Calling All Experts->Unsolved Mystery: Sudden Battery Voltage drop to Empty

[sfiii]

Okay, I've got the puzzle for you! I"m an experienced e-trike builder, been doing it since 2006. I keep a detailed log so buckle up.

Current set-up: one power main goes to 2 Phaserunner motor controllers to 2 hub motors.

I generally run off of a pair of LiFEPO4 bricks at one time. They run in parallel and so are connected to each other. They are exact twins in every way possible. Each pack has 25A fuse on the positive line. I have experienced a sudden failure twice over the last 2 years with my main pair and with no warning. The second time is the bigger mystery.

New pair of 36V 30Ah batteries with 50A BMS' from known Amazon seller incorporated
into my personal trike power system on 10/6/18. Also, due to ahem a little accident involving saltwater, also new motors and motorcontrollers. New system running at max 50A, vs old system was closer to 42.

All is well for one year. On 10/9\19, driving on the flats at full speed at dusk, about 2.5 miles into the journey, sudden system failure. V flashes on the CycleAnalyst (CA) display and the the whole system turns off as I coast to the shoulder of the road.

As it turns out, my automotive key switch (rated to like 35or 40A) has fried, after 4.5 yrs of faithful service. To be clear, I was running ALL power through that switch.

I bypass the switch, but the CA prevents me from turning the system back on, because one of the batteries has been suddenly drained to below 29V.

Upon closer examination, I find that one of my two batteries has one of its two Anderson connector pins pushed back, preventing full contact on both pos and neg terminals. I am able to limp home on the remainimg battery. I failed to record which of the two batteries drained, whether it was the one with the bad connection or the one with the good.

After recharging them, the Left Battery was showing 39.7V at rest and 36.2V under load, and the Right battery was showing 40V at rest and 37V under load. This is before extended cell balancing. Four months later, using only the CA as a measuring device, the L batt showed 0.77 ohm resistance and likewise the R batt the exact same when connected separately. I did not document if this was at rest or under load. As you can see, my concern at the time of the incident was that that connector problem on one could have caused the twins to have an imbalance.

This first time, I am theorizing that the switch frying caused the sudden battery drain of the only battery that was solidly connected. But it is a chicken and egg problem as I dont know what happened or what caused what, only that there was no overheating issue or excessive system loading at the moment of failure (at full speed on the flats I am drawing about a third to half of max Amps)..

What happened? Your theories and explanations are very welcome here.

Fast forward now to the bigger mystery.
I installed a new Glide Cruisers switch which has a proper 50A relay and off I went, fast forward ANOTHER year of frequent journeying ( my Frankentrike is my principal mode of transport) with no issues.

On 11/3/20, I am on an unplanned (yes) bike camping trip. So i have borrowed a sleeping bag and I dont have chargers with me, just barely enough power for the total trip between my main set and my backup set.

I sleep overnight in the bag, no tent. It is California after 9mos of no rain so everything is bone dry even the sleeping bag has hardly any dew on it at dawn. I am 35mi away from home. I bid friends adieu (need to get home in time to vote!) and start along a road with gentle rolling hills. I am running at 25A at 15mph (vs 50A and 24mph) to conserve power. About 3miles into the return journey on a mild uphill running at half throttle with barely warm batteries, V starts flashing again oh no! total power failure. CA turns off.

I check connectors, seem okay no signs of sparking and pin alignment is solid. I try each battery separately and neither one can power the CA. Thats bad! The pair was at 37.9Ah when they failed and they showed a normal voltage before the sudden fail. Recall that they are 2x30 so thats 60Ah of range and I regularly get a comfortable 56 Ah out of them in the rare moments that I take them all the way.

Everything including the switch still works so I plug in my backup pair of batteries and pedal hard to get home with half of the banked power available.

Upon returning home, I get out the multimeter but its Radio Shack ass has finally kicked the bucket. I order a Fluke off ebay and charge all the batteries, keeping a close eye for overheating or swelling.

Fluke arrives a week later, by then cells are fairly balanced.

When I plug in my back up pair, the Phaserunner LEDs remain solid, as they should;, but when I plug in either my L or R battery that failed on the trip, the Phaserunners flash a 1—6 pattern, which is supposedly a Motor Hall Sensor Fault. Huh?

However, they are now able to power the CA, lights and horn so I decide to take them on a test run around the block. Both batteries work as tested separately, and after each is run for a minute, the flashing error code stops and is replaced by a solid ' on' LED.

The R battery measures 42.6A via Main or BMS wires after top up and the L battery measures 42.5A via Main or BMS wires after top up. The R later showed 39.9V at rest and 36.2 V under 25A load with .127 ohm resistance showing on CA; the L similarly showed 39.9V at rest and 36.1 V under 25A load with .126 ohm resistance. At 41 cycles, this is still a fairly new set of batteries.

Okay experts, what happened, especially the second time? I had ZERO equipment failures and no fuses blew. Why did my remaining 18ah of range suddenly evaporate in a flash? I think this could be any ebikers worst nightmare. It would be like filling up your gas tank and getting out to the middle of Death Valley and then discovering that your tank had a hole in it and your tank is empty!

I am not an electrical expert, but I'm not dumb and I've been dinking with this tech for 14 years. I am at a loss.

Segue

 

[goatman]

sounds like a bms issue

1-6 sounds like one of your motor cables needs replacing or a bad connection at the connector throwing an intermittent fault code. if i get 1-6 with a bad cable and no-go, i unplug battery, wiggle motor cable, plug battery back in, 1-6 is gone, go home and replace cable, no more problems, 2 completely different phaserunner bikes i have showed that symptom and new cable solved it each time

 

[jonyjoe303]

are you able to check cell voltage in realtime while charging? I always had balance issues with my lifepo4 at the top end. If just going by overall voltage when charging, if one cell is higher then the rest, you'll reach full voltage, and the battery will not be anywhere near full.
Are you charging to 3.65 per cell? by quick calculations from your figures I got a max of 3.55 volts when you top off the pack. Maybe thats not high enough to properly balance the pack.
Sudden battery voltage drop to empty with lifepo4 from my experience is due to an undercharge battery.

 

[sfiii]

joneyjoe303, It is quite possible that my cheap Chinese chargers are not doing what's optimal. I DO have a single Cycle Satiator and could be individually topping off and balancing packs to a custom maximum.

I have not yet broken into the pack to see if the cells are currently balanced, because as I mentioned, I have been getting a full expected range out of them so have had no reason to believe they were signicantly undercharged. Additionally, when my LiFEPO4 set does get low, they behave differently. It's true that this is usually only near the bitter end for 2-3miles, but there IS a distinct drop in voltage under load. I didnt see that happening this time. Maybe it was off my radar due to the fact that I was trying to conserve energy?

I find it interesting that both times the system failed, it was about 3miles into the journey. Neither time did I have any reason to believe I was near the end.

I just had the humorous thought that maybe i topped off the wrong pair, which in theory could happen now, but when the failure happened the first time, I only had one pair on board.

So while I am not clear how an imbalance would have a 1 in 100 chance of suddenly depleting the batteries so low that they wont even power the Cycle Analyst, I will definitely look into greater topping off and balancing care.

 

[sfiii]

sounds like a bms issue

1-6 sounds like one of your motor cables needs replacing or a bad connection at the connector throwing an intermittent fault code. if i get 1-6 with a bad cable and no-go, i unplug battery, wiggle motor cable, plug battery back in, 1-6 is gone, go home and replace cable, no more problems, 2 completely different phaserunner bikes i have showed that symptom and new cable solved it each time

on the first comment re:bms, please elaborate.

in regards to the motor cable suggestion, if this were the case, why did the system work fine with my back up pair? Additionally, even 5 days later after the failed set had been recharged, the Phaserunners still 'recognized' these batteries, individually and as a pair, and made the fault code flash. Only after I drove on each battery for about a block or two did the fault code stop flashing for that particular battery. The back up pair never made the Phaserunners flash, not at any time.

Finally, since both Phaserunners are flashing and stop flashing at the same time, the only place the connector fault could be is not a motor cable, but only at or before the power splits from one into two to go to each motor controller.

Are you suggesting that a lapse in power due to a lousy (anderson) connector could trigger the fault code? If thats the case, why would that cause the battery to drain? Where would the stored energy go?

 

[goatman]

fault code could be when your messing with the plugs and moving the harness, if there is an issue with motor cable youll notice the motors might start giving a different sound/feel, if youre not experiencing that then???

i dont do bms, this is where amberwolf would step in and start talking about gate drivers and other things i know nothing about

 

[sfiii]

fault code could be when your messing with the plugs and moving the harness, if there is an issue with motor cable youll notice the motors might start giving a different sound/feel, if youre not experiencing that then???

i dont do bms, this is where amberwolf would step in and start talking about gate drivers and other things i know nothing about

Well one of the two motors that Justin sold me has always made a weird sound everytime after it warms up a little, and I have had a problem with the two motors not putting out equal power always since the beginning of this new getup in 2018, much to my annoyance (veers slightly to the left).
However, I dont see how a fault with one motor would make BOTH phaserunners flash with the same code? I guess technically theyre connected thru the power main (2 batt to 1 line with CA shunt then to 2 controllers) but to me that would negate the point of having two motor controllers to manage two motors if they are not acting independantly of each other.

Its true that I never had this problem with the old equipment--the old motors, controllers, old drainbrain and LIFEPO4 batteries similar to what I use now. 10years with the old equipment this never happened and now twice in two years

 

[sfiii]

are you able to check cell voltage in realtime while charging? I always had balance issues with my lifepo4 at the top end. If just going by overall voltage when charging, if one cell is higher then the rest, you'll reach full voltage, and the battery will not be anywhere near full.
Are you charging to 3.65 per cell? by quick calculations from your figures I got a max of 3.55 volts when you top off the pack. Maybe thats not high enough to properly balance the pack.
Sudden battery voltage drop to empty with lifepo4 from my experience is due to an undercharge battery.

I replied to you but did it wrong so just a little note to let you know to look at thread and fell free to comment further if you like. Thanks for weighing in already!

 

[amberwolf]

So while I am not clear how an imbalance would have a 1 in 100 chance of suddenly depleting the batteries so low that they wont even power the Cycle Analyst,

The batteries only have to have a single cell (group) drop below the BMS's LVC to cause the BMS to shut it's output off, which will then be unable to power anything as it will not provide any voltage (with zero load on the output, not connected to anything but a voltmeter, there might be a "ghost" voltage, but as soon as any load is applied, that drops to nothing).

If the BMS is of a type that only reenables the output after a charger is plugged into the BMS, then without doing that it will appear to have "died", even if it's just a single cell that dropped below LVC for just an instant, and then recovered as soon as load was removed to appear to be the same voltage as all the others (because the cell has higher internal resistance than the others).

 

[amberwolf]

in regards to the motor cable suggestion, if this were the case, why did the system work fine with my back up pair? Additionally, even 5 days later after the failed set had been recharged, the Phaserunners still 'recognized' these batteries, individually and as a pair, and made the fault code flash. Only after I drove on each battery for about a block or two did the fault code stop flashing for that particular battery. The back up pair never made the Phaserunners flash, not at any time.

That sounds like the batteries causing the error were *higher* in voltage than the ones that didn't, and the fault code is specious, actually meaning the batteries were above some limit in the PR firmware but not one that causes shutdown. Thus when the batteries run down a little, the code stops.

 

[amberwolf]

Anyway, what you need to do is to do a balance voltage check, at the BMS balance connector, on all packs (both the "working" and the "not working as expected"), in several states:

First, at full charge, if they're already there, preferably before it has sat there balancing, just when it has finished the main charge. Just test each cell voltage at the balance connector (preferably at the actual solder joints on the PCB, so you know if there is a connector problem--testing at the back of the connector doesnt' tell you what the BMS actually sees). If it's already balanced you won't see the problem, whcih isn't useful.

If you don't see any voltage differences between cells, then you'll need to do more extensive testing, below:

Second, during discharge at the current you typically see problems at, or above, for as much of the discharge curve as is practical. You may have to setup a static load (heater, etc) to do this test "in the lab", unless you can setup the trike offground and use something to put a load on the motor wheel(s) like your mechanical brakes, sufficient to cause that kind of current draw while testing all the voltages.

Third, at full discharge, after the BMS has turned off the output.

Fourth, during recharge, for as much of the charge curve as is practical.

These four tests will help you see which cells have higher internal resistance, which tells you which ones are a problem. During discharge, the ones that drop lower in voltage have higher resistance, and during charge the ones that rise higher in voltage have higher resistance.

For this kind of extensive testing, it is probably easier to get RC hobby balance checkers that display multiple voltages at once (like the Battery Medics, or simlar), and then wire up a harness that connects those to the same type of connector as the BMS balance connector, with a female on one side and male on the other, so it just "taps into" the balance lines.

But you can do it with a regular single voltmeter, it's just a PITA to keep moving the probes from one pair of balance wires to the next for the entire discharge and charge cycles, while either videoing it or writing down the voltages so you can later compare them for each pass across the connector.

 

[amberwolf]

I find it interesting that both times the system failed, it was about 3miles into the journey. Neither time did I have any reason to believe I was near the end.

Is there a hill or slope or wind or anything else at that point that increases the load on the batteries, above what it was for the rest of the journey?

 

[amberwolf]

1. If it was a BMS issue, why did both trip simultaneously?

If both were paralleled, and one pack was for whatever reason shutting off it's BMS output, then the other pack would suddenly have to handle the full load (when it had only been taking half), and if it then had a cell that sagged low enough to hit BMS LVC, that pack's BMS would also shut off. The same would happen if the spike in current was enough to trigger the BMS overcurrent.

 

[amberwolf]

3. If I tripped the BMS, why didn't I need to do a 'Force Start' to get them going again? I just charged the packs normally and then it was as if nothing had ever happened.

Are there separate charge and discharge ports? If so, then the charge port would always be "on" when the pack is above it's safety-minimum-charge cutoff point, even when the pack has shutoff the discharge port due to overcurrent or LVC.

Even fi they're they same port, the charge FETs allowing current into the pack would still be "on" under those conditions, and the discharge FETs in series with them, though "off", would still have their internal diodes allowing current flow *into* the pack (just not out of it).