Upgraded phase wires, but lost torque at low speed?

[Jan-Erik-86]

Hi guys,

My bike is a 26" with a 1000W DD hub motor. Specs unknown, but top speed with wheel off the ground and fully charged 48V/13s battery is around 60 km/h.

I originally had a 26A KT square wave controller that i shunt modded to give around 36A, and at the same time i added some additional 1.5mm2 phase wires to the stock 0.75mm2 wires. This worked perfectly fine for a few months until i got my hands on a 36A KT sine wave controller, and did a swap to get a more smooth and silent ride.

The power felt about the same with the new controller and it worked fine, but as i now only had the 3x 1.5mm2 phase wires, and noticed they were getting hot, i decided to replace them with 3x 4mm2 wires, that could also be significantly shorter due to a new location of the controller - but this is where things that confused me started to happen...

Instantly on the first test ride with the new, and shorter 4mm2 wires, i felt significantly more vibrations in the bike when accelerating from low speed. I thought it could just be my imagining, or that the upgraded wires had less loss, causing more motor power, which again naturally caused more vibrations. But after a day of riding i did notice that the pulling force i once had at low speed have been significantly reduced, but comes back at higher speed.

At around 7-20km/h i see the power is being delivered to the motor, but it feels more like when you choke a petrol engine - you push the gas pedal, fuel is entering the engine, but power just isn't there - until you get to around 20+ km/h where it starts to return again. I know below 20 km/h is a low-efficiency range for the motor, but it's a speed i often used before to pull me up short, but steep and curved hills. Before i could by in PAS lvl 2 (~350W) up those hills, while now i feel lvl 3 (~600W) is suddenly needed.

Does anyone have any idea where to start troubleshooting this?

 

[pwd]

It sounds like a poor connection on one of the hall sensor cables. Does it change the behavior if you unplug the motor's hall sensors from the controller?

 

[E-HP]

What kind of connectors are you using for the phase wires between the motor and controller?

 

[Yungyakz]

check solder connections and what conectors are you using. test hall sensors

 

[Jan-Erik-86]

Thanks for the replies guys!

The stock wiring from the motor is still connected, I simply hooked onto the phase wires right outside the motor. This is how I had it previously as well though, with the 1,5mm2 phase wires, and it worked fine.
When I replaced the 1,5mm2 wires with 4mm2 wires, I simply cut off the old wire and left a little bit of it outside the motor, around 10-15cm or so, simply so I wouldn't have to re-solder directly onto the stock 0,75mm2 phase wires that is right next to the sensor wires.

For connectors I use simply crimp spade connectors designed for 4-6mm2 wires. I know this is not ideal and XT150 connectors are on the way - either way, it did work fine previously with the 1,5mm2 wire but git a bit hot on high power for an extended period. It does not get hot now after I replaced the wires.

The way it is now is like this:
Controller board -> 30cm 4mm2 -> spade connector -> ~80cm 4mm2 -> solder connection -> 15cm 1,5mm2 -> solder connection -> 30cm 0,75mm2 until motor winding.

As for the hall sensors - I haven't touched those wires at all when I replaced the phase wires, but I did have a look at the connector and it seemed fine. It's a brand new gold plated connector that I just soldered on a couple of months ago. I'm not sure how to further test the hall sensors? Do I need a special tester/tool for this?
As for disconnecting and testing without hall sensors - I don't think I can do this as my controller is not a sensor-less controller?

 

[dogman dan]

Its your shitty spade connectors, or the soldering you did on one of the phase wires. Sounds like it gets full power on two phase wires, and 50w on the other. It will spin no load, but put a load on it, and it stutters because it's " running on two cylinders" instead of all three.

 

[Jan-Erik-86]

I see, thanks. The new solder joint were already checked and are good, but I will recheck the connectors, and possibly just solder them while i wait for the XT150 to arrive. Will recheck the old solder connection between the 0,75mm2 and 1,5mm2 wire though, and maybe just solder the 4mm2 wires directly while I'm at is if it looks questionable.

Wouldn't they normally get excessively hot if it was a poor connection though? The old spade connectors used to get hot before with the smaller wire, but with the larger connectors for the 4mm2 wires it does not appear to heat up anymore (the remaining part of the stock 0,75mm2 however does heat up quite quickly, but not the soldered part). The watt-meter is reading 1500W+, so the power must be dumped somewhere at least...

 

[Tolkano]

Your phase wires are 1.5m long and you have all these spade terminals connecting it together? Wtf is going on

 

[Jan-Erik-86]

Is that really so bad?

My measurements were a bit off, it's closer to 125cm long from PCB to motor winding. There's only one spade connector the entire way, and it's a larger type, but either way that's only temporary until the XT150 arrive.

Either way, while i was inspecting all the connectors for poor connection, i accidentally damaged the LCD+throttle connector, so it seems it will take some time before i get it all up again to do more testing - but at least all the other connectors looked good.

 

[dogman dan]

Hmm, usually bad contacts look good, but do eventually get hot and reveal themselves. Once I had a rubbed wire that looked fine, but on the back side the tire had rubbed all but one strand of the wire away.

Keep looking, and looking again. It could have been just the throttle wires, half voltage there would be similar. Try another throttle would be my next thing to try.

 

[Drunkskunk]

Sounds like a bad connection, and I would also suspect spade connectors.
But it could be other things. the Symptoms are of an intermittent or high resistance connection. You're getting voltage across, but failing to get enough current across.

You might check for a bad crimp, but it could also a cold solder joint. Cold solder joints can look normal and feel snug, unless you're inspecting real carefully. But they won't conduct well.

It can also just be a bad section of wire. A few years back I bought some high flex silicone shielded 8 gauge to make a battery cable. I spent weeks trying to figure out why my bike would shut down as soon as I turned the throttle. Turns out one section of the wire had some kind of alloy problem, and had huge resistance.

 

[Jan-Erik-86]

So after several hours work I've redone the LCD+throttle connectors, and used thicker wires time so they won't fail again. 24awg wire is space efficient for signaling wires, but man are they fragile...
Also inspected the hall connectors more carefully, and indeed there was a bad connection. Not in the connector, but in the wire itself right behind the connector. These were the same thin 24awg wires. Replaced those as well with thicker ones.

Will be able to test it more tomorrow, but just did a quick 5 min test ride, and it appears to have solved the issue.
It pulls more evenly again, less vibrations, and it had no problem pulling me up a hill from near stand-still.
Spade connectors on phase wires still in place, but they do not run hot and i did measure the current of each phase wire to be close to identical during load.

 

[dogman dan]

It was running on two phases, mostly. Usually more stutter though, when that happens. And occasional failure to start from a full stop.

 

[Tolkano]

Use Anderson's, they're basically the only choice if you don't want to solder