Halls effect sensors...the plague from hell

Warren

100 kW
Joined
Oct 5, 2010
Messages
1,628
I have gone over 54K miles on my two electric assist bikes. One of the halls started acting up on my mid-drive MAC bike at 8K miles. At 9K it stopped working all together. I replaced the cheap sensored controller with a cheap sensorless, and rode another 21K miles. It sort of worked, but was never as efficient, and growled unhappily on startup much of the time.

My big DD cargo bike worked perfectly for 22K miles, but has now started acting up intermittently over the last 1,200 miles. Being a very high power DD, when one of the halls quits, it is like one of my old 4 cylinder cars when a plug would fail...rough running, noisy, low torque. As with the MAC, the problem is happening more frequently, for longer, and I eventually will have to give up,and do something.

Here is my dilemma. I am old, with the failing eyesight, and shaky hands that come with age. I don't feel confident to replace the halls, and understand that there is a good chance of ruining them with static from my hands when doing the installation.

I searched around and found someone who had put external halls on a skateboard outrunner motor. Getting halls near the magnets on a hubmotor would probably be impossible with the spokes in the way, but what about a simple magnet ring, like people use for pedelec? It seems like an external encoder would be a great aftermarket product for somebody....Justin? :)
 
As long as the magnets on the encoder line up with the magnets on the rotor, and the halls on the sensor unit line up with the originals (or are at the same electrical angle, which may correspond to a physical angle) there's no reason it wouldn't work.

You would probably need one magnet on the outside for every magnet on the inside, at the same polarity (they reverse on each successive magnet). At a guess, you'd need 46 magnets, as most of these DD motors use 23 pole pairs, and arrange them so tehy are north, south, north, south, etc., so they trigger the new external hall sensors in the correct pattern and timing for the phases inside the motor. YOu could probably just use little button magnets like the ones on PAS sensors or badge holders, and glue them on the outside of the side cover, whereever is convenient. The hall sensors could then be placed in an arc on a bar on the frame next to those.

The static is easy to control; grounding wrist straps are a couple bucks that just clip to a grounded object, or plug them into the ground prong hole of an outlet. As long as you're wearing the strap, you are grounded and don't build up any static.
 
Intermittent halls problems are always going to be in the wiring to the motor. Some have gotten fed up, and simply solder connected all the motor wires. Might be worth it to try that before some kludge.


By now though, you might have a legit halls failure. Test them before you take apart the motor. To make the test really easy, buy a motor/controller/ throttle tester. Best 20 bucks I ever spent.
 
dogman dan said:
Intermittent halls problems are always going to be in the wiring to the motor.

Are you saying that there is no intermittent failure mode for a hall? That would be fantastic. It would mean the halls on both my bikes are actually OK, and the problem is in the wiring.
 
I have had intermittent hall sensor problems on thumb throttles as the halls go out, but I have never had them go out on my hub motor. I agree that checking the connectors and wiring is the first step you should take. Make sure all the contacts are good and solid. Also, have you tested your halls like this to make sure something is wrong with them? https://www.ebikes.ca/documents/HallSensorTestingFinal.pdf

:D :bolt:
 
The problem with testing is that the problem is intermittent. I had the problem the first time 31 miles into a 60 mile ride, for a few miles. The next ride it happened 7 miles into a 40 mile ride, again for a few miles. Then it didn't happen for hundreds of miles. The other day it happened three times in the middle of a 51 mile ride. Yesterday I had no problems on a 41 mile ride. It needs to either break, and stay broken, or I need to wiggle the wires, and connectors while it is happening. I will definitely try that the next time it happens.
 
e-beach said:
I have had intermittent hall sensor problems on thumb throttles as the halls go out

I have never had a problem with the halls in my throttles. The throttle on my mid-drive started failing intermittently after a year. I determined there was a break in the line along the handlebar run. I assume it happened from my moving the bike around grabbing that part of the bar. I never bothered to fix it. If it stopped working, a quick touch of the cable fixed it...for almost 30K miles. My hobby is riding, not fixing e-assist bikes. :)
 
Warren said:
... My hobby is riding, not fixing e-assist bikes. :)

Maybe just buy a new motor and sell/give/trade your existing motor to someone else who might like fixing it? 22K miles may not be as far as you'd like, but it sure ain't bad.
 
The Crown 4080 costs $800, plus shipping, is out of stock, and I'd have to rebuild the wheel. This motor has never been run outside its design specs for volts/amps/temperature. I get that the bike building hobbyists on this forum think this is acceptable. I think it is ridiculous. Something as failure prone as hall sensors should be designed as a plug-in replacement. I shouldn't have to dig them out of the motor, glue in new ones, and re-solder a bunch of tiny, fragile connections, all while wearing a ground strap.
 
Warren said:
The Crown 4080 costs $800, plus shipping, is out of stock, and I'd have to rebuild the wheel. This motor has never been run outside its design specs for volts/amps/temperature. I get that the bike building hobbyists on this forum think this is acceptable. I think it is ridiculous. Something as failure prone as hall sensors should be designed as a plug-in replacement. I shouldn't have to dig them out of the motor, glue in new ones, and re-solder a bunch of tiny, fragile connections, all while wearing a ground strap.

I didn't know what motor you had. A quick survey of your thread before posting didn't turn up a reference. Neither did a quick google search.

On an $800 motor, I agree. Repair and maintenance should be easier. And reliability should perhaps be better to begin with.

On a motor that is just one item representing perhaps half the price of a $225 kit - gluing on replacement Halls or simply replacing the motor after 20K miles seems reasonable. I'll spend that in tires by the time I reach 20K miles.

If you had a replacement motor, couldn't you just swap out the center core and leave the outer portion attached the spokes/rim?
But right now it seems like your best time/effort/money option would be to simply hire someone with younger eyes to do the hall replacement repair for you.
 
How have you determined the problem is 100% the hall sensors?
 
e-beach said:
How have you determined the problem is 100% the hall sensors?

No. Dogman Dan suggests intermittent hall problems are caused by faulty connections. I hope he is right. He has tons more experience with this stuff than I do, and that would be a lot easier to fix. I just did quick 27 mile ride, at 25.6 mph average without a glitch. When this thing is working, it is like a Stradivarius. When it's not, it is like a cigar box banjo with two strings. It is gutless, and riding at WOT on the level it will show 300-500 watts at 15-18 mph, with me pedaling for all I'm worth. Up hill it slows do to 8-10 mph, sounding like the coffee grinder at our local supermarket. Downhill, pushing the regen button gets 50-80 watts, and no discernible slowing. The problem happens with no warning, and is instantaneous, like flipping a switch. One second it is powerful, and smooth as silk. The next it is wheezing.
 
Dogman's advice is not hard to follow. Your motor probably only has 2 connectors attached to the short wiring leads coming out of the motor. I would suggest pinching the connectors with the thumb and a finger, and with the other hand, gently but firmly push the wires into the back of the connector.

Quite honestly, I don't think it is your Hall sensors anyway. If you motor only has three of them, then your controller can not read the sensor sequence properly if one is bad. Your motor would just stop.

So check your wires for insulation abrasion where you might have rubbed against something and exposed the copper inside the insulation causing a short.

Push the connectors back together like I mentioned above.

Get a kid to follow the instructions on how to check the halls from the link I posted earlier in this thread.

:D :bolt:
 
Warren said:
I think it is ridiculous. Something as failure prone as hall sensors should be designed as a plug-in replacement. I shouldn't have to dig them out of the motor, glue in new ones, and re-solder a bunch of tiny, fragile connections, all while wearing a ground strap.
Your'e right--you shouldn't. It's completely unnecessary PITA. :/

If the systems were correctly designed then when a motor is run within it's specs the hall sensors would never have a problem at all.

But they have no protections against the several possible things that could feed excess voltage/etc into them, and the wiring/connectors/etc system is not well-designed to prevent any of those things from happening, or to prevent intermittent breaks of contact, etc.

Even the way teh wires come out of a hub motor, and the way the motor torque is transmitted, could be much better designed.

But they're designed to be as cheaply manufacturable as possible, instead.

There are a few exceptions of certain parts of certain models of certain brands, etc., and some aftermarket stuff and some OEM bike stuff is better.

But not much of it. Even most of the stuff Grin Tech sells is just slightly improved...though they're working on changing that.



I've seen so many "stupid" failures, caused by manufacturing mistakes, or plain carelessness of assembly, that if it were designed differently they *might* have worked anyway, or might not even have been possible to build wrong, etc. Not that it doesnt' happen in other industries...but in this one, it's rampant.


It wouldn't even cost much more to do it "right", but it would require changing things throughout the industry to keep everything compatible (because changes are needed to controllers to make the protections against overvoltage/etc in hall lines complete, in addition to those in the motors themselves).
 
Warren said:
Dogman Dan suggests intermittent hall problems are caused by faulty connections. I hope he is right.
It's certainly possible--it's more likely than a defective hall sensor itself, if it's not overheating, or electrically damaged.

Wires can be broken inside insulation, anywhere from a contact crimp or solder all the way back to the solder connection on the hall itself or the PCB inside the controller. It's not easily detectable, because you can't really see it, and usually can't feel it.

Sometiems you can find it by wiggling the wires along their lengths, just a bit at a time, until the problem occurs (or the continuity buzzer on a multimeter stops, if you have one connected from one end of the wire to the other).

The most common place for problems is wherever wires bend sharply, or are pulled, or enter or exit a hole.


Sometimes it's just corrosion on a contact surface, and unplugging and replugging a connector will fix it--sometimes it requires multiple cycles to scrape the contacts clean.


Sometiems it's a bad crimp in a contact pin in a connector, either wire broken at the back fo the pin, but held on by the insulation or the pressure of other wires in the cable, usually making a connection, but not always, and sometimes it corrodes or oxidizes there. This is especially a problem with JST connectors, or the "molex" or "tamiya" or similar types, that are open-frame connector plugs, with no cable support or strain relief.
 
I've never had a hall failure either, just a hall wire failure where the insulation partially melted and became brittle causing intermittent shorts between wires. I could be wrong, but I don't think a hall failure can be intermittent...either it's dead or still good. I've never run a Craplyte though and remember reading something way back about them using some 2 cent halls to save a few pennies on the cost of each overpriced motor.

Assuming it is just a wire or connector problem, I wish you luck that it's something outside of the motor and not where mine was, where the wire harness goes thru the axle, which meant a harness repair that's more involved than just changing out hall sensors.
 
Neither did I until yesterday. After 17,000 miles on my front hub motor, it seems that I blew a hall on my front hub motor. The other two test to 0v - 5v as I spin the motor slowly backwards, but the one on the yellow wire stays at 5v the whole time.

:oops:
 
If it stays at 5v, most likely there is a broken connection between the controller and the hall inside the motor. Usually it's just a problem at the connector contacts, but it can be a broken wire anywhere along the way.

Since you've measured 5v, then the break is somewhere between that point and the hall inside the motor.
 
amberwolf said:
If it stays at 5v, most likely there is a broken connection between the controller and the hall inside the motor. Usually it's just a problem at the connector contacts, but it can be a broken wire anywhere along the way.

Since you've measured 5v, then the break is somewhere between that point and the hall inside the motor.

And here I was thinking, I got return voltage so A) positive and negative leads are good. B) got a return voltage so return lead is good. But, C) the Hall sensor was staying at 5v so it is bad because it is not latching/switching. :oops:

In any case i checked all wires and connections and they look good so I am going to have to open the motor. At least it had the courtesy to happen during a pleasant time of the year around here and not the dead of winter.

:D :bolt:
 
e-beach said:
And here I was thinking, I got return voltage so A) positive and negative leads are good.
Unfortunately because the 5v is provided *on the signal line itself* by the controller, it doens't even mean the positive and negative of power supply are good. (those have to be checked separately, though if other halls are working then they're probably good).



B) got a return voltage so return lead is good. But, C) the Hall sensor was staying at 5v so it is bad because it is not latching/switching. :oops:
IT could be that, but since the sensor just grounds the output, then a sense line at 5v could be disconnected from the sensor completely, and just pulled to 5v by the controller.

Equally, a 0v hall signal can be a broken wire at the controller end, so there's no pullup to 5v, and thus nothing for the hall to ground/leave open.
 
amberwolf said:
....
.......Equally, a 0v hall signal can be a broken wire at the controller end, so there's no pullup to 5v, and thus nothing for the hall to ground/leave open.

I switched the battery, controller, and throttle to another motor and they all ran fine on the different motor. What ever it is, is probably in the motor. I ordered 10 - SS41 Honeywell halls from Mouser. (6 of them would cost me $22.00 while 10 cost me $23.00. :roll: ) Anyway, a few on the shelf.

:D :bolt:
 
Doesnt' hurt to have them around, but it *is* usually a wiring problem rather than an actual sensor failure. (even when a lead breaks off the hall right at the body, it's actually a wiring failure rather than component, becuse if you could connect to the sensor inside it's likely still working ;) ).

FWIW, most of the sensor failures that just randomly occur, with no wiring/axle/controller issues prior to that, are likely static-electricity induced (ESD), because it's VERY unlikely that the people installing them in the hubmotors during manufacture are using any form of ESD control.
 
e-beach said:
amberwolf said:
....
.......Equally, a 0v hall signal can be a broken wire at the controller end, so there's no pullup to 5v, and thus nothing for the hall to ground/leave open.

I switched the battery, controller, and throttle to another motor and they all ran fine on the different motor. What ever it is, is probably in the motor. I ordered 10 - SS41 Honeywell halls from Mouser. (6 of them would cost me $22.00 while 10 cost me $23.00. :roll: ) Anyway, a few on the shelf.

:D :bolt:

Which SS41 halls did you buy? There appears to be different SS41's. Are these standard for most hubmotors? I would not mind having 10 around just in case for spares 8)
 
Just the normal Honeywell SS41's Mouser, (no extra letters) they are standard enough for my specs. I just took a look at the SS41F's and had I went with them, I would have saved a couple of dollars. :roll:

:D :bolt:
 
Back
Top