Cheap motor works fine but controller keeps breaking

[S0up13]

Hi there,

I upgraded to a 48v 1000w brushless, gearless hub motor with hall sensors that runs on a 26-32A controller (around 1500w). My battery is a 48v 17.5ah with a 30A BMS cutoff. I figure I am pushing the battery cutoff limit but I use a 3-speed switch to keep it on low, which has run fine for days on end, sometimes nearly 100k without a problem.

The problem is, after a short amount of time (weeks or days), it keeps on shutting down and does not work unless I replace the controller. The bike rolls fine when it is not working, which I have read would point to a blown MOSFET. I have some experience building and maintaining ebikes so I have taken care of the basic troubleshooting steps (yes I have connected the switch lock and tried recharging the battery) to narrow down that it is the controller.

I weight 150lbs and my area is very hilly (lots of steep inclines in every direction). I cannot determine why this keeps happening. I purchased the 1000w motor in a DIY kit for an incredible $250 CAD (around $185 USD), so when the square-wave 2-mode controller blew I figured it was just part of a cheap kit. I am now on my second sine-wave controller that has become unresponsive. The replacement controllers are the ones bought for around $55 CAD/ $40 US, orange in colour, 26-32A, three-mode. They are very popular on Amazon these days.

I am wondering if the motor could be wound improperly or cheaply built in a way that is somehow blowing the controllers and/or my throttle? Or is it my battery causing voltage sag and high currents are blowing sensitive controller parts? Are the cheap controllers to blame? Could it be any or all of the above?

Any ideas on how to test/troubleshoot are appreciated. I have a digital multimeter and a bench DC power supply. Thanks.

PS This is my first post as a member but I have read countless threads from ES over the years to help me in my tinkering. I am grateful to the many contributors and brilliant mad scientists out there who share their knowledge with the world.

 

[amberwolf]

There are a lot of possibilities, so the below is a mix of info and questions that might help narrow down things.

You said here:

The bike rolls fine when it is not working, which I have read would point to a blown MOSFET.

which is a bit confusing: if the bike rolls fine and its' a DD hub (not geared) then a blown FET is usually eliminated unless you mean it only rolls fine if the controller is unplugged from the motor, but otherwise is draggy or cogging.

This part

I figure I am pushing the battery cutoff limit but I use a 3-speed switch to keep it on low, which has run fine for days on end, sometimes nearly 100k without a problem.

do you mean 100kilometers? Or 100,000 kilometers (or miles, etc)?

Note that on most controllers, the 3speed switch only changes the max throttle level the controller actually reads--scaling the throttle input from probably something like 50% on low, 75% middle, 100% high. Most of these the throttle only controls the speed of the motor, not how much current it takes--so on hills and during acceleration before it reaches the speed set by the throttle position it is still pulling just as much current from the battery on low as on high. Some more advanced controllers (FOC) use the throttle to control torque (current) and so the switch would affect battery current drawn instead, but yours probably are not like that.

If the BMS is designed and working correctly (which you can test if you have something to show battery current and voltage, and can force a situation that exceeds the current limit) then if you reach the current limit it should shut off it's output, at least until the excess current stops; some require resetting by disconnecting the load (unplugging from the controller) and/or connecting to the charger.

Some BMS dont' actually monitor current or shutdown on overcurrent, but may be damaged by it--when they are damaged, usually the FETs are stuck on, so the BMS cannot protect against anything, and so the pack can be overdischarged, allowing cells to go too low in voltage. But if the controller is made for the pack voltage (48v / 13s, 52v 14s) and isn't an autodetecting LVC, it will shutdown (but stay powered on) and stop allowing you to use the motor until the pack is recharged (or recovers from load removal) up above that LVC. If the pack is unbalanced it could still overdischarge a cell group (or groups) but the entire pack shouldn't be overdischarged. If the controller does not have the proper LVC for the battery used, then it cant' do this protection.

Or is it my battery causing voltage sag and high currents are blowing sensitive controller parts?

The battery doesn't directly cause voltage sag--the current demanded by the controller initiates this process, and a battery that has too high a resistance / insufficient ability to deliver that current will then sag in voltage. So as long a hte battery is capable, of the current the controlelr asks for, it won't sag too much. (they all sag some, more the less compatible the battery is with the system, or the lower quality the battery is).

You can test for voltage sag with a wattmeter, as most of them have a Vmin and Amax that let you see about what the lowest voltage was at the highest current, and then go test the system under worst-case and normal riding situations to see what happens. Even a regular voltmeter will help, but isnt' as fast a response and you have to be watching it at the time, which can be hazardous on a ride (unless you use a camera or phone to video it during testing/riding).

The problem is, after a short amount of time (weeks or days), it keeps on shutting down and does not work unless I replace the controller.

By "keeps on shutting down" do you mean it actually powers off, (and if so, is taht at the battery or just the controller), or does it just stop allowing the motor to run?

What is the battery voltage when this occurs?

If you can measure temperatures, what are the motor and controller temperatures at that time?

By "does not work", do you mean it wont' allow the motor to run / not respond to throttle, or not power on at all, or some other thing?



Is the controller mounted in the open air where it can be cooled, or inside a bag, box, etc with no airflow around it?

Do you have a wattmeter or other instrument you can put between battery and controller to see watts, amps, volts, etc, while riding? Anything to measure motor or controller temperatures (inside) while riding? If any measurements have ever been made, what were they, and under what specific conditions?

Damage (from heat, scrapes/pinches, etc) to the motor-controller wiring (usually at the axle or inside the motor) might intermittently allow a short between phases or from phase to hall that could damage the controller--it might not be immediate-fail damage, either, but could be enough that it eventually fails. Certain other wiring faults could also do this, such as anything that can short the 5v to ground even intermittently.

Winding damage (from heat or vibration or factory mistake) that intermittently allows phases to short to or thru the lamination stack (stator teeth) could also do the above, and not be detectable except during the instant they temporarily create the short (except with HV-isolation testing equipment, which can usually detect these via arcing at the hundreds of volts they can test at).

Controller internal overheating can cause capacitors to fail; often they swell up visibly in these cases, but not always. Insufficient capacitance lets voltage and current spikes thru the system, and can damage any components on the supply rails those caps are supposed to filter.

Similarly, it can damage the LVPS (low voltage power supply), usually the parts that get hot on their own anyway, like any big input resistors, or voltage regulators, etc.

If the controller has a display, and it stops powering on at all, the display may have failed, ro the wiring between the two. (guessing there isnt' one if you have to use the KSI (ignition) wire. )

 

[S0up13]

Thanks for your detailed reply.

When my motor stops, I mean that it stops responding. Even before disconnecting the controller from the battery, it still walks/rides normally.

My controller is mounted outside with good airflow, and the current climate in Ontario has been -10C / or 14F.

UPDATE: After leaving my battery on full charge for a while (I usually pull it at 80%), today I tried my last controller that had broken and lo and behold! it worked! So I must have missed something here. I am thinking that it is the battery LVC or high-current shutoff that must take a while to reset: last time I left it for a day before switching out the controllers. But I am using a cheap battery purchased from Amazon and perhaps the BMS needs to be plugged in for a while on full charge to reset? I am planning to upgrade to a QWW 48V 50A battery soon. I did have a throttle replacement in there recently but that was just a cheap throttle.

I will check the temperature while riding and watch my voltage meter on the handlebars so I can try to narrow it down if this happens again. I have a meter testing current output but it requires a shunt that I have never worked with before so I have avoided installing it. Now I have a reason to find it.

 

[amberwolf]

Given the new info, it's likely the battery simply isn't capable of what the system needs, especially in the cold temperatures. (batteries generally are less capable when cold)

Some thoughts based on many other similar problems reported here on ES:

If the battery is like most of them, it's made of cheap unmatched cells that do not charge or discharge at the same rates and have different capacities. That means that under load some cells sag in voltage more than ohters, and some will at some point sag below the BMS LVC, so the BMS then shuts off it's output (which means there is no power to the system at all) to protect the cells from damage that could lead to a fire.

Some BMSs will reset as soon as the cells "recover" from the load, but some require the pack be recharged to a certain point, so that the cells' voltages rise above some cutoff point before they will reenable output.

Some BMSs don't properly reset at all, or do so under what seem like random conditions (I have one here that actually has to be *drained* more (thru the charge port!) to reset it when it gets stuck like that, but normally it resets once cells rise above LVC). Some have a reset pad on the board, or a reset procedure, some don't. Finding out if they do is difficult, as few BMSs have any info pages out on the web, and virtually none come with a manual or spec sheet (especially if they are alread built into a battery).

The imbalance problem, if you have one, is not something you can fix without replacing all the cells in the pack with well-matched ones (which are not going to come from the cheap-pack or cell places, because cell matching costs a lot more in time and labor and discarding cells that don't match other cells).

It's easier to replace the whole pack with a new one that has well-matched cells to start with, which unfortunately is somewhat difficult to find--most pack makers think that making sure all the cells are at the same voltage when they build the pack is "matching cells", but what it really means is making sure they all have *the same* (not similar!) capacities and internal resistances, etc. Some places may do this (you can ask EM3EV, for instance; I can't think of another pack maker that I trust would reply honestly instead of just saying whatever will get you to buy something).

The BMS will "balance" the pack by making all the cells the same voltage, but it doesn't make them the same capacity or properties, so it has to be done every time you charge the battery. If you leave the battery on the charger after it has "stopped charging", the BMS will continue balancing and the charger will turn on and off periodically until the BMS is done balancing, then it will stay off. Depending on the imbalance and the BMS's balancing capability (usually very small) it can take hours to days to even weeks to balance a pack the first time around. Usually if the pack is not run down to empty the next balancing will only take a few hours at most, but a badly-matched pack can become badly imbalanced the harder it is used and the more empty it is drained to.


BTW, if you still have power to the system when the motor stops working, then the battery itself is not the (whole) problem. If you measure the battery output with nothing connected to it, it may show a voltage, but with the controller plugged in it would drop rapidly to zero, if the battery's BMS output has turned off.