Maximizing Copper Fill

[hias9]

Hub motor stators usually have a piece of wood and paper on top of each stator slot.
Is this additional insulation or too keep the copper windings where there are?

Would it be a good idea to remove these pieces and add maybe 1-2 strands of winding wire to increase copper fill?

[markz]

Slot liner insulation protects the winding conductors during insertion into the stator slots.

The slot liner insulation also provides additional insulation between the motor winding wires and the stator laminations.

The slot liner insulation also provides extra insulation at the sharp edge where the motor windings exit from the motor stator slots.

Phase to phase insulation can also be used to add additional insulation where large voltage gradients are present between adjacent windings.

[larsb]

you'll find that those extra wires don't give a large increase in fill if it's not a (really) crappy motor winding.

The other thing is that the windings closest to the magnets are more affected by eddy losses. I’ve seen some tests where it wasn't worth it to pack the final layer closest to the slot opening as the lower resistance was offset by the increased eddy losses.

[markz]

There was a discussion on about if you could make a square wire if that would be beneficial.

Could be this one, search term "fill square" 10 pages worth
3 posts with "square" in same one
viewtopic.php?f=30&t=102934&p=1505953&h ... e#p1505953

Read the quote
viewtopic.php?f=31&t=65031&p=1197760&hi ... l#p1197760

Iron Fill
viewtopic.php?f=30&t=90037&p=1445084&hi ... l#p1445084

[ekline309]

Square wire is great in theory but in my experience it is a huge pain, especially when you are moving from one layer to the next. This is the best article I've ever found about maximizing copper fill: https://ieeexplore.ieee.org/document/651289

[larsb]

I’ve done some square wire winds, it’s really dependent on the way you shape the wire. This is a 60% winding.

F5428A80-C891-4365-9FAF-5BCA42ABB9B2.jpeg (83.96 KiB) Viewed 603 times

From there it’s hard to get a large increase in fill whatever method is used. 90% or whatever is mentioned in the paper won’t ever be wound on a motor coil.

[Chalo]

Makes me wonder where there's anything in using hexagonal wire. It would at least be able to pack in more densely, like square wire but easier to work with.

[larsb]

I don’t find rectangular wires to be that difficult but it’s true they cannot be wound as quickly as a softer/lower diameter multiwire wind. The thicker wires have the benefit of being bulletproof in comparison to the super thin wires on many chinese rc motors.

Anything less than a true 30% is a poor winding and anything above 60% is remarkable in my book.

[hias9]

What are your experiences about the possible bending radius when using rectangular winding wire?
When rewinding a DD hub motor with 36 or 51 slots and 3 turns I would guess that using a single 5.5x3mm wire would not be possible, but one would have to use 3 5.5x1mm or even 6 5.5x0.5mm in parallel instead.

Does anybody have experience with this?

[Hummina Shadeeba]

i dont know where you got your square wire but i got a small amount from MWS. Its SUPER expensive. it was a huge pain to not have it twist and end up with worse than round wire if the wire used is anything smaller than 16awg and even that size was a pain.


most interesting to me is the square wire starts as round wire that already has the enamel and then they form it into the square shape. i asked about possibly crushing wires into the stator and without my adding any details MWS told me it was "child's play". squishing a wound tooth so as to get more wire in or further down in there away from the magnets sounds doable but a pain. More so the advantage seems in being able to crush the endturns of the stator after already wound so you have more room in the case and can use a longer stator.

[SafeDiscDancing]

Maximizing fill is always a good thing.

I just did a 104% rewind and it clearly shows itself in heat reduction.

Now I just need my "social reality" to permit my ability to enjoy it.

(right now I'm in a lockdown)

[Hummina Shadeeba]

Maximizing fill is always a good thing.

I just did a 104% rewind and it clearly shows itself in heat reduction.

Now I just need my "social reality" to permit my ability to enjoy it.

(right now I'm in a lockdown)

how are you testing heat?

[SafeDiscDancing]

how are you testing heat?

When I went from an 11 turn wind to a 14 turn rewind it dropped the temperature drastically as is defined in the Grin simulation.

Basically I'm saying the Grin simulation is spot on and 100% valid if you put in good data.

Justin knows his shit.

[Hummina Shadeeba]

how are you testing heat?

When I went from an 11 turn wind to a 14 turn rewind it dropped the temperature drastically as is defined in the Grin simulation.

Basically I'm saying the Grin simulation is spot on and 100% valid if you put in good data.

Justin knows his shit.

If ur dropping the kv I think that’s likely much more so going to be the cause of the temp drop and not the 5% copper.

Can u add 5% more copper in the simulator or u altered the resistance of the motor in a way that would be adding 5%?

[SafeDiscDancing]

If ur dropping the kv I think that’s likely much more so going to be the cause of the temp drop and not the 5% copper.

Yes. The motor was designed to be operated at 36 volts and I am running 12S Li-ion which runs from 48 volts to 43 volts.

So the rewind simply adjusts the motor to a more realistic situation.

The lessoned learned is you need to know what you are doing.

But it was done because of an ERPM limitation of Sensorless mode so there is more upside if the controller was designed to handle it.

This setup only pushes the ERPM to 24,000 and I know there is room for that going way higher.

[calab]

The chinese sellers were right all along, motors ARE designed for voltages.

If ur dropping the kv I think that’s likely much more so going to be the cause of the temp drop and not the 5% copper.

Yes. The motor was designed to be operated at 36 volts and I am running 12S Li-ion which runs from 48 volts to 43 volts.

So the rewind simply adjusts the motor to a more realistic situation.

The lessoned learned is you need to know what you are doing.

But it was done because of an ERPM limitation of Sensorless mode so there is more upside if the controller was designed to handle it.

This setup only pushes the ERPM to 24,000 and I know there is room for that going way higher.

[larsb]

No, but i guess you´re trolling a bit there.

[larsb]

i dont know where you got your square wire but i got a small amount from MWS. Its SUPER expensive. it was a huge pain to not have it twist and end up with worse than round wire if the wire used is anything smaller than 16awg and even that size was a pain.


most interesting to me is the square wire starts as round wire that already has the enamel and then they form it into the square shape. i asked about possibly crushing wires into the stator and without my adding any details MWS told me it was "child's play". squishing a wound tooth so as to get more wire in or further down in there away from the magnets sounds doable but a pain. More so the advantage seems in being able to crush the endturns of the stator after already wound so you have more room in the case and can use a longer stator.

got he wires from wires.co.uk. It' relatively expensive compared to round wire. I have to add that i never used square wire as it's known to be a bitch to wind. i've tried 5x2.5mm rectangular, 2x1mm etc, whatever maximises the fill factor on a given stator. For a 12slot 14pole motor the thickness should be <1mm to avoid getting too much skin effect if you plan to use the full rpm potential.

[hias9]

I plan to use it on a direct drive hub motor, so skin effect is not really an issue.
But because of bending radius I doubt that something like 5.5x3mm would make sense if the stator tooth width is only about 7mm. So instead I guess I would have to use 5.5x1mm 3 in parallel or 5.5x0.5mm 6 in parallel.

[larsb]

The radius will be the width of the tooth minus some mm, the wire bundle can be flattened a bit. I had to do it on some winds, not really because the wires were thicker and hard to bend but because the copper fill grew so much and the motor wasn’t spaced for the copper.

A 12s14n motor like the 80100 maxed at about 7000 rpm has a frequency of roughly 800Hz, i estimated ~1kHz when looking at skin effect. What would your hub motor have at your max speed?

[hias9]

130kph would be 1100rpm and with 16 pole pairs it would be less than 300Hz. So I doubt that skin effect is a big issue here.
Bending radius would be less than 4mm in this case. So I doubt that this would be possible with 5.5*3mm wire without damage.
So you think 5.5*1 3 in parallel would be fine or do I even need something like 5.5*0.5 6 in parallel?

[Hummina Shadeeba]

“heavy build” stuff that’s double coated shows better flexibility on a tight radius

Page 9 shows thicker insulation doing better. they sell thicker insulation rectangular wire
https://www.hca.hitachi-cable.com/produ ... ire_en.pdf
You’d be doing a slightly tighter bend than in their test but then again even if the enamel were to crack that crack would be touching wire that wasn’t cracked and you could rely on that other non-cracked wire’s insulation.


Or maybe could add the smallest spot of kapton tape at the bend.

[hias9]

Do you mean page 8 (the page which has two pictures on it) ?
The two wires which are compared there have the same insulation thickness (only slightly different at the edges).

I also found this pdf some weeks ago, but did not find any place where to buy that high reliability Hitachi wire in smaller quantities.

[Hummina Shadeeba]

Yea page 8. It says .03 compared to .04mm insulation thickness.

If anyone knows a place to get reasonably priced rectangular wire please post it. MWS wire is crazy expensive and only place I’ve found.

Or knowing it’s insulated prior to shaping maybe there an easy way to make our own.

[hias9]

It's not a comparison of .03mm vs .04mm
The layer thicknesses are the same in both cases (only the value in parenthesis which is the minimum layer thickness at the corner is slightly different).
It's a comparison of the Hitachi wire against common wire of this size.

I highly doubt that it's insulated before shaping.