Rewind the HSB36 motor. The epic beginis

[Leslie]

One of my HBS36v GM motor has done me some hard work and after some encouragement to load the ebike to its max, one night the wife brought back one very hot smelling motor.  

I hosed The HBS36 down to cool it.

After a week the hall sensor that was replace failed and I had to put the bike into the dry dock for some work

I found the windings were very dark and the stator had melted plastic insulator melted from the teeth.



Before and after shots.





This motor was pushed time and time again way past its intentional designed limits so in essence ive been very happy with it performance.  It was still functional before I undressed her but as you can see it was on its way to the grave.





HBS WYE Star configuration.


After some good advice from Rolf here as to the winding order, observing the motor and surfing google I came up with this winding pattern.  It is the same as Rolf suggested but drawn out as it is configured in the HSB GM motor.









I gots my high grade enamel copper and some off cut winding paper. $33 dollars so far. The guys at HQ industrial windings have offered to dip this for $10 when I've finished

It was very hard to get these windings off. I used a small flat head screw driver to part the winding down centre of the slots and drilled a hole in between them with plenty of copper between each slot and the drill bit as to make sure I didnt flare the slot layers apart.  Then used pliars to pull out the windings.  

Once I broke the first winding I knew then there was no turning back.  


As you can see the plastic core insulator around the outside of the slots is melted from overheating.  I will remove this and replace it with the paper I got from HQ.

The grade of copper I got is rated above 180~200 deg C (in background) and it has a nice redish tone to it..  The copper on this hub is about 5 shades darker than it was when the hub was new due to the excess heat from overloading the hub.

I haven’t got a good count of the turns yet but last count was 7t6p on the HBS36.  I think 42 strands surround each tooth. I will get a certain on this before I finish.


The paper in between the coils and stator was still in good shape and in good faith GM should observe this for future builds as indeed the paper out performs the plastic by a good margin.


IMO paper in place of the plastic mouldings would still be an improvement on the already hardy design.  I really wouldn’t say what happened was due to a flaw in the design, not at all, I just I see room for better design.

GM should look at the data coming out of this work very closely because they will not find many GM motors that has been pushed so hard in such a way.

This damage has not been done by a 86v over volt this has been done by sheer loading of weight by means of a trailer carrying lots of shopping.




This is how I plan to tackle my problem.


I’ve done a bit of restoration and removed the melted plastic stuff.



Another test wind with some lower temp WW4016 enamel magnet wire "I have to experiment with". This is a 7p 8 turn wind with flat layered incrementing winding style from top to bottom on the first coil. I am contemplating doing a 8p 7t.





High temp winding paper cut into rectangles folded into V's with a 5mm nick in the middle to allow half flaps to overlap each other in the centre of each tooth.

I have a 1kg of that higher grade enamel wire for the real job.  

I want some feedback from a few members here before I do anything.

I shall succeed.

[Bikemad]

the wife brought back one very hot smelling motor....I hosed her down to cool it.

I don't suppose your wife appreciated such extreme behaviour.
I hope you are ashamed of yourself!

Alan
 

[Leslie]

the wife brought back one very hot smelling motor....I hosed her down to cool it.

I don't suppose your wife appreciated such extreme behaviour.
I hope you are ashamed of yourself!

Alan
 

 

 

Ouch

Shall fix that one.

 

[Leslie]

I tell no lies.  This wasnt the load that near killed my baby either.  It was 2 weeks after this and it was worse.  
 



There's two 40 litres of pottong mix in that lot too.

I could of hosed the wife down.



As you can see I have a lot more than a trailer full of shopping to deal with.

This bike is a tank to start with and those 4 X 24ah SLA's weigh in at 7.2 kg each. 28.8kg total.  

Every 2 weeks the loads got bigger.  She road this load 14kms. 28kms round trip.

As you can see What I mean by being happy with the motors performance.

Outstanding really.

[Leslie]

I am still stuck on how many turns to parallel winding I am going to put on this.

All I have to go on ATM is a crystalyte 400 series.

I counted 51 teeth on the motor and near similar is not the same dimention stator.

Each .5 winding has the ability to cary 2.2 amps. 8p can draw 18 amps  At best two phases are open total 36 amps.

The original motor had 6p copper making that 26 amps.

My controller will push 40 amps however I keep comming up with 7 turns 7p as the perfect balance.

I wish someone could tell me how many turns are on the HBS48 1000 watt motor. Or any of the HBS motors for that matter..

No one can ever answer my questions. 

[Leslie]

Determined to get more accurate data I wound 8 series teeth with seven turns with a single strand winding and got myself 160uh.

I tested around 840 uh over 34 windings originaly.

I took my 160~165uh and divided it by 8 teeth and multiplied this by 34 and came up with 680uh.

This is still out of range and too low for my likes.

I want to be in the 700uh+ ball park.

One more go with 8 turns.

I did witth the same copper piece 7 teeth with 8 turns and did the same math above but over 7 teeth and got 864uh.

This is closer to the orginal figure of my hub.  Given the copper is different too I wont fuss just yet.

So now to decide.

7 turns or 8.

Im pretty sure my original hub configuration was only lacking a little torque agaisnt the wind and such so I will stay with 8 turns and add two more parralel windings to give me more torque.  This will make a similar speed possibly a little more and but closer to constant.

Next I will do a test run on a few teeth with my good copper to see any differences.

And then she gets all dressed up for the show.


 

[Electrobent]

"My controller will push 40 amps however I keep comming up with 7 turns 7p as the perfect balance."

some questions:  How are you coming up with this number?  Math?

What does the p mean?  Number of wires?

Why don't you need the paper on both sides?  Is is just to protect the wire from the sharp edges of the stator poles? Is that what the plastic that melted on yours was for? 

Did you build some jig to hold 8 spools or how are you getting the multiple strands to stay even? 

Good Luck!

[e-lmer]

7 turns 7p

I think the turns is the number of times the parallel (p) bundle of wires
wraps around one of the poles.

So 7-7 means take 7 strands, solder one end together and to a lead,
and wrap all seven around the pole.

The parallel wires allow you to send more current that you could through
one wire.  This configuration gets you 49 turns around the pole, and divides
the current between the 7 strands (so 50Amps is 7ish Amps per wire.)

[Leslie]

The wire can only handle 2.2 amps each this is where resistance or (impedance is the better definition n) applies.  I was lucky to even trip my ohm meter on two phases so some guess work is going to be implented of how long 7p7t would actually be vs the data on the resistance of the chosen copper around 17 slots.

The more parallel windings takes the resistance down

1/R_Total = 1/R₁+1/R₂ +1/R₃ +1/R₄ +1/R₅ +1/R₆ + 1/R₇      

1= origin; as the point where all current comes from so 1 (origin) must be divided by each and every parallel R (ohms) to gain how much resistance the windings will have.  

Not enough impedance (resistance) and this behaves closer to dead short and heat appears over the controller Fets or winding depending on how much the wire can handle.  

7p windings with an ideal impeadnce will draw 15.4 amps. Two phases may be operational in parallel giving me 36amps max.

This is has not much to do with turns as this is when voltage steps in.  Like resistors in series inductors in series are similar and work as voltage dividers and convert voltage into current and in turn MF magnetic flux.

To get a hold of the whole thing one must have adequate impedance over the whole phase/s adequate inductance to covert to MF which is the driving force and adequate voltage and current and silicon fet drivers initiate all qualities of a functional motor.



The problem I face is that I may do more efficient windings than the original job so the inductance may be higher at the expense of impedance.  This is not a bad thing though.  To gain MF in this way makes the motor more efficient, so less, more efficient turns will give me more speed and less impedance per inductance, and more turns efficiently will give me more power and inductance per impedance.

So the more I think the more I like the idea of 8t 8 p if I wind better.

8 turns will increase impedances and take the max current down per winding and 8p, it will add another +1/R₈  to the resistance to a 1/7 ratio but will provide 2.2 more amps head room.  

Maybe 7p 8t is another option giving me some handing resistance to play with.  

If that hasnt confused you, youre doing better at this than me.

[Leslie]

The reason voltages are not being considered into the resistances with my rebuild is that I do not plan going over 48v.

Even if I was to wind the most efficient inductors over 17s + 17s2p or 34s WYE slots  in between a positive and negative PWM supply, a 6t winding would still perform in spec to a 48v supply.

This is mainly the reason why people can and do run their stock GM motors at 84v.

These hub motors have long phase lines with plenty of inductance to duct voltage through the phases.

It is the load and heat I am concerned with.


My only limitation would be the size of the magnets and space to wind copper, being turns or parallel windings.  I still must have my cake and eat it.

Can any one test the inductance of a Pie. @ 96v, wind tears would be the only blur you would witness..

[Leslie]

« Last Edit: Today at 05:50:53 AM by Bikemad »
(formula corrected)
 
  
Formula was already correct for equal resistance wire.

Those sub chars are used to lable resistors in P as you may have different value resistors/


But if youre bored enough to label each and all my parallel wires so you know the difference, go ahead as they all look the same to me.

So others don't think I am a total noob

The original formula I posted was.

1/R= 1/R+1/R+1/R+1/R+1/R 1/R+1/R you just need to count the wires and put it into a calculator and viola.





I have discovered that 8 turns over 3 slots spools out close enough to a meter of wire so more to come..


Hey Bikemad!

Seeing you’re feelin physicsy and all.

What does the very simple E=MC² mean to you.

One condition, No google.
 
My own thesis of “Origin” wraps this up and spits out dark matter into space particles.

[Leslie]

I tried and tried, I pulled those wires until they broke.

I started and finished and started again, I spooled.  

There is no way in hell I can wind 8t 8p or even 8t 7p on these lil teeth.

I would get three done at 8t 7p and some mishap would set me back to point A.

The lesson learned. "Never take on a GM hub builder and better them".




As soon as I went back to 8t 6p "the original winding config" the jobs started to proceed ahead.

This may be as good as a GM wind but I doubt it.



Based on my diagram above the job here looks just like the hub did before I undressed it.

The windings may be a little shorter so I may get some more top speed.  

The copper insulator sets a yummy red tone to the copper.  When this is dipped it will look very pretty.

This hub with a sensorless contoller should be good up to 180 deg C.

[e-lmer]

The key has got to be to keep the parallel windings parallel when
you wrap, every time a wire crosses it costs you room.

I see in your pic that the group is in a bundle of wires
instead of a flat band.

You might try taking your 7 and run them straight along
a strip of masking tape to keep them parallel, then peel the
tape by inches as you wrap.  It looks like 7 wires would make
about a turn and a half per level, with the band turning over
at each end of the spool

Here is an example of what I mean wrapped on a wrench.

[Bikemad]

Elmer,

That sounds like a very good, but time consuming suggestion.

If you were to have seven spools of copper wire and feed them through the eye of a suitably sized darning needle, you could simply pull the wires through and apply the masking tape in one easy operation and wind it onto another spool ready for use.

If the stator was mounted in a swivel vice like this:
it would allow the stator to be slowly turned around the centre axis of each tooth in turn, allowing the home made copper ribbon cable to be wound neatly onto the stator without becoming twisted.

Anyway, you seem to have invented a rather unique tool, an electromagnetic wrench which will hold your nuts in place, but I don't imagine there will be much call for it.

Alan
 

[Leslie]

Thank for the advice.

Naaa its too late.

I lose my turn count almost every second tooth and have to unwind. Its "turn start turn finsihed". uhh ohh was that start or finish?

I almost don't count now I just turn until it looks big enough and test the inductance, one turn too less and it reads .01mh one turn too many and it reads. .03mh.

When I do a series of 3 teeth it sits on .07mh for about 2 seconds then down to .6mh.  So I think Im getting about 21.2uh per tooth.

This is around 720uh per two phases in series.. The current may be an issue, as my controller can load 40 amps and this hub will be lucky to pull 30.

I might remove a shunt pipe out of the controller and choke the circuit.

One or so nicks on one of the windings that don't touch anything, inductance readings don't lie.

Charing and removing the enamel on this copper is difficult a lighter doesnt heat it up hot enough.  The propane torch is a lil better.  This motor will be allowed run somewhat  hotter and I hope after dipping it and giving it a good soak will make sure the job lasts.

I am prepared for failure but am optimistic this  will work.