Help! Voltage drop when hall sensors connected

[BKero]

Hi guys,

I'm having a problem with my ebike. The hall sensor +/- to the motor dip from 5V to 1.7V when I plug the hall sensor connector in.

My configuration is as follows:

Lyen 12-FET controller
3 phase brushless motor with 3 hall sensors (5-pin hall sensor connector)
48V LiFePO4 battery, CycleAnalyst 2.3

I recently switched over from Lipo cells, and re-arranged from a rack-mounted case to a triangle case. I didn't alter any wiring of the working system.

When I put the entire thing together, the throttle no longer worked. Curious as to why i started taking a voltmeter to things. I originally found that the throttle was dropping from 5V to 1.7V when the motor hall effect sensor connector was connected to the controller, but this morning I looked at the pinout of the motor hall sensor connector and found it too had 5V/GND pins.

My CycleAnalyst and multimeter both show that my battery voltage is normal.

Do you folks suspect that it could be caused by a short in the sensor board inside the motor, or would this be something inside the controller?

 

[Ykick]

Sounds like something's shorted. Another motor would be handy to help narrow it down to interior of your motor?

Any butchered wiring in/out of the axle or controller? Do you have a DVM? Use Ohm function check for short circuit between all the hall sensor wires?

 

[BKero]

Sounds like something's shorted. Another motor would be handy to help narrow it down to interior of your motor?

Any butchered wiring in/out of the axle or controller? Do you have a DVM? Use Ohm function check for short circuit between all the hall sensor wires?

I just took an ohmmeter to the hall wires, and you're right. 0 ohms between +5V and GND on the motor side. That's bad. Also bad is that I see 0 resistance between any of the phase wires! I don't think that is normal, is it?

I suppose I'll be tearing down the heatshrink on here to find where the short is, and might need to crack open the motor to find it. The design is through-axle, but there isn't much keeping it in place.

http://imgur.com/l3aCxz7 is a picture of the wires going through the axle.

 

[Ykick]

How much power have you been pushing into this motor? A Lyen 12 FET carries the potential to make a motor very hot. Very hot when parked has been known to cook-off and damage Hall sensors.

Only legit Ohm resistance check of phase wires - between Hall circuit and/or motor frame, axle.

 

[dnmun]

yes, the phase wires are all connected at the wye to each other. the 5V supply to the hall sensors is a voltage delivered to an IC. the IC is not a resistor inside the hall sensor package, but an active device. the value you get from an applied device depends on what that applied device, your voltmeter, is doing to the IC.

first verify the 5V rail when you plug in the hall plug. if it remains at 5V then we will check further.

 

[BKero]

How much power have you been pushing into this motor? A Lyen 12 FET carries the potential to make a motor very hot. Very hot when parked has been known to cook-off and damage Hall sensors.

Only legit Ohm resistance check of phase wires - between Hall circuit and/or motor frame, axle.

At peak I've been pushing around 3500W at 84V for short bursts. I was able to touch the phase wires afterwards. The casing was warm, but not hot. The shell of the motor was also warm to the touch.

In particular the controller was this one: 12 x 4110 MOSFET Extreme Modder Controller LYEN Edition.

The odd part is that I was able to ride the bike successfully for a 40 mile journey at a lower load/voltage the day before. The problem just arose when I started bending cables to move the connectors from the rear rack case to a triangle-mount case.

 

[BKero]

yes, the phase wires are all connected at the wye to each other. the 5V supply to the hall sensors is a voltage delivered to an IC. the IC is not a resistor inside the hall sensor package, but an active device. the value you get from an applied device depends on what that applied device, your voltmeter, is doing to the IC.

first verify the 5V rail when you plug in the hall plug. if it remains at 5V then we will check further.

I tested the voltage of the 5V/GND from the controller, which appears as ~4.85V. Upon connecting the hall sensor connector that voltage drops to around 1.75V.

I took an ohm-meter to the 5V/GND pins from the hall sensor connector going to the motor. That revealed that the resistance between the two is 0 ohms. I'd suspect that means that two of the wires are shorted somewhere along the cable.

Also interesting is that they are about 1.5Mohm resistance from the axle.

I measured the resistance between A, B, and C hall effect sensors.

A to B is infinite
A to C is 30 Mohm
B to C is 31 Mohm

 

[dnmun]

is the controller open so you can measure the 5V rail?

if not, then measure the voltage between the red wire and the black wire to the throttle from the controller when you plug in the hall sensor plug.

can you see the 5 wires to the hall sensors on the motor side of the plug and can you see wires touching?

 

[BKero]

is the controller open so you can measure the 5V rail?

if not, then measure the voltage between the red wire and the black wire to the throttle from the controller when you plug in the hall sensor plug.

can you see the 5 wires to the hall sensors on the motor side of the plug and can you see wires touching?

I've measured the 5V rail on the throttle. It too is outputting 5V but drops when I plug the hall effect plug into the motor side.

I can see the connector for the hall effect sensors, but none of the wires there are touching. I gear that the shielding has gotten warn off somewhere in the cable, maybe in the axle.

If that is not the case, then I fear that a hall effect sensor might have blown, and must be replaced.

 

[dnmun]

i doubt if the IC is damaged. but if your 5V rail measured at the throttle is dropping to 1.75V also then it is a problem in the 5V regulator.

 

[amberwolf]

I doubt it has anything to do with the controller, since:
A) it only drops voltage when the hall connector on the motor is plugged into the controller per this

I've measured the 5V rail on the throttle. It too is outputting 5V but drops when I plug the hall effect plug into the motor side.

and

B) the 5V is directly shorted to ground somewhere on the motor/motor cabling side of the hall wiring per this:

I just took an ohmmeter to the hall wires, and you're right. 0 ohms between +5V and GND on the motor side. That's bad.

So the problem is somewhere within the wiring between the halls inside the motor and the connector at the end of the cable coming out of the motor. (or one of the actual halls is internally shorted, though I've never seen one as a dead short of 0 ohms before I suppose it's possible).

 

[Ykick]

40 mile journey at a lower load/voltage the day before.

That's a long run time with a rather large controller. Lower voltage and whatever you mean by "lower load" has little to do with anything. "I rode it one day and the next day Hall sensors were toast." is a very common description for the way Hall sensors expire. While the hub spins it's better able to shed heat. When it stops turning, it just cooks-off.

This thread might be of some use to further qualify motor sensors:

http://endless-sphere.com/forums/viewtopic.php?f=3&t=57053&hilit

Before cracking a motor open (which isn't that bad, btw) I would try an external 5V supply and resistor to provide the power for testing Hall sensors and eliminate the controller from the equation.

Be aware older motor sensors tolerate very wide voltage range between 5-30V and I've done that same test with 12V battery but I've also learned/discovered some newer "P" Hall sensors have internal pull-up resistor which greatly limits device operating voltage to not much more than 7V.

Here's a quick search of posts relating to motor Hall sensor troubleshooting:
http://endless-sphere.com/forums/search.php?st=0&sk=t&sd=d&sr=topics&author_id=14677

 

[made_in_the_alps_legacy]

The problem just arose when I started bending cables to move the connectors from the rear rack case to a triangle-mount case.

Do I understand correctly ? seems a good indication to me... that you cabling needs a closer look since :

the 5V is directly shorted to ground somewhere on the motor/motor cabling side of the hall wiring .

 

[dnmun]

if the 5V of the motor side of the hall plug was shorted directly to ground then the 5V would be at 0V and not at 1.75V.

the 1.75V is to be expected if there is 5V on the 5V rail in the controller and the current flows out of the controller through that current limiting resistor 51 ohms to the halls on the controller power supply 5V to the halls.

so it is not a direct short imo. i asked him if the 5V on the throttle remained at 5V when he plugged in the hall plug and he indicated that it has also fallen to 'low' voltage but did not say what voltage so i am assuming that the current load is too high for the 5V regulator to supply so it is either a 5V regulator problem or the hall sensor ICs themselves are now using so much power that they pull down the power supply voltage to them.

if he was still around there might be some way to isolate the problem but there is just limited info and in the case of the power to the throttle is really an unknown so no way to tell if the 5V regulator died or if it is the halls. i would not expect the 5V regulator to just die for no reason, but observing the actual voltage on the throttle i thought was the best way to proceed. so that is where it ends.

 

[amberwolf]

I guess I was just going by what he'd reported; I always have to assume that what is reported is accurate, otherwise I can't help anyone if I have to question everything they do and have them mad at me (or stop responding).

I used to have to do that when I did tech support, and got tired of people being angry at me for questioning them, when in fact most of the time they had indeed not done what they said they had, and so any help given them based on that would be unlikely to actually help, and be a waste of our time....

 

[dnmun]

i thought your advice was spot on. people do not appreciate your skill and i find that most people have already made up their mind so when you try to figure out how to analyze the problem in a logical fashion because of your experience then they just decide that what you are asking is not germane in their opinion so they just drop out and never reply.

i am like you and feel that the search, or analysis of the problem, is more interesting than using the device. it is like a personal quest to logically remove the variables based on your learned skills, which others do not appreciate as they should.

i also think there is some kinda short but not sure where it is on the motor side. maybe checking continuity to the motor case would help him but if it had shorted to the phase wire then one of his halls would just show a dead hall signal imo.

 

[amberwolf]

Based on the bending of wires done during rearrangement of stuff, I expect it's inside that cable where the bending itself was done.

I have seen a number of cables in ebike stuff (one brake, two throttles, three halls, and something else I forget what) that have wire splices well-inside the outer black jacket, as if when they were creating that cable they had spools of each color wire, and when one ran out (or broke) they just spliced in another spool, and continued with the cable manufacturing. But when they do the splice, there's no insulation on that area at all, except for the black jacket itself that contains all of the wires of that cable. And all of them were just "knotted" or "folded" together. No solder, etc. Just small enough that there's no lump, about the same diameter as it would be with insulaton on the wire itself. There's pics of one of these in one of my Fusin posts or threads, from the DayGlo Avenger days I think.

So if something were to happen that breaks the very thin insluation on *another* wire inside that cable at the same point (like rubbing of those splices on the outside of anothe rwire's insulation, if they aren't all supertightly fitting inside the jacket adn are at a bend), a direct short could occur. It might not even be permanent, only intermittent, which would explain various problems people have that are fixed when they redo their cabling.

I've yet to just cut open such a spliced cable all along it's lenght to see if it has other splices or damage, becuase it is unlikely, but I suppose it's also possilbe thre are two splices in proximity, and pulling or bending of the wire could bring them into contact (or if the wires are not tightly bound by the jacket, nad you pull on the jacket, it could pull some wires harder than ohters, especially if hte splices are near one end).

Just speculation....