T2

User avatar
weber
Site Admin
Posts: 2384
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Junkyard motors parade

Post by weber » Sat, 24 Aug 2013, 17:39

peskanov wrote:Image
It's interesting to read about your efforts, but why would you be winding a motor as 2-pole when it is to be used with a VFD?
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Post Reply