Expert diagnosis required - Hall faults on Sabvoton 72150 / Crystalite shuddering

[Eddie.K]

Patient history: older Sabvoton 72150, Crystalite MKII, 92V, Cycle analyst (10 KW system). It's always had a hall issue where it'd fault if you accelerated hard for more than a few seconds, or if you dumped the throttle off too fast instead of ramping it down gradually. Then few years later, if you go over 60km/hr the power would flutter off and on then fault. All issues i learnt to live with.

Recent Events & Symptoms: 1st January - it'd just fault as soon as you apply any throttle. I replaced all 3 halls (using ss441A unipolar sensors, I read plenty of discussions on this so hoping I got that right!), reassembled, then motor just buzzes/humms, won't turn! Halls all return a signal in sequence as wheel rotates, 3.08v, 3.08v, 4.85v, again I read lots of posts on this and appears to be "normal" on sabvotons). Tested motor - same resistance, nothing down to earth, cogging the same when shorting out 2 phases. Perhaps I mixed up sensor signal wires or sensor / positions, so started swapping yellow/green/blue...first combo I tried the wheel went backwards! Interesting. I tried another combo that went backwards again, then the final combo it went forwards...great...I reassembled...BUT then under any load it moves you forward but shudders badly, then eventually gives a hall sensor fault if you apply more throttle.

Diagnosis: no idea!!! I'm sure there are some good suggestions amongst the group

Treatment: Do I try a different type of halls? Considering buying a new sabvoton 72150. my current one is missing the programming cable (tried making a new one from a usb rs232 and a rs485 chip but couldn't get it to communicate).

The patients family is quite emotional and hoping there's a specialist witch doctor out there who can give us some hope.

 

[xfrankie]

1) The programming cable is IMO 100% needed. Here they sell it on its own for 5 bucks+shipping. Definitely worth a shot before buying a whole new controller.
You have to have a way to make the ESC perform its "auto-hall-test", whether with the PC cable or bluetooth dongle doesnt matter. Without the auto hall test, I dont think it is ever going to perform properly. I find the PC software very useful for diagnosing the ESC's mental problems.

2) Sabvotons are extremely fucky when it comes to hall sensors (ask me how i know...). Check everything, and I mean !everything! every connector, every part of signal wire, every last place where the connection might be the tiniest bit janky. Push everything together, hardwire/solder whatever you can along the signal path, eliminate any possible breaking point.

Multimeter shows continuity over a cracked part of wire? JK LOL doesnt matter, not gonna work...
You would be very surprised how the tiniest amount of play in a connector is enough to brake the ESC's senses.

Even a slightly improper settings can shut it down. My old thread here is I think a good example of how touchy it is during set-up.

The hall "signal" so to say, is indeed normal to be at around +3 Volt, even if most other ESCs work with +5Volt. Not sure why one of your signals is at +5 while the other two are +3 though, they should be equal... When turning the wheel the signal must switch between 0V and +3V but i guess you know that already.
In the PC software when controller is powered on and you turn the wheel by hand, the "hall status" window must gradually show numbers from 1 to 6, in any order, but only those numbers. If you see anything else (a 7 or a 10 or w/e) then you dont have sufficiently strong/reliable "signal path".

One last note: if you can find a phase-wire combo that makes the motor spin backwards without problem even under load, keep the combo and change the motor spin direction in the programming software (but again, you must have the connecting cable for that).

hope this helps at least a little... just keep trying

 

[Eddie.K]

Sounds like a lot you've learnt through experience - Thanks Frankie!

For $5 will try the usb cable. My sabvoton is old, before the controllers had a usb socket, it just has 2 wires into a plug...hopefully the usb cable has the chip inside it that my old girl needs. I think you are right, I need to get it connected to the pc and do a hall test / calibration next. Good tip!

Regarding checking hall sensor wires and connections...I was hard wiring the new halls all the way back to the controller and then found a loose/broken solder joint inside some old heat shrink on the green wire!!! Made me wonder if i even needed to replace the halls, but no looking back now. haha. Will double check all of these again. Thanks again Frankie, really valuable info!

 

[Eddie.K]

Update - New Sabvoton 72150 arrived (couldn't get the old one to communicate), I run the hall test, wheel rotates as it should, eventually it stops and fails the test - I've done about 20 tests, twice it's said "test ok", but the hall angle hasn't updated. The first time it started at 66 deg and stayed there. I hit "OK" and "Paramater store". Try to use throttle and wheel shudders/runs rough. Retry and test keeps failing. Restored factory settings, hall angle now 245 deg, after another 10 or so tests failing, one tests ok, but again the hall angle stays the same, and motor just shudders.

hall sensors appear to be fine...I rotate the wheel and hall status changes nicely in the software (numbers range between 1 and 6, not hitting 0 or 7 and not faulting). voltage toggling between 0 and 5 volts.

Could it be phase wires pinched inside the shaft leaking current? I previously tested continuity / resistance and nothing between phases, but perhaps it needs some load to arc across a gap??? I've see electricians using a megga tester to apply a test load to phases of industrial motors and check nothing leaks...thinking i could disconnect phase wires, apply the battery voltage (82 volts) and see what happens on the other 2 phases...slightly concerned i might damage something :/ It's a Crystalyte Crown TC80 so expensive to replace.

Any suggestions????

 

[Eddie.K]

Update, as may help someone someday. I cut all the phase wires and pulled them out of the axle to inspect (my last resort before replacing the motor)- they all had insulation that was cracked and broken and flaking away! Likely why the motor shuddered under load. Will replace and try again.

 

[amberwolf]

Teflon insulation does that as it ages. Aging can be calendar aging, or caused chemically or environmentally. It also gets damaged easily when being pulled/pushed around as it's fed thru the axle, as the wiring sometiems has to be forced thru it.

I'd recommend using some other kind of insulation, or else ensuring there is double-insulation between each wire and anything else around it (including any other wire), where the second insulation is of a different kind. And either lubricate the axle wire tunnel, or use cable small enough to easily go thru it, or both, to avoid damaging wires as the cable is fed thru.

Just a bit of info to consider: inside the axle itself, the wires can heat up under instantaneous currents more than elsewhere, as they have no airflow to pull the heat away, and the axle steel is not a good heat conductor, even if the wires didn't have multiple layers of thermal insulation between them and it.

 

[larsb]

I replaced all 3 halls (using ss441A unipolar sensors

I think Honeywell ss411 bipolar sensors are the most commonly recommended here, they have a more stable switching than ss441. It’s not necessarily the cause of any problems but for next time :wink:

 

[Eddie.K]

I replaced all 3 halls (using ss441A unipolar sensors

I think Honeywell ss411 bipolar sensors are the most commonly recommended here, they have a more stable switching than ss441. It’s not necessarily the cause of any problems but for next time :wink:
A8561757-94D1-4F2A-B327-EB5679E2CEF8.jpeg

Ah, Thanks for that Larsb! certainly worth trying ... it's the only thing left as since my last post i replaced the 3 phase wires, tried another set of 3 halls sensors, new hall wires, and got the same shuddering issue. Note I went with the ss441A following other threads on here saying they are the most common! haha :lol:


I've sadly abandoned that motor now but will have a go with those halls one day when i feel like tinkering or building a spare bike....Am now onto a new QS205 and new Sabvoton 72150 $$$$ , another round of issues getting that working but almost there.