Tesla LDU Bench Test: turns slowly, cogs, but has good Hall signal

[jangell]

I've been trying to get an LDU to work for some time now. I had some issues with my wiring to the Hall sensor, but I've gotten that fixed and am properly reading a speed signal when I rotate the motor by hand.

Unfortunately, I'm still having trouble performing a low-voltage test. Here's a video demonstrating the problem, as well as the tests I performed:



I followed Damien's video (https://www.youtube.com/watch?v=ZvlzGm709zg) about how to test the diodes. Everything seems fine there.

I manually rotated the motor and am getting a good speed signal in the OpenInverter web interface.

But my low-voltage test (using the settings in Damien's video, but with two 12v car batteries like in this Boxter EV video: https://youtu.be/odFYeZ2goPs?si=FwcYoXdzVp6s7WUv) runs very slowly, and it cogs every few degrees. Boxter EV's video shows that the motor turns at a pretty decent rate, but mine is turning much slower. The speed graph shows that the Hall sensor still working, though (I forgot to show that in the video).

I should note that this is the exact same result I saw when the Hall sensor wasn't working, so either the problem is not at all related to the Hall sensor, or something is still wrong with the Hall sensor.

I'm wondering if there's something wrong with my inverter. You may notice the lack of an electric stove coil in my circuit, which I'd overlooked in my earlier tests. At this point I figure that whatever damage I've done is already done. I'm only running at 24v, but I don't know if that's enough to damage it or not.

Is there anything else I can test or try? Or should I start looking for a new inverter?

Thanks!

-- Joe

[ekohn04]

I assume you've already read the threads about switching your A/B wires on the 4 pin encoder connector. I can't recall if that would affect the hand-turn test you're doing or not. Are your encoder wires shielded? I was bench testing with a brand new harness that did not have the cables shielded for the encoder which I thought would be ok for bench testing, but was experiencing the cogging that indicates an encoder issue. I ended up using a Tesla LDU harness with the encoder cables shielded and that fixed my issue.

[jangell]

Yup, I've seen the "swap the A/B wires" threads. I did that a few times trying to track down the complete lack of signal for my original hand-built harness. That was more out of desperation, though, since I had no signal at all. I was using a shielded cable there, made from a butchered shielded ethernet cable.

I'm now using a factory harness from a Tesla, which got me an actual speed signal. The entire part of the harness from the Hall sensor to the 23-pin Ampseal connector is original, including the shielded Hall sensor cable.

All that said, I did not try swapping the pins on this factory harness. I assumed it was good, but who knows if was paired with a motor that also had swapped pins on its Hall sensor.

I don't show it in the video (mostly because the iPad isn't in frame), but the cogging is happening at such low speeds that it hasn't registered yet. It eventually turns enough to show a spike, which seems to be about 10 degrees of rotation (I'm just eyeballing it). I don't know many pulses per rotation the Hall sensor detects, so I'm not sure if that's right or not.


I also compared my video to Damien's at 41:24 where he rotates the motor. This should go right to that part of the video:

It looks like we're getting similar readings. I'm turning mine a bit faster than he is, I think, so I'm getting more consistent speeds with fewer gaps.


I don't suppose there's a way to plot the encoder A and B pins so I can verify that both are working? I'm also wondering if the best way to try swapping the pins on the Tesla harness is to disassemble the connector, or just cut the wires and swap them.

Thanks

-- Joe

[johu]

I assumed it was good, but who knows if was paired with a motor that also had swapped pins on its Hall sensor.

The OI software seems to spin the field in the opposite direction so it always needs AB swapped from factory harnesses

I don't know many pulses per rotation the Hall sensor detects, so I'm not sure if that's right or not.

36 pulses per channel (numimp=36) but 4 edges, so resolution is 2.5°.

I don't suppose there's a way to plot the encoder A and B pins so I can verify that both are working?

No...

[ImportFanatiK]

I am having this exact phenomenon with my sport LDU. I just made my first attempt to spin it about an hour ago and ran into this.
The motion and speed on my LDU shaft looks identical to yours. moves very slowly and very jerky / pulsing.
I am running 2 car batteries (25v - same as you).
I am also using a shielded cable for the encoder signal.
I did the manual spin test from Damien's video (as you did) and the encoder appeared OK.
I am also watching those same two youtube videos you posted as a reference.
I just posted in my thread below:
viewtopic.php?t=4418

Glad to know I am not alone with this situation.
I will swap my A/B wires next and see if that helps.

[jangell]

I had a chance to work on the motor this morning. I swapped the wires and it worked pretty much perfectly:



It cogs a little at the very start, but I'm putting that down to it not having anything attached to the shafts and that I was very slowly increasing the power.

This didn't work immediately. I swapped the encoder wires, and then got no signal at all, even when turning the motor it by hand. That was pretty frustrating. I wound up taking the casing off the inverter, pushing all the connectors together a bit, and trying again, and it finally worked. I'm not sure where the loose connection, and I imagine that will be a pain to hunt down properly.

In any case, I clearly should not have assumed that the "swap the wires" advice didn't apply to factory harnesses. Thanks for clearing that up!

-- Joe

[P.S.Mangelsdorf]

I had a chance to work on the motor this morning. I swapped the wires and it worked pretty much perfectly:

Congrats!

It cogs a little at the very start, but I'm putting that down to it not having anything attached to the shafts and that I was very slowly increasing the power.

Yeah that's something that can be sorted out in the tune with full pack voltage. Check out some of the LDU tuning threads and parameter guides, especially JonVolk's tuning information

I wound up taking the casing off the inverter, pushing all the connectors together a bit, and trying again, and it finally worked. I'm not sure where the loose connection, and I imagine that will be a pain to hunt down properly.

Odds are very good that it was the connection from the inverter harness to the external header. That connection has very little slack, the connectors don't have a good hold, and is easy to bump loose.

[ImportFanatiK]

FWIW - I swapped my A/B today and it fixed mine too.