High absolute current repeatedly

[leoms]

I use a Flipsky 75100 ESC with a mid-drive motor on a bike. I have the absolute current limit set to 200A based on ESC specs (with "slow ABS current limit" set true, so it's checking filtered current against the limit). In general my setup never exceeds this limit: motor current is limited to +/-60A, battery discharge limited to 80A (at ~80V). But once in a while I do get absolute current above 200A. The odd thing is that this happens repeatedly when it happens. For several seconds after the first fault, almost any acceleration results in the ESC cutting for exceeding the limit. I haven't fully investigated what gets it to stop but it's some combination of waiting a few moments and restarting the ESC. This has always happens at low speeds. It does not seem to result from the high speed or high total power output (i.e. fairly high speed full throttle). It's possible it's correlated with high motor current (low speed full throttle) but often I can gun it at low speed and there's no problem.

Any guesses or debugging tips? Or more info that would help? Thank you so much in advance!

-Leo

 

[Hummina Shadeeba]

I always set the max amps to something realllly high. Eskaters do too so as to avoid any risk of losing power

 

[scianiac]

Well first I would recommend against using slow ABS current, it's a top 10 way to blow up an ESC. You should isolate more when it happens, you didn't say anything about your setup here so I'm just guessing, does it use hall sensors for low speed? do the faults happen above or below the hall sensor transition speed if you do have them? If it's not in hall sensor mode that means the observer is losing the motor position most likely, why it would do it under those odd circumstances I don't know, normally it happens at high current levels especially with low quality hard hardware like a Flipsky.

 

[amberwolf]

I use a Flipsky 75100 ESC with a mid-drive motor on a bike. I have the absolute current limit set to 200A based on ESC specs (with "slow ABS current limit" set true, so it's checking filtered current against the limit). In general my setup never exceeds this limit: motor current is limited to +/-60A, battery discharge limited to 80A (at ~80V). But once in a while I do get absolute current above 200A. The odd thing is that this happens repeatedly when it happens. For several seconds after the first fault, almost any acceleration results in the ESC cutting for exceeding the limit. I haven't fully investigated what gets it to stop but it's some combination of waiting a few moments and restarting the ESC. This has always happens at low speeds. It does not seem to result from the high speed or high total power output (i.e. fairly high speed full throttle). It's possible it's correlated with high motor current (low speed full throttle) but often I can gun it at low speed and there's no problem.

Is that 200A limit a battery current, or a phase current?

If it's a battery current your BMS should trip and power should shutdown; if that's not happening maybe it's actually a sensor fault (or firmware bug) and not a real 200A+ excursion.

If it's a phase current, maybe there is a connection issue between controller and motor in either the phases or the position sensors, so that the controller either can't correctly read the sensors under those conditions just for that moment, or the phase currents are not flowing as it is intending to drive them and results in a current spike on a phase.

Another potential issue, if it's always high current / low speed issues, maybe the position (hall) sensors are reacting to the noisy motor environment (whcih gets worse the more current is flowing)

 

[leoms]

Thanks everyone. Sounds like the most likely cause is noise in the ESC's understanding of the motor state.

amberwolf, to answer your q, the 200A limit is absolute max current (experienced in the ESC, as I understand it). I'm attaching a screenshot of the current limits I'm using.

The motor has no sensor. Maybe this all adds up: no sensor -> ESC has bad understanding of motor state at low speed, plus there's always highest current at low speed -> ESC gets current spikes at low speeds. Sounds like you'd both guess that adding Hall sensors might help?

scianiac, why would using slow ABS current limit be bad for an ESC? My understanding is that high current causes damage through heat, and I'd expect heat to accumulate on a timescale long enough that filtered current would catch issues. Is that not correct?

 

[A-DamW]

why would using slow ABS current limit be bad for an ESC? My understanding is that high current causes damage through heat, and I'd expect heat to accumulate on a timescale long enough that filtered current would catch issues. Is that not correct?

ABS current limit is an absolute maximum, i.e., a fault state, a protection mechanism.
You don't want a slow response to an absolute limit.

Read:
VESC Tuning tips

 

[amberwolf]

amberwolf, to answer your q, the 200A limit is absolute max current (experienced in the ESC, as I understand it).

Max current *where*?

Phases (output to motor), or Battery (input to controller)?

These are very different things, and involve different systems in the controller, and different things in the firmware, and different conditions / hardware / etc that could cause faults.

So it's important to know which limits you are setting affect which parts of the system.

scianiac, why would using slow ABS current limit be bad for an ESC? My understanding is that high current causes damage through heat, and I'd expect heat to accumulate on a timescale long enough that filtered current would catch issues. Is that not correct?

Any absolute limit is a safety cutoff feature. You want as fast a response to that as you can get, normally.

Slow response to whatever sub-limits there are is probably ok, but if you have too slow a response to an absolute limit you can get hardware damage either instantly or cumulatively.

 

[scianiac]

Any absolute limit is a safety cutoff feature. You want as fast a response to that as you can get, normally.

Slow response to whatever sub-limits there are is probably ok, but if you have too slow a response to an absolute limit you can get hardware damage either instantly or cumulatively.

This, it's not the heat that will kill the FETs the temp cutoff is pretty good at preventing that, it's blowing the FETs up by dumping hundreds of amps through them in a massive current spike, or a voltage spike.

With no hall sensors it sounds like yes the controller is losing track of the rotor position due to noise or tuning or a combination. It's possible tuning some settings will fix it and that tuning thread provides good advice on that. Ideally if you can have some data capture that is easy but you want to turn off slow ABS current and then lower the abs current until it trips every time. So like if it trips every time at 150A that's not exactly stable but if it trips at like 80A well that's not that bad if it's set at 60A. The idea that you have the phase current set to 60A and it's fine most of the time and only occasionally tripping a 200A absolute means that for all your know the current is swinging wildly all over the place between 60A and 200A and the PI controller is doing everything it can to keep it under control. I mean it's probably not all of the time, you can feel if a controller is running really unstable because the motor does weird things you can feel and hear. But ideally testing your tune with a much lower absolute current and/or data capture to see how stable it is is a wise idea.