DIY 10awg phase leads through axle in 9C hubmotor.

[AussieJester]

Maxxis Hookworms with slime filled inner tubes are also good choice if we are talking tires...

KiM

 

[methods]

I go with anything that is a "downhill" tire in a 60 compound
They have sidewalls that are amazingly stiff
Kevlar belted and heavy as hell
Widest I can find

I run them with thick downhill tubes and no slime

I ride in the cement arroyo where the shit-heads break bottles every day after school.
Never had a flat on that combo.

-methods

 

[Hyena]

I think that the mere mortal folks around here could do well by running 12AWG through the axle using your methods

Exactly what I was thinking!
Damn Luke, everything about this is overkill - 4ga power wires, running a casual 72v for your parents and I bet you're setting up the controller up with a modest 80amp current limit too

Still, I bet your mother will enjoy riding it to church on Sundays :lol:

I'm building my old man an ebike for his 60th and thought even 12S lipo was plenty of power for him (and he's ridden motor bikes all his life)
I offered to build a mobility scooter for my 88 year old grandma but he said what ever voltage I run it on he's going to come and remove half the batteries. :lol:

 

[dequinox]

I have the marathon plus on my rear wheel, and even though I've busted about 9 spokes, I've gone 500 miles without a flat. They truly are a remarkable tire, and totally worth the price.

 

[dogman dan]

Back to the wires issue, the big diameter bearing is the design improvement all motors need. Custom covers so we can mod motors would be nice!

It not only allows any size wire to go through without going though the axle, it allows the wires to enter the motor without getting anywhere near the dropouts and getting cut. the large axle size at the bearing also allows an integrated tourqe arm, so the axle can be normal size and round. Amazing how nuts don't strip when 50% of the threads haven't been machined away to make a flat sided axle. This is how heinzmann did it, no doubt patented, but there has to be a way to make a new design different enough to get around that.

 

[liveforphysics]

Back to the wires issue, the big diameter bearing is the design improvement all motors need. Custom covers so we can mod motors would be nice!

It not only allows any size wire to go through without going though the axle, it allows the wires to enter the motor without getting anywhere near the dropouts and getting cut. the large axle size at the bearing also allows an integrated tourqe arm, so the axle can be normal size and round. Amazing how nuts don't strip when 50% of the threads haven't been machined away to make a flat sided axle. This is how heinzmann did it, no doubt patented, but there has to be a way to make a new design different enough to get around that.

Yep, running bigger bearings and a spacer around the axle with holes in it is sure a better way to run phase leads. For a 9C, it would require a custom side plate to be made.

There are a ton of 9C motors in the world, but since you can slip 10awg through them with the method I showed here, I don't think anyone would feel the need to spend more effort to get bigger phase leads than 10awg.

 

[El_Steak]

Wow, this is exacly the info I was looking for! Thanks.

Were you able to remove the cover with a knife/screw driver to pry it off or did you use a gear puller or other special tool?

 

[Chalo]

I'm curious as to whether you've done a calculation to reckon how much resistance/power you save in this case. It's true that the 9C motor leads are pretty small, but they're also pretty short. We accept small but short conductive paths at connectors and in PCB traces without taking heroic measures to enlarge them (usually).

In my former life as a rocket surgeon, I used milspec teflon-jacketed wire that packed a whole lot of conductor into a small total diameter. Teflon-covered wire would be easier than rubber-covered wire to pull or push through a crowded hollow axle. The silver-plated strands in that wire were superb at taking solder.

Such wire is usually pretty expensive, but this supplier has it for cheap:
http://www.surplussales.com/wire-cable/Wire4.html

Chalo

 

[Chalo]

BTW--

There is a similar alternative to the Schwalbe Marathon Plus that appears to have more specific protection against pinch flats. The Michelin Pilot City (be precise when looking for it, as Michelin makes other tires called "Pilot" and "City") has the sub-tread urethane belt of the Schwalbe Marathon Plus along with smaller belts of the same stuff in the sidewalls, to cushion the rim in case of bottoming. I've been able to find it for lower prices than the M+. Note that the Michelin tire has a strange oval cross section, such that its height off the rim is much greater than its width would normally suggest. The Schwalbe has a more traditional approximately round cross-section. That should only matter if you have a tire clearance limitation, such as a 700c wheeled bike with limited room under the caliper brakes.

http://www.outsideoutfitters.com/ps-1145-1084-michelin-pilot-city-26-tire.aspx

Both my wife and I have bikes with the Schwalbe Marathon Plus, and they have accumulated plenty of mileage with no flats to date (knocking wood now). But the Michelin Pilot City on the back of my wife's Dutch bike seems to be turning in an equally good performance, with no flats yet and no noticeable speed penalty compared to an unarmored tire of the same size.

 

[TMaster]

I've always wondered about this situation with a brushless motor. There is a difference because you are really sending the motor an alternating current instead of direct. I'm not sure how the current/amps works in this situation, but there are 3 wires to spread the load instead of 2. The smaller gauge of wire may go futher in the amperage capability in this situation. I just don't know, maybe someone does?

 

[Takemehome]

Wow, this is exacly the info I was looking for! Thanks.

Were you able to remove the cover with a knife/screw driver to pry it off or did you use a gear puller or other special tool?

No special tools required. I just pry it with a screw driver equally all around until it pops up.
To close it back, use a piece of wood and tap on it with a hammer.

 

[liveforphysics]

Wow, this is exacly the info I was looking for! Thanks.

Were you able to remove the cover with a knife/screw driver to pry it off or did you use a gear puller or other special tool?

I took the screws out of the side with the wires, then bonked the wheel against the axle on the other side against a block of wood. Second bonk and it lifted the cover, which I was then able to get some prying tools under and worked it off.

I'm curious as to whether you've done a calculation to reckon how much resistance/power you save in this case. It's true that the 9C motor leads are pretty small, but they're also pretty short. We accept small but short conductive paths at connectors and in PCB traces without taking heroic measures to enlarge them (usually).

In my former life as a rocket surgeon, I used milspec teflon-jacketed wire that packed a whole lot of conductor into a small total diameter. Teflon-covered wire would be easier than rubber-covered wire to pull or push through a crowded hollow axle. The silver-plated strands in that wire were superb at taking solder.

Such wire is usually pretty expensive, but this supplier has it for cheap:
http://www.surplussales.com/wire-cable/Wire4.html

Chalo

Just worked out the numbers. If the factory wires are 16awg, and you just plug them in from the supplied pigtail, then at 25amp battery current with 50% duty cycle, phase current would be 50amps, and V-drop across the wires would be 2.53v. This is only 126watts, but it's 126w being wasted as heat that I would rather be putting towards moving the bike rather than heating it. Of course for larger input currents it will be dropping more voltage and doing more heating.

I mainly did this as a reliability thing rather than performance thing, as I've heard so many folks say the wires get blisteringly hot and melt/short under high current loads.

That teflon wire is nice. I don't think you would fit 3 passes of 10awg through though.

I've always wondered about this situation with a brushless motor. There is a difference because you are really sending the motor an alternating current instead of direct. I'm not sure how the current/amps works in this situation, but there are 3 wires to spread the load instead of 2. The smaller gauge of wire may go futher in the amperage capability in this situation. I just don't know, maybe someone does?

The motor only uses 2 wires of the 3 at a time. Of the two wires being used, it's sending current in through one end, and back out through the other, so the resistance is twice the value of the single conductors resistance value. For a simplified example, if you have 3 wires going to the axle each having 10mOhms of resistance, then at any point during the motors operation, the voltage drop across the wires would be equal to a 20mOhm resistor being in-line.

The way the current works on the phase leads is pretty neato. The motor windings work as an inductor to filter the DC PWM signal we send so that battery current gets multiplied as the voltage is dropped. If you pull X amount of amps from the battery at Y voltage, you've got XY watts of power. When the controller sends z PWM% to an inductor, it ends up cutting the voltage by Z%. Input/output power need to be the same (minus 1-2% from controller inefficiency), so to keep power in matching power out at a lower voltage, phase current has to be multiplied. To find phase current, take y*z% to get the phase voltage due to PWM, then stick a variable to repesent phase current next to it, and set it equal to XY (watts) and solve. That's at least a rough explaination of how it works...

 

[TMaster]

Very Nice! I went and checked and the motor wires are just about 16 AWG. The controler wires are about 12 AWG though. I guess they couldn't fit them in the axle hole haha!

Does anyone use the e-brake's? I was acutally thinking about doing this with some 12 Awg wire. I don't have my setup running yet, but Im already thinking about mods. I guess you would need the motor halt wires if you were using a LVC or the e brake.

 

[Chalo]

I'm curious as to whether you've done a calculation to reckon how much resistance/power you save in this case. It's true that the 9C motor leads are pretty small, but they're also pretty short.

Just worked out the numbers. If the factory wires are 16awg, and you just plug them in from the supplied pigtail, then at 25amp battery current with 50% duty cycle, phase current would be 50amps, and V-drop across the wires would be 2.53v. This is only 126watts, but it's 126w being wasted as heat that I would rather be putting towards moving the bike rather than heating it. Of course for larger input currents it will be dropping more voltage and doing more heating.

I mainly did this as a reliability thing rather than performance thing, as I've heard so many folks say the wires get blisteringly hot and melt/short under high current loads.

Just using P = I^2 * R as a guide, and .004 ohms/foot as a conventional value for 16 awg wire, I get 50A^2 * .004 ohm/ft = 10W per foot at 100% duty cycle. 50% duty cycle on three one-foot-long leads equals 15W. That's annoying, but it's an order of magnitude less than what you came up with. Where's the discrepancy coming from?

If folks are burning their motor leads, I'd look at connectors and crimps first.

Chalo

 

[liveforphysics]

I'm curious as to whether you've done a calculation to reckon how much resistance/power you save in this case. It's true that the 9C motor leads are pretty small, but they're also pretty short.

Just worked out the numbers. If the factory wires are 16awg, and you just plug them in from the supplied pigtail, then at 25amp battery current with 50% duty cycle, phase current would be 50amps, and V-drop across the wires would be 2.53v. This is only 126watts, but it's 126w being wasted as heat that I would rather be putting towards moving the bike rather than heating it. Of course for larger input currents it will be dropping more voltage and doing more heating.

I mainly did this as a reliability thing rather than performance thing, as I've heard so many folks say the wires get blisteringly hot and melt/short under high current loads.

Just using P = I^2 * R as a guide, and .004 ohms/foot as a conventional value for 16 awg wire, I get 50A^2 * .004 ohm/ft = 10W per foot at 100% duty cycle. 50% duty cycle on three one-foot-long leads equals 15W. That's annoying, but it's an order of magnitude less than what you came up with. Where's the discrepancy coming from?

If folks are burning their motor leads, I'd look at connectors and crimps first.

Chalo

Oops! Thank you for the catch Chalo. I made some mistakes somewhere!

If you just plug it in as it comes, the pigtail is ~3ft, which means 6ft of wire in the circuit. Using a voltage drop calculator for 6ft of 16awg at 50amps, it gave 2.53v of drop, but trying a different V-drop calculator gives me 1.5v of drop (maybe they try to factor in connector resistance or something???)... And working it out manually based on 4.016mOhm/foot gives me 1.248v dropped at 50amps, which would only be 62.4w total heating.... That doesn't seem like much, but through much real-world testing we know people heat them up to insulation melting temps in just a few minutes of hard loading. This is the whole problem i wanted to solve, and I think I've managed to accomplish that nicely.

 

[ZapPat]

I just finished some 9C hub motor upgrades myself... but not just because I wanted to I must admit. I had a hall sensor fail a couple days ago, so had to open the motor up to replace it. Once it was open, I could see that I was pushing it *very* close to it's limits regularly, and this time it was too much for it to take. The coils have a dark color to them and the hall protection plastic film strip had melted into the coils. The smell as the hub was opened was really gross, and it made my lab smell bad for hours. This hub was on the back of my 2WD 20" ebike FYI.

Once inside I could see that the stock teflon phase wires had started melting to the point of fusing together as they entered the axle from the inside. Well I guess it was time to upgrade them while I'm changing the hall sensor. I decided to try Method's suggesting and double up the stock phase wires. They are 1.5mm^2 each, so a little bigger than 16AWG, but smaller than 14. Doubled they are almost equivalent to 12AWG. With just a little persuasion I was able to fit the 6 phase wires plus the 5 hall wires through the axle.

I then decided there was enough room left to pass two more small teflon wires for a temperature sensor. The slippery teflon helped much since things were getting pretty cramped in the axle. I installed the small 10K NTC chip thermistor close to the windings, but touching the laminations only. I'll now be able to use the eagle tree data logger to know what's happening in there. Too bad it wasn't there before...

OK - So now I have a 9C hub that will burn up it's coils long before it's phase leads start melting. That doesn't sound great, because I had already managed to almost destroy it with the stock leads. Hummm... the solution is to make some holes in those covers! 48 holes later I was ready to reassemble the hub. No active cooling for now, but I don't think I need it right now because I'm using this thing in the snow right now. So this now is a snow-cooled hub!!

We have about a foot of pretty dense snow here now, so time for a test. I headed out right into the field behind my place, plowing throught the snow as I went along. The bike managed to go through it, but sometimes slowed quite a bit because of deeper snow drifts. Anyways, I did a bit more than a thousand feet like this before heading back home. When I got back the hub was steeming hot, but still working well. I charged the bike back up and noted that I had used about 7Ah from my 3p12s Lipo pack - so about 45V * 7Ah = 315Wh. So for under 1km I used 315Wh --> that's over 300Wh/km in about 5 minutes of riding!!! No wonder it was steeming after 5 minutes of this! Outside temp was about -7oC BTW.

The mods seem to have worked well, but only time will tell just how well. I'll have to install the logger to monitor the motor's temperature soon... I just hope that eagle tree made the logger compatible with low temperatures. I'll also be looking into buying some moisture protection spray for inside the hub - I'll have to read up that 'improved weatherproofing' thread for some ideas.

 

[Hyena]

So for under 1km I used 315Wh --> that's over 300Wh/km in about 5 minutes of riding!!! No wonder it was steeming after 5 minutes of this! Outside temp was about -7oC BTW.

Yikes, no wonder things are getting toasty in there! Good thing the temp is below 0, it sounds like it's the only thing that's been keeping your wires from completely melting.

Luke, is the outrunner style heatshrink you used better than the average stuff from say hobbyking? I've got plenty of the hobbyking stuff here and thought I'd have a go at rewiring my GM hub. I'm not pumping huge current (~50a) so 12ga would suffice and it should make the fit easier.

 

[liveforphysics]

So for under 1km I used 315Wh --> that's over 300Wh/km in about 5 minutes of riding!!! No wonder it was steeming after 5 minutes of this! Outside temp was about -7oC BTW.

Yikes, no wonder things are getting toasty in there! Good thing the temp is below 0, it sounds like it's the only thing that's been keeping your wires from completely melting.

Luke, is the outrunner style heatshrink you used better than the average stuff from say hobbyking? I've got plenty of the hobbyking stuff here and thought I'd have a go at rewiring my GM hub. I'm not pumping huge current (~50a) so 12ga would suffice and it should make the fit easier.

The hobby king stuff is pretty damn strong once it's shrunk down. I think it would work just fine. Keeping the wires from chaffing from some method to immobilize them (like glue) would be a good idea to avoid a potential insulation issue later down the road.

 

[lifepo4ever]

so for my 72 volt set up with ping battery there no enougth amps to toast my 9c motor ? is that right

 

[El_Steak]

12 AWG is relatively easy to do if its enough for you. Its a bit more than twice the size of the original wires so I guess it should be good for 80-100amps in short bursts.

I bought the Teflon wire from Powerwerx. It was expensive as I had to buy a minimum of 25 feet and have it shipped to Canada. I also bought some small gauge Teflon wire from them and used it for a temp sensor inside the hub.

If anybody plans on doing the same mod and needs some Teflon 12AWG wire, I can stuff 4 feet of it in a padded envelope and ship it for cheap.

*EDIT* no more 12AWG available

 

[TMaster]

Nice! I'll be doing the same tonight. I paid bookoo dollars for the teflon wire though :/ . From your pics it looks like theres plenty of room for the wires!!! I hope to put 1 layer of heat shrink over the entire bundle. How long did you make the phase wires? The wire is not that flexable, I'm debating on how long.

 

[El_Steak]

In my case the shrink wrap doesn't go all the way down. I tried but it was too tight. I only had pretty thick shrink wrap though.

My phase wires are about 12 inches long. As you say, they are not very flexible, so you need enough to make a nice curved "bend" at the exit.

It took me an hour or two of fiddling before I got it. Here's how I did it (might be a better way):

1- Removed all the black silicone at the base of the axle
2- Cut the connectors off all wires (Anderson, Molex)
3- Carefully pulled the small hall wires out of the axle from the inside to avoid breaking any hall-related connection inside the hub.
4- Carefully pulled the black rubber sleeve out of the axle
5- Unsoldered and removed the 3 phase wires (remember where each color goes! - take a picture!)
6- Pushed 2 big teflon phase wires all the way down the axle and inside the hub.
7- Pushed 2 small teflon wires all the way through (for temp sensor)
8- Pushed the hall wires from the inside to the outside.
9- Pushed the last phase wire. Thats the toughest part.
10- Reconnect everything and apply high-temp silicone generously.

I tried doing the 3 phase wires together at first, but couldn't pull the hall wires through the axle afterwards. They are too flexible and "bundle up" in the axle.

 

[TMaster]

Cool, well the 40W Iron i have is not doing the job. The copper wires are sucking the heat right up. I'm going to bust out the 80W iron.

edit: Ok I managed to unsolder them

 

[TMaster]

Took me like 5 hours to do the whole job. I took my time meticulously though. I used Black Teflon coated wire. I heat shrunk the halt wires sperately. Everything fit well. I belt sanded the ends of the wires domed so they would slide down easier. I was not able to use heat shrink around the bundle of wires inside the axle. The teflon is pretty dang tough though. I went a little overkill on the heat shrinking of everything else. I used 4mm bullet connectors on the ends and soldered in about 6 inches of silcone 12AWG wire at the end for flexability to the controller. I took a bunch of pics of the process. Instead of posting that many here is a zip file with all of them.

 

[El_Steak]

Nice Job !

Its odd the the inside of the motor (stator) is different than mine / LFP / Methods.