Strange issue with e-scooter controller, and battery voltage

progrock

100 W
Joined
Oct 2, 2017
Messages
219
So, I;m trying to help a friend fix his 60V scooter. The battery had been drained and left that way for some time,... I took the battery from him, and brought the charge back up to about 60V and returned the battery. The scooter had been running fine before the battery died (months ago). He reported to me that while his charger now worked, and that the battery was fully charged (about 67 V), the scooter would not work. I did a basic investigation of the scooter.... all the connectors, and the battery again. When the battery was unplugged, the output read the correct 67 V.... but as soon as it was plugged into the controller, the voltage suddenly dropped to around 11 V (measured at the battery's output connector both connected and unconnected). So... I removed the controller, and attempted to investigate further. I followed the instructions for testing blown mosfets on ebikes.ca (http://www.ebikes.ca/documents/BlownMosfets.pdf) as well the similar instructions http://www.lsdzs.com/en.php/article/show/id/50.html...

When measuring the reistance between the 3 phase wires and GND... the Green and Blue phase wires measured ~10k Ohm... while the yellow measured 8k Ohm. (I beleive that means yellow's bottom bridge mosfets may be having problems)

When checking continuity between positive power and the phase wires, all 3 had 0 continuity... which I thought meant ALL the top brideg MOSFETS were broken.

These results seemed to make sense to me, considering the voltage drop when the battery was connected to the controller... and a have a relatively large number of good mosfets laying around, so I figured I might as well replace the 15 stock mosfets with the new, better ones (The stock ones were Si Tech S80N10R.... planning to replace them with CSD19536KCS FETs).

I began by taking the controller apart, and removing all the current FETs (luckily I have a a Hakko FR300 solder sucker.... made this much easier/cleaner/quicker). After removing them, I used a transistor tester device I have to check each individual FET.... while I planned to replace all of them, I figured I might as well keep the good ones.... BUT... ALL 15 FETs tested fine... this seemed odd... so I then tested most of the FETs manually with a simple DMM.... same results, all fine. I investigated the rest of the circuit board, and the pieces the FETs mount to, etc.... couldn't find any obvious problems at all... surprising the controller, while rather simple, is very clean and in pretty good condition.

I'm now very confused at what has gone wrong.... where I messed up in the tests.... and also, I need to figure out what is actually wrong, causing the battery to drop to almost nothing when plugged into the controller, and why the scooter will not work at all. My only though right now is, I took the controller from my friends place, back home to fix it here, but I don't have the scooter or battery on hand. When I found this voltage drop, the controller was still connected to the motor (at the very least the phase wires, but I believe more than that). I'm now worried that I should have tested everything, once disconnecting the controller from the motor completely. Also worried that the actual issue could be the motor now. I'm really not sure, the FETs seemed to make the most sense to me... but I've literally triple checked them, and can't find anything wrong with them. I'm likely still gonna replace the FETs with the better ones (I have quite a few CSD19536 around), but I 'd like to figure out what actually is wrong, and fix it... before putting the FETs into a controller that may possibly never work again. I suppose there's also a small chance that there was an issue with the FETs, but removing all of them, and reinstalling could have unintentionally fixed it... tho I find that somewhat unlikely. I'm still quite confused at the tests results from checking the resistance/continuity between the phase wires and neg/pos power wires, before I even opened the controller.

Anyone have any advice?... anything I should look into next?... also pointing out my mistakes may be useful,hopefully I'll be less likely to repeat them next time.... any and all help is very welcome. It is definitely annoying not having all the pieces together at my place to make things easier... but I could not get the rest of the parts (scooter/battery) between the apartments at this time, so for now I can only have the controller... would like to see if there is anything else I can find wrong with it today, then get it back together... supposed to return to my firend's with it tomorrow, and if lucky have the scooter fixed with the fixed controller... if not so lucky, will have to attempt to troubleshoot/fix whatever is still broken when I am there.

EDIT: BTW, just check the Phase wires resistance between the pos/neg power wires again.... still getting 10k, 10k, and 8k (green, blue, yellow to neg power).. and infinite resistance between blue/yellow and pos power, and jumping between infinite or jsut REALLY high resistance between green and the pos power.... more or less same results when the FETs were installed, but this is with them completely removed. Not sure if that's exactly as it should be, or if that means something. The yellow being 8K seems strange, as well as the Green not being completely infinite, while the other 2 are seems a little strange, tho maybe it is normal.. I don't know... figured it was worth mentioning... not sure what to do next... would like to run more tests before reassembling, but not sure what I should be doing.
 
I wish I had a friend that could do all that to my controllers! But with all that, it's probably the battery. The way it drops voltage under load after sitting discharged for a long period would make that pretty suspect.
 
It's possibly it's the battery... I need to get an ammeter in there... but seeing how its not running a motor.. or really anything.... I didn't think it was dropping voltage due to load, was thinking it was more likely a short or something... but I could be wrong.

At least I'm learning more as I go through all this... but definitely a disappointment if there was absolutely no need/reason to remove all these FETs... still not sure what I ant to do yet.. might end up just reinstalling thew stock ones for the time being... not sure if there's a reason to use the CSD19536's on this... maybe I'll use some IRFB4110's... tho I have less of them. THo if these stock one's are indeed 100% fine, they'll do the job fine too.... I just feel like, if I'm gonna ut this much work into it anyway, maybe I should throw in better FETs... even if there is technically no point to doing so at all, haha.
 
FWIW, it is possible for FETs to measure perfectly good in a static measurement test, but when in operation at voltage, they don't switch correctly, or have leakage currents (stuck on) or high on resistance, etc. Usually the failures are so bad they're obvious in even a static measurement though.

However, it's much more likely that the cells in the battery are not balanced, or are damaged from overdischarge or other issues, and the BMS is shutting the pack down to prevent problems.

The pack might have charged enough for the charger to shut off, but it almost certainly isn't balanced. Sitting a long time (long enough to shut off it's output) probably drained some cells (whichever ones power the BMS) a LOT more than the rest, so it's probably badly unbalanced, and may take weeks to rebalance via the charger and BMS.

You can open it up and check it's cells individually, and manually recharge the low ones individually if you have a single-cell charger for that chemistry, if you want to speed up the process.

There are hundreds (at least) of battery test/repair threads you can look thru for details of the processes.
 
Thanks for the advice amberwolf (again, haha, very appreciated). I'm gonna replace the FETS I guess (might as well) and see what happens. I guess I should concentrate on the battery if this is not the issue.. it's by far the most likely things next to them. I have a 10s iCharger....so wouldn't be too hard to balance charge the cells as needed... I was hoping that the BMS would kick in on full charge and balance for me... but maybe I was wrong there. Probably should have used the balance charger in the first place I suppose... guess I was more likely expecting too much from the BMS. And of course, while I have some pretty nice bestech BMS's in the batteries I built... this one probably has one of the cheapest most generic chinese ones you can find. The battery had a couple of surprises.. felt light for it's size.. I was expecting some really cheap no name batteries.. but instead it has LG 2.600mAh cells, in a 16s6p configuration. BUT the battery holder has space for close to double that, the entire center is just empty spots, with something like 4 unused cheaper batteries to hold the structure. I suppose it's better off that it's decent cells, just funny how much space is wasted.

SO.... anyone have any advice on any other tests I can run before I dive into taking apart the batteryand balance charging manually?.... just to make sure the probability that the correct issue is identified before I tear it apart.

Wish I knew more about the controller... it seems very simple (but for that matter pretty clean design)... since I'm drastically upgrading the FETs... almost tempted to upgrade the main Capacitors while I'm at it (I do have some better spares laying around... THO my supply of those is somewhat limited). The controller has 2 solder pads almost joined (seems like an obvious place to short to adjust some setting)... next to it it says "60/120" .... it's a 60v controller by default, I'm wondering if that would enable up to 120V possibly? (then it would make more sense to upgrade the Capacitors with the FETs... but if it's just gonna always be a 60V, I suppose there wouldn't be much purpose... wish I knew who made the controller, or ideally had a data sheet for it... it's a 15 FET, 60V, 2000W... even has a place for a 3rd shunt... seems like it could have the potential to handle a lot more than 2kW... especially if I'm right that the solder pad would turn it to a 120V max... tho that seems like a big jump, but 15fets, and 2 shunts with option for 3rd seems nice.. better than adding solder to them... could make sense. (One funny side note, the throttle connector is labeled "Turn the" ... most of the other labels make more sense, that one was the oddest.... but since i don't know chinese, maybe that's the more literal translation of the term used for throttle, but just guessing)
 
progrock said:
I was hoping that the BMS would kick in on full charge and balance for me
The BMS will balance it...eventually. Remember that most of these have a few mA of balancing current, maybe a few dozen mA at best, to drain the high cells. So they drain a bit over a few minutes to hours, then the BMS allows charging again, then in a few seconds to minutes the high cells reach HVC again, BMS shuts off charging and starts balancing again, and this cycle repeats for hours, days, or weeks, depending on how big the difference is between the high and low cells.

Even the best BMS units have a very small capacity to do balancing, relative to the size of the cells they protect--they're not intended to fix a pack that's got a problem; they're intended to prevent minor imbalances from growing serious enough to result in other problems. But to do that the pack has to be charged fully (and left on the charger to balance, which most people don't do) on a regular basis, and/or has to be built of cells that are already well-matched (not something most pack builders bother with).


I was expecting some really cheap no name batteries.. but instead it has LG 2.600mAh cells,
At least, it has cells *marked as* LG 2.6Ah cells. ;) It doesn't mean they are--there's LOTS AND LOTS of fake cells out there, and probably just as many that might be real but are actually recycled garbage cells literally pulled from trashed used dead packs and resold as new or built into packs that are sold as new. Sometimes they leave the orginal wrappers/markings on them, and sometimes they rewrap/remark them as some "better" type (whatever is popular and sells for more, at the time).

I don't trust that any pack is built of what it says it is, until the actual cells are tested thoroughly and proven to match the cell manufacturer's specifications. ;)



next to it it says "60/120"
hall angle--60 or 120 degrees. unless you havea motor that you know requires the other setting, leave it as it is.

wish I knew who made the controller, or ideally had a data sheet for it..
There arent' any for any of them; sometimes you can figure out who made a controllr by it's case markings, but not usually.


. it's a 15 FET, 60V, 2000W... even has a place for a 3rd shunt... seems like it could have the potential to handle a lot more than 2kW...
Sometimes an "extra" shunt pad set is for a higher current controller, and sometimes it is because they might use three shunts of a higher resistance each, and sometmies a pair (or single) shunts of a lower resistance each.

FWIW, a 2kw controller at 60v is only about 30A.


If the battery can handle more than that, you could take a shunt wire out of a dead controller, and solder it into this one. If there was 30A with two shunts, then there'd be around 45A with three (assuming the third (new) shunt is the same resistnace as the two originals.

More current is harder on a battery, so if it cant' handle it it'll probalby degrade a lot quicker and have much less range than it used to, prior to the shunt mod.
 
Amberwolf, again, thank you.

So... I decided to just take apart the battery again, and inspect it... unless I somehow missed a group, and I don't think I did... it's actually pretty much perfectly balanced.... every group within .02 V of each other... practically all of them showed up as 4.173 V.

AND very good point about the labels totally not confirming legitimate LG cells. I forget how crazy all these fakes are.... I've been using ONLY Sanyo 20700A cells directly from brand new Bosch 6.3Ah 18v packs..... the packs are so beautifully engineered.... and these cells are just incredible.... almost too strong ;) Based on where I am getting those cells from, plus the uniqueness and quality of them.... I never have to think twice about whether they are legit/good cells or not. BUT I've watched many a video, and seen many a discussion about the extensive amount of fake name brand cells... seems like you can't get a good deal on batteries without there being a high chance of them being fakes.... and many/most of these fakes are truly horrifically bad cells, many times having close to 1/10 of the mAh they are supposed to.

Either way, they are apparently balanced, and charged. And while the battery pack is not connected, shows the correct ~65 V charge.... but as soon as they are connected to the controller... drops to below 10 V instantly. I'm still suspecting some kind of short somewhere... just not sure where to look next. Wish I had a spare controller... not to mention a spare engine... to test these things on... still have no real answers... seems like everything I test checks out (many times after spending plenty of time dismantling it. I guess I'll have to do a closer inspection of the scooter.. AGAIN... hopefully something shows up.

All I know is this scooter worked 100% perfectly when my friend got it... was fully charged... and that charge allowed him to do quite a bit of riding before it finally died. It was then left in storage for between 3-6 months... I think closer to 3 than 6.

As for figuring out the maker of the controller... no clue, I def can't read chinese (assuming that's what it is)... The one unique thing I saw on the controller was that the throttle connector is labeled "Turn the" (All labels show presumably chinese, and then I'm assuming their direct English translation)

For now, I'm gonna keep looking around for anything obviously wrong... hopefully I'll make some progress.

EDIT: I might also see if there is any way I can confirm these are legit LG's or not... They are labeled LGEBM261865....As in the high drain LG M26... which I believe should be rated at 10A per a cell.... a 16s6p of those should be a pretty solid battery... and based on my friend's description of how he originally got it, and how long that 1 charge lasted... would make sense.
 
progrock said:
unless I somehow missed a group, and I don't think I did... it's actually pretty much perfectly balanced.... every group within .02 V of each other... practically all of them showed up as 4.173 V.
Easy to be sure you get all the groups. Measure at the BMS connector, and use a pencil or marker to dot off the pairs as you check, and also write down the voltage of each pair as you go, numbered 1 (most negative) to 16 (most positive).

FWIW, 16 x 4.173 = 66.768, so for an unloaded battery not connected to anything, it shouldn't lose over a volt and half between the cells and the BMS output (which you said is reading 65v). (it might if you were testing under a load causing some voltage drop across the BMS FETs, but you're not).


Either way, they are apparently balanced, and charged. And while the battery pack is not connected, shows the correct ~65 V charge.... but as soon as they are connected to the controller... drops to below 10 V instantly.
That means the BMS is shutting down to protect the pack.

This can be from overcurrent (short on the output, or even just too high an inrush current drawn by controller capacitors; fixable with a precharge resistor circuit (lots of examples on the forum under "precharge")

But more likely it is from undervoltage (meaning, once a current is being drawn from it, even a small one), one or more cell groups drops in voltage down to below wahtever the BMS's LVC is.

THe latter is easy to check, just put your meter on the first cell group, then while watching the voltage connect the battery to the bike. If the voltage is stable, that group is probably ok. If the voltage drops, that group has a problem. THen move the meter leads to the next group, and repeat the test (letting the batery reset it's output between tests).




I'm still suspecting some kind of short somewhere...

THat's unlikely, but easy to test. Use your meter on Ohms, and without the battery connected, hook the red probe to the + input of the bike, and the black to the - input of the bike. If the controller is connected you'll get a lowish reading for a moment, then as the caps charge up the reading will increase until it is probably several megohms.

If it stays low, in the tens to hundreds of ohms, you do have a short somewhere, and can start disconnecting things from the + input wire of the bike until the problem goes away. Whatever you last disconnect is where the problem starts.



All I know is this scooter worked 100% perfectly when my friend got it... was fully charged... and that charge allowed him to do quite a bit of riding before it finally died. It was then left in storage for between 3-6 months... I think closer to 3 than 6.

It's bad for a battery to be run to death and then left uncharged; it's likely that one or more groups of cells (probably those closer to the negative battery output as the BMS is usually powered by those) are damaged by the BMS draining them down, and are causing the problem. You can load test the battery (see the many many testing / repair threads) to find the problem.

That kind of thing has killed a number of batteries, which have been diagnosed and tested and somtimes repaired, in various threads around the forum over the years.
 
AS always, thank you for the advice/recommendations. Ha;f the problem is that I am doing the work on the battery and/or controller at my apartment.. while the rest of the scooter is at my friends place. You're probably right about there being an issue with one of the battery groups (or more) under load... tho it'll be good to have a few things to test in case that is not the issue. Sadly, while I have a few dozen different common connectors, the majority of the ones used on this scooter I do not have... so I might end up changing a few connectors, that way I can do a few different tests with other equipment. Sadly I don't have a 16s battery to test the scooter with (that would be ideal to quickly prove it definitely is the battery)... BUT I do have a bike with a controller that can run on 60v at my place... so I might just change the connector on the battery and try it with my bike... at least then I can test the individual groups under load, etc while at my apartment, where the majority of my tools/testing devices are located.

Also, thanks for the advice on checking for a short, and other things... I'll run through the different tests and hopefully have a more conclusive answer of what is wrong. Will post an update once I get through all this.... while the battery is the most obvious place to look, I just want to get a conclusive test to confirm this (TBH was hoping that the batteries weren't gonna be balance when I opened it up again... would be much easier knowing that a few 18650's needed to be replaced and would be done with this.... but hopefully I'll know for sure after these tests)
 
Back
Top