Re-winding a hubmotor

Farfle/Zoombies, how are you measuring the rshunt value?
My measurement are always off... Either with a clamp DMM or a RC Charger the measurement is never spot on.
 
liveforphysics said:
You should be a "proud papa" Farfle.

Thats a badass machine, and you made it yourself. Nothing you can buy can match that level of satisfaction. Savor it my friend.

I'm just spit-balling here.....Farfles build is epic....but I thought you didn't like Schwing-Arms which lacked triangulation???
dunno.gif


I am close now to doing what Farfle did.....except i want to press SNF cartidge ball-bearings into the bottom bracket and use the inherent strength of that part of the bike as the swivel point.

@ Farfle: love the jacket around the big triangle, looks very Moto-Retro.
 
Farfle said:
The controller is a steveo built 24 feet with 8 shunts.
Yeh you should run a known current through the shunt and verify it because 4 shunts is ~.2540 mOhm from what I saw with the last one I did. Unless you already did this.
I have had people argue about this with me and sying I had my shunt off because of my avitar ;). Edit post corrected
 
That's what I am doing tomorrow, at the ebike shop there is a load bank with a calibrated shunt and CA. Ill just run a 36v battery thru it and measure the amperage and drop across it simultaneously.
 
Arlo1 said:
Farfle said:
The controller is a steveo built 24 feet with 8 shunts.
Yeh you should run a known current through the shunt and verify it because 4 shunts is ~2000 mOhm from what I saw with the last one I did. Unless you already did this.
I have had people argue about this with me and sying I had my shunt off because of my avitar ;)


2000mOhm would be 2Ohm's bro, and at 100v it would be impossible to pass more than 50amps. :)

I would think you meant 2000uOhm, but that's still way too high. Perhaps 200uOhm? (which would be semi-reasonable) I've got mine at 86uOhm at the moment. It makes low current resolution poor, but makes it possible to read >200-300amp current levels effectively.


Also, for all you guys who use anything but shunt alloy metal on the shunts, you are not getting correct readings when temp increases happen on your shunt (and huge temps happen on the shunt). If you took a 500uOhm shunt, and soldered it down to a 200uOhm shunt (as is pretty common), and then calibrate your shunt to be correct at room temperature, when you're actually riding, the difference in resistance (and hence your current reading on your CA) will be massively influenced by a soldered shunt's temp.

At 25degC, typical 60/40 solder is about 13x10^-8 ohm-m, at 120degC (and a shunt easily reaches this temp), the solder is about 24x10^8. Meaning, your 500uOhm shunt soldered down to 200uOhm and calibrated will be reading about 40% higher than reality, even if it was calibrated perfectly at room temp.

This is why folks use shunt metal for shunts. It's whole special attribute is being an alloy that holds consistent resistance with temp change.
 
Arlo1 said:
Farfle said:
The controller is a steveo built 24 feet with 8 shunts.
Yeh you should run a known current through the shunt and verify it because 4 shunts is ~2000 mOhm from what I saw with the last one I did. Unless you already did this.
I have had people argue about this with me and sying I had my shunt off because of my avitar ;)

****darn it, got tree'd by Luke on the shunt value****

Might wanna check your math there Arlo :D 2000mOhm is a 2 ohm shunt, I believe you meant 2 mOhm. All my Lyen controllers have shunts that seem to be somewhere a little lower than 4 mOhm each. My 18 FET which uses 3 shunts measured in right around 1.25 mOhm. 12 FET version measured in around 2mOhm.

Calibration was done through multiple discharge / charge cycles then adjusting the value based on the difference using basic algebra. I use an iCharger206B so it's only as accurate as that, but when charging batteries at low C rates the charge is typically over 98% efficient and I charge at really low C rates like 0.2C my Cycle analyst always is in line with the ebikes.ca simulator for motors I have on it and also the ebikecalc17 simulator so I trust my shunt values.

****edit****
good point on the solder Luke. I added a piece of 12 gauge copper in parallel to my shunt to lower to to around 0.5xx mOhm and my CA puts me in the ball park, but I'm just using it as a gauge of when stop beating on it. I know copper isn't great due to it heating and my internal case temps reach over 60C when I'm being on it so I'm sure that copper is skewing the readings a bit. Good news is I can remove it in a few mins if I ever want to because of how I installed it. My numbers are also consistent on discharged AH vs amount charged back into the pack for sane riding. When I ride hard there is of course some error, but at that point I'm also burning watts into the wires and connectors because they get warm so that skews the reading as well.
 
Thanks so much for posting that about shunts LFP. I was wondering why the hell my readings were so off.

So I guess I just go to mouser, find shunts with the value and size I want, and drop them in the board?
 
Ok i have it writen down i think at my shop and i will look when i get home.
 
zombiess said:
I added a piece of 12 gauge copper in parallel to my shunt to lower to to around 0.5xx mOhm and my CA puts me in the ball park, but I'm just using it as a gauge of when stop beating on it.


2cm of 12awg is 0.0001ohm, or 0.1mOhm, 100uOhm.

3cm of 12awg is 0.00015ohm, or 0.15mOhm, 150uOHm.

Be very careful with copper... on a cold day, your stuff get's hit harder.
 
liveforphysics said:
zombiess said:
I added a piece of 12 gauge copper in parallel to my shunt to lower to to around 0.5xx mOhm and my CA puts me in the ball park, but I'm just using it as a gauge of when stop beating on it.


2cm of 12awg is 0.0001ohm, or 0.1mOhm, 100uOhm.

3cm of 12awg is 0.00015ohm, or 0.15mOhm, 150uOHm.

Be very careful with copper... on a cold day, your stuff get's hit harder.

I've already noticed that when everything is cooler it hits harder. At first I thought it was the motor warming up like my 9C use to, but now that you mentioned this I'm wondering if it's just the copper I added to the shunt since I don't think I've ever managed to get my motor to even 100C yet. On long mild rides where the controllers internal ambient temp stays 45C or below the amount of kick and throttle response I get from the bike is much more than when it's over 60C and it stays that way for the entire ride. I knew it wasn't the greatest move to put copper in parallel, but I did it as a test for trouble shooting another issue several months ago and haven't had the controller off the bike since then. With the warning you just gave me I think it's time to remove the controller from the bike and pull the copper shunt off. I have some proper shunts I can parallel with it If I want to stay in the high shunt range and reduce the heat, but I'm pretty happy riding at 85A if I feel the need for speed. The stock shunts will dissipate 9W at 1.25 mOhm but it's only for short bursts.

Maybe I'll try some hard rides with the covers off the controller for airflow and see if it makes much of a difference in throttle response. Controller ambient temp will be lower and I'll be able to measure it so I'll then be able to tell if it's the motor warming up or the shunt warming up that's causing the change in response. Simple test to try.
 
Ok, we have a 20 inch DX32 at the shop, so ill just run that and an ml-75. Going to move the shock mounts down to keep the geometry where it is height wise.
 
Farfle said:
Ok, we have a 20 inch DX32 at the shop, so ill just run that and an ml-75. Going to move the shock mounts down to keep the geometry where it is height wise.

Awesome, 20" wheels is going to rock!
 
Ok so the last controller with normal nichrome shunt material was .2540 mohm i corrected my post it was on a 36 fet i was running!
 
Ok, Waiting for the next order of CAs to come into the bike shop so that we can get our calibrated load tester running, but here is my plan to calibrate the shunt: Ill run a 48v pack into the load testers CA thru the shunt on the controller. Ill record the voltage drop across the shunt, and the amperage on the load testers CA and the that should give me the V=IR values that I need, correct?
 
Grrrr, on the way home, It started to rain about halfway there so I hammered it all the way home, what I didn't realize is that it takes a phenominal amount af power to go 68 mph, the motor was drawing 140 amps for the last half of my route (10 minutes) and as I slowed down into my driveway, the motor shuddered and bucked backwards, I pushed it the rest of the way up my driveway, diagnosis is that I cooked the yellow hall sensor. The phases are all fine, but I haven't popped the cover yet .when this robotics tournament is over I'm going to ventilate the covers, and replace the halls, any way that you can protect them, Optical sensors?
 
Before you ventilate the cover, have you thought about filling it about 1/3 full with the thinnest mineral oil (Think Dextron III ATF type) or silicon oil you can find? Since you can machine your axle, you could provide a nice vent through the solid side of the axle. I think the oil will stop this heat soak fratricide of halls dead in it's tracks.
 
What about running double hall sensors. I believe Lowracer did this on one of the 540X motors had has. I think he basically put 3 more sensors on the other side of the magnets (in parallel?) so that if one sensor failed, the backup would still be working.
 
That's what the second set of sidecovers is for, i'm hoping that a new set of sensors, the ventilation and moving into a 20" wheel will solve it, but if that can't get me the performance levels I want, I have a gallon of in-ox mx3 ill use to do an oil cooling setup.
 
Looks good cuz ya don't want a baked pie
 
Moving all updates to the motor to the race bike build thread here:

http://endless-sphere.com/forums/viewtopic.php?f=6&t=34880
 
Farfle said:
ok, here we go,

16 strands of 23ga is 4.01 mm^2
one strand of 11 AWG is 4.17 mm^2
two strands of 13 AWG is 5.26 mm^2 :D (probably wont fit :( )
two strands of 14 AWG is 4.16 mm^2 :)
three strands of 16 AWG is 3.93 mm^2 :(

im thinking two strands of 14 gauge for being the easiest to pull, and still get a good copper fill, but ill order dome 13guage first, and see if i can get it to fit :D


what's the objective on deciding what gauge wire to use ?
the more copper fill, ok, so you got here 14g and 11g giving same copper fill and i am sure smaller gauges at some points will give also same copper fill, but what gauge should we aim for, lets say not 11g because the insulation might crack, 14g wont crack, 18-28g wont crack either, and assuming some wires between 18-28g will give same copper fill, then how do we choose what to use :?

you would've ordered more than 3 sensors LOL
 
Back
Top