High Output Battery Pack

Brent_C

100 µW
Joined
Aug 1, 2021
Messages
8
Sorry this is so long. The TL;DR version is: I have two motors and a high-output battery pack that seems to be overwhelmed by what I'm doing with it, and I'm not sure the best way to fix it.

Long Version:
I have an Organic Transit Elf that I have modified. I swapped out the stock mid-drive motor with a XOFO DD45_STD motor (from Grin Tech.) and installed a Single Side All Axle motor on one of the front wheels. To support the extra load, I built a 14x4 battery pack of Molicell P42A cells and a Daly 120A BMS. The 21700 cells are rated for 45A continuous output, so I should have 180A of theoretical output, but I limit my output to 100A because that is plenty and I believe the batteries will have a longer service life if I don't drive them to their limits.

I use a Cycle Analyst connected to a Phaserunner to drive the front motor and a Baserunner to drive the rear motor.

(Why do I need so much juice? Well, it is a fairly heavy vehicle and I always travel with my 75 lb dog. And, my home and work are both on top of long hills without much of a bike lane.)

With 26" wheels, and both motors in the 9-10 rpm/V the 52-ish volts keeps me moving right along with traffic, most of the time. And, the output is high enough to handle the load of the hills I need to climb ... but ...

I have my BMS configured to fault-out if:
a) The sum voltage drops below 40V, or;
b) Any cell drops below 2.8V, or;
c) The difference between the lowest and highest cell is > 0.5V.

On a full charge, it performs great!

As soon as I get down to around 75% - 80% charge, it seems that pushing hard up a hill puts one of the cells below 2.8V and the voltage differential close to 1V. I believe this is because it is discharging too rapidly for the pack to balance, so the weakest cell takes a beating and the BMS shuts down before it can rebalance.

When I recharge, everything comes into balance, and ... here's the weird thing ... it isn't always the same cell that drops lowest. That would make sense - just a bad cell. But, it is a different cell everytime (or at least not always the same cell).

So, I have a few questions:

I'm betting that I can configure my Cycle Analyst and/or controllers to prevent this condition. Maybe I just need to limit the current to a lower number. My hill-climbing performance would suffer, but that may be the price I have to pay. Thoughts/advice?

Also, I'm not against replacing the battery entirely if I could find something that can handle that load. (Most use a 40A or 50A BMS, which wouldn't cut it.) Would probably need one custom-made. Like I said, I did this myself, but I suppose it is entirely possible that the quality control of the cells is inconsistent. If someone makes the pack for me, I am not responsible for rebuilding the pack if there is a bad cell - and that's appealing. Does anyone have a recommendation for a company that could sell me a suitable pack? (52V, 100+A output, around 16 Ah)

Is there something else I am not considering?

Any help/advice is greatly appreciated.
 
Brent_C said:
I have an Organic Transit Elf that I have modified. I swapped out the stock mid-drive motor with a XOFO DD45_STD motor (from Grin Tech.) and installed a Single Side All Axle motor on one of the front wheels.

Do you still have the original motor? (I think they used Ultramotors?) If so, do you have any future plans for it?


To support the extra load, I built a 14x4 battery pack of Molicell P42A cells and a Daly 120A BMS. The 21700 cells are rated for 45A continuous output, so I should have 180A of theoretical output, but I limit my output to 100A because that is plenty and I believe the batteries will have a longer service life if I don't drive them to their limits.

This cell?
https://lygte-info.dk/review/batteries2012/Molicel%20INR21700-P42A%204200mAh%20(Gray)%20UK.html

If so, this is a test curve from that page, showing discharges at up to 30A, which should be less than what your cells see. It drops to 2.8v at around 3.7Ah (out of 4.2) at that rate.
Molicel%20INR21700-P42A%204200mAh%20(Gray)-Capacity[1].png


With 26" wheels, and both motors in the 9-10 rpm/V the 52-ish volts keeps me moving right along with traffic, most of the time. And, the output is high enough to handle the load of the hills I need to climb ... but ...

I have my BMS configured to fault-out if:
a) The sum voltage drops below 40V, or;
b) Any cell drops below 2.8V, or;
c) The difference between the lowest and highest cell is > 0.5V.

On a full charge, it performs great!

As soon as I get down to around 75% - 80% charge, it seems that pushing hard up a hill puts one of the cells below 2.8V and the voltage differential close to 1V. I believe this is because it is discharging too rapidly for the pack to balance, so the weakest cell takes a beating and the BMS shuts down before it can rebalance.

When I recharge, everything comes into balance, and ... here's the weird thing ... it isn't always the same cell that drops lowest. That would make sense - just a bad cell. But, it is a different cell everytime (or at least not always the same cell).

WHen you say it comes into balance, do you mean that without the BMS doing any balancing that the cells just bulk-recharge back to the correct top-balanced levels?

Or is the BMS rebalancing them? (and is it a resistive-shunt top-balancer, or a capacitive charge-shuffler type that balances all the time, whether charging, discharging, or just sitting there?)

It's an important distinction; the latter is perfectly normal to end up with a balanced pack at the full end of charge, or some time later. The former doesn't usually happen, where cells just happen to end up balanced at full when they were very different at empty.

If it's the former, what you are experiencing could be an artifact of teh BMS itself either measuring thigns wrong due to noise or connection issues, or problems in it's internal programming or even hardware design. Do you have a way to measure the voltage of the cell groups under those conditions, realtime, with other instrumentation, at the same time as the BMS? (you may need to disable the BMS LVC shutdown condition temporarily for the test so it doesnt' shutdown and remove the load before you can see the results on your secondary measurement device(s)).


I'm betting that I can configure my Cycle Analyst and/or controllers to prevent this condition. Maybe I just need to limit the current to a lower number. My hill-climbing performance would suffer, but that may be the price I have to pay.
Does the CA control throttle input for both controllers? If not, you may need to change the current limit in the other controller itself as well, so load is still distributed proportionally between the two.

There is also a way in the CA to "rollback" throttle (reducing load on the pack from the controller) when voltage sags past a certain point, which means it won't limit current all the time (which changing the current limit does), just during these events. You may also be able to do that in the PR and/or BR.

[ Batt->Vlt Cutoff ]*
Low voltage rollback. When the CA detects the battery voltage falling below VltCutoff, it will gradually scale back the power draw to prevent the voltage from dropping lower. This protects lead acid batteries from sulfation, nickel batteries from cell reversals, and with BMS protected lithium batteries it can keep the battery above the BMS trip point to prevent the battery from abruptly shutting down.



As for replacing the pack, you could just add another row (or more) of parallel cells. Personally I'd recommend doing something along the lines of overbuilding any pack you use so it has more capacity and more capability by 25-50% than you actually need, worst case (high headwinds at full speed on a high load hill/etc, for instance), so that even as the pack ages it will have the capability you need (since it will lose both capacity and capability over time, and the harder it has to be used, the worse this will be and the faster it will happen...the easier you are on it the longer all parts of it will last and the more reliable it will be).


You can also get large-format EV cells, even used most of them are more capable than the cylindrical stuff. I'm using 10-year-old (or older!) EIG NMC 20Ah cells on the SB Cruiser trike (also large, heavy, and used to haul big dogs or full loads of groceries or dog food, or cargo as large as an upright piano, in one case). They aren't made anymore, though some places (like Jimbob01 here on ES) still have some NOS laying around. I like them because they just bolt together with wires or busbars, no fiddling around with DIY interconnect solutions.

Mine I use in a 14s2p setup, for 52v 40Ah; they're 5c cells so even 1p will give 100A when new with just typical voltage sag--mine sag at least twice as much as normal now under high loads, but I only have an 80A setup at present. However, I can still maintain 20mph easy up 7th St on North Mountain up past The Pointe at the top--it's fairly steep and long; don't know the slope but google maps probably has info on that.

However, I'm really thinking of large-EV stuff that places like Batteryhookup and the like get as salvage units. You might have to bypass failed cells, but the modules come factory-assembled and wired, so all you really need is your own BMS for them (whatever BMS they come with probably only works with the original application's other electronics). Sometimes they have other large-format-cell packs, like server rack backup or powerwall units, etc., I don't know if those will be as high-current capable as the other stuff, but most likely they would be.

There are also cells like the Nissan Leaf modules; those are probalby getting up in years, though, too, for the ones I have seen around the forum.
 
What power levels are you seeing normally during ride? To me it seems strange with the cutouts since the molicells are strong cells and your power needs should not be too high, max 4kW? 14x3.6x4x30 is about 6kW so shouldn’t be an issue. (The cell max rating at 45A is not what you can use it at, already at 30A the cell gets to about 50deg C in tests on a single cell and that is higher than what’s good for the longevity)

If the cells are diving to 2.8V during your ride and it’s about the same for all of them then the answer is simple, the pack can’t take the load. What about the original pack, is it still connected?
 
amberwolf said:
Do you still have the original motor? (I think they used Ultramotors?) If so, do you have any future plans for it?

In fact, I do. I originally pulled it out because I blew a hall sensor (a careless short). At the time, I didn't know it was something that simple, and I needed a FAST repair, so I just replaced the motor. So now, it is just collecting dust. Might need a little work, but you're welcome to it. I've held onto it because ... "some day", but if you have an immediate use, I would rather see it go to a good home.

amberwolf said:
This cell?
https://lygte-info.dk/review/batteries2012/Molicel%20INR21700-P42A%204200mAh%20(Gray)%20UK.html

If so, this is a test curve from that page, showing discharges at up to 30A, which should be less than what your cells see. It drops to 2.8v at around 3.7Ah (out of 4.2) at that rate.

That's the cell, and good point about the test curve. I've been doing a little math which pushes the limits of my understanding. (I'm a software guy and know just enough about electronics to build something, blow it up, rebuild it, then hopefully just repeat that once or twice before getting it right.)

In response to Lars' post, I used Grin's simulator (https://ebikes.ca/tools/simulator.h...n=true&add=true&blue=Lbs&hp=0&mid=true&eff=75) and found that the simulated performance pretty well matched what I've been seeing. On a 7% grade, I can maintain about 24 mph. According to the simulator, that should be about 2300 watts.

So, it does seem that the system should easily support that. Right?

If I'm reading the discharge curve correctly, I should expect somewhere around 12 Ah from my pack (14s, 4P) before getting close to that rapid downward curve. Or, put another way, I started up that hill on a full charge and was drawing 40A, I should really only be able to sustain that for about 18 minutes before the battery performance started to rapidly decline to the cutoff point of 2.8V.

And actually ... that's probably not far off from what I'm getting. On my commute, I spend about 7 minutes climbing steep grades, each way. Plus, I have at least 3-4 traffic lights, at least one of which tends to have me starting out up hill. What I am experiencing is that it handles the commute to work, just fine, and gets me ALMOST home before cutting out.

So, I think that my commute has me using 2300 watts for about 7 minutes (my hill climb) and about 1800 watts for about 5 minutes (the rest of the commute). Those are really rough estimates, based on the simulator and my actual speeds/distances.

Given those estimates, I should consume about 418 Wh. (I get there with 7 / 60 * 2300 = 268, plus 5 / 60 * 1800 = 150)

That is a bit over half of the theoretical capacity of the 16Ah battery pack. So, if I did this correctly, the fact that it gets me to work and ALMOST home looks about right.

Am I way off, on my math? That seems REALLY low performance. With a single motor and a 48V, 16Ah battery, I was able to JUST make my 22 mile commute that ended with that same hill. Granted, most of this was out of town, cruising at a constant 16ish mph. My current commute is about 3 miles and is all in town with traffic.

The thing that bugs me about this is that most of the cells have plenty of charge left, both individually and cumulatively when the BMS cuts off because one of the cells has dropped below 2.8V and the differential is > 0.50V. Makes me think my math is off.

amberwolf said:
When you say it comes into balance, do you mean that without the BMS doing any balancing that the cells just bulk-recharge back to the correct top-balanced levels?

Or is the BMS rebalancing them? (and is it a resistive-shunt top-balancer, or a capacitive charge-shuffler type that balances all the time, whether charging, discharging, or just sitting there?)

Yeah, I didn't state that very clearly. It is a resistive balancing BMS. I just meant that I think that maybe the discharge is so rapid that one of the cells is entering a fault condition (low voltage) before the BMS can dial up the resistance fast enough to prevent it. There are things about that that don't sound right, but I'm just making wild speculations.

It's an important distinction; the latter is perfectly normal to end up with a balanced pack at the full end of charge, or some time later. The former doesn't usually happen, where cells just happen to end up balanced at full when they were very different at empty.

amberwolf said:
Do you have a way to measure the voltage of the cell groups under those conditions, realtime, with other instrumentation, at the same time as the BMS?

Yeah, I have access to full data feeds from the BMS, Cycle Analyst, and both controllers. So, I can log what's going on across the board and hopefully see what the culprit is. That's my to-do for this weekend. It would be good to do this anyway, just to see if there are other opportunities for optimization. I can correlate this with my GPS and mapping data to see exactly where I was, how fast I was going, the grade, etc. for any given set of logged data. Should be interesting.

amberwolf said:
Does the CA control throttle input for both controllers? If not, you may need to change the current limit in the other controller itself as well, so load is still distributed proportionally between the two.

I currently have it set to send the same throttle voltage to each controller. (Just the throttle input voltage "cropped" with min/max outputs

amberwolf said:
There is also a way in the CA to "rollback" throttle (reducing load on the pack from the controller) when voltage sags past a certain point, which means it won't limit current all the time (which changing the current limit does), just during these events. You may also be able to do that in the PR and/or BR.

I think that might be my stopgap solution, if my logging experiments don't yield enlightenment.

amberwolf said:
Personally I'd recommend doing something along the lines of overbuilding any pack you use so it has more capacity and more capability by 25-50% than you actually need

Definitely! If I find I need to replace/upgrade the battery, I am going all in. I've been looking at Headway cells, for a while.

Thanks for taking the time for all of the advice! Let me know if you want that motor. I'd be happy to send it to you.

Cheers,
-B
 
larsb said:
What power levels are you seeing normally during ride?

Yeah, I'm going to go out and do some logging to find out what is actually happening. Anecdotally, the power is great ... until I hit the cutoff. I think that the voltage is dropping to around 42V, but I haven't confirmed that. (My cutoff is set to 40V.) I don't yet know that the current is doing. Logging should tell me that.

larsb said:
The cell max rating at 45A is not what you can use it at, already at 30A the cell gets to about 50deg C in tests on a single cell and that is higher than what’s good for the longevity)

That's why I have the total output limited to 100A. It isn't clear to me if it will cutoff immediately at >100A or if there is a "grace period" whereby you can briefly exceed the current cutoff.

larsb said:
What about the original pack, is it still connected?

The original pack has a 40A BMS which cuts out pretty readily, even under relatively modest load, when using both motors. In fact, when I had just one motor (the DD45_Std), I had a problem with the 40A BMS cutting out if I rapidly accelerated up a hill, so I had to dial the controller back a bit to limit performance to avoid the cutout.

But, I have plugged in the old pack and turned off one of the motors (I can turn them on/off independently). It performed, as expected for either motor.

Cheers,
-B
 
Brent_C said:
In fact, I do. I originally pulled it out because I blew a hall sensor (a careless short). At the time, I didn't know it was something that simple, and I needed a FAST repair, so I just replaced the motor. So now, it is just collecting dust. Might need a little work, but you're welcome to it. I've held onto it because ... "some day", but if you have an immediate use, I would rather see it go to a good home.

I cant' say "immediate" but I have a trike (SB Cruiser) that presently uses an old Stromer version of the Ultramotor (gutted the original built-in controller and rewired to use with external controller) on the rightside rear wheel, and I have an A2B Metro version of the Ultramotor to put on the leftside rear wheel (hopefuly next week)...and I had originally planned for one in the front wheel, but that ended up in the right rear when the previous (MXUS 450x) had problems leading to failure of trust. I still have the rim and spokes to put one in the front wheel, and dropouts to build the fork into one that can hold a "rear" wheel axle type if necessary.


(I'm a software guy and know just enough about electronics to build something, blow it up, rebuild it, then hopefully just repeat that once or twice before getting it right.)

I'm a hardware guy and I still do that. ;) (I'm more of an "artistic" type that "sculpts" my way thru projects).

My best math is when I get it right and end up with 5 when adding a couple of pairs. Wait....er... :oops:

So, I think that my commute has me using 2300 watts for about 7 minutes (my hill climb) and about 1800 watts for about 5 minutes (the rest of the commute).
If your CA shunt is between the battery output and both controllers, you can set the main display to show watts, and actually monitor that realtime. Or you can log it's output and look at it later (Grin has a page to help chart the logged output).


Given those estimates, I should consume about 418 Wh. (I get there with 7 / 60 * 2300 = 268, plus 5 / 60 * 1800 = 150)
The CA also logs accumulated Wh internally, so you can measure that with certainty as well even without external logging.

It also gives you a wh/mile calculation, averaged over your trip. If you really do 418wh over 6 miles (3 each way?) that's about 70wh/mile.

Am I way off, on my math? That seems REALLY low performance. With a single motor and a 48V, 16Ah battery, I was able to JUST make my 22 mile commute that ended with that same hill. Granted, most of this was out of town, cruising at a constant 16ish mph. My current commute is about 3 miles and is all in town with traffic.
More motors uses more power, though not as much as you're seeing. But there is a definite difference between traffic riding and constnt riding. On my old CrazyBike2, I would get 30wh/mile+ in traffic, but cruising for long stretches without changing speeds (same top 20mph speed) would get me about 22wh/mile. Pretty big difference.

It's even worse with SB Cruiser, partly because it weighs at least 20-25% more, and partly because it's as aero as a brick with a parachute. (the latter does a bit of evening out between the two riding styles, but the weight makes a big difference during startups from a stop even on flat ground).


The thing that bugs me about this is that most of the cells have plenty of charge left, both individually and cumulatively when the BMS cuts off because one of the cells has dropped below 2.8V and the differential is > 0.50V. Makes me think my math is off.
This is why I recommend using another simultaneous measuring method on the cell groups so you can see if the measurement the BMS is making is real, or an artifact or ghost.


Yeah, I didn't state that very clearly. It is a resistive balancing BMS. I just meant that I think that maybe the discharge is so rapid that one of the cells is entering a fault condition (low voltage) before the BMS can dial up the resistance fast enough to prevent it. There are things about that that don't sound right, but I'm just making wild speculations
.

The BMS does not do anything except turn the discharge port on or off in response to cell-group-level LVC (or other limits). It doesn't change any resistances anywhere.

The resistive balancing only happens at top of charge, during charge. it doesn't do anything during discharge.




amberwolf said:
Do you have a way to measure the voltage of the cell groups under those conditions, realtime, with other instrumentation, at the same time as the BMS?

Yeah, I have access to full data feeds from the BMS, Cycle Analyst, and both controllers. So, I can log what's going on across the board and hopefully see what the culprit is. That's my to-do for this weekend. It would be good to do this anyway, just to see if there are other opportunities for optimization. I can correlate this with my GPS and mapping data to see exactly where I was, how fast I was going, the grade, etc. for any given set of logged data. Should be interesting.

To do this test, you will need some other *cell level* measuring setup. That can be 14 voltmeters, or RC-hobby-type cell monitors (like the Celllog), etc. Best is if you have something that can log the voltages from all of the cell groups at the same time, so that you can compare that to the moment the BMS detects an LVC on a group.

Then you can find out if it is a BMS artifact, or if there is something strange going on with the cells (or their interconnects, etc); if the same cell is detected with the same voltage drop by both, it's real; if not, it's an artifact of some kind.

The CA and controllers cannot help you monitor this. When the BMS LVC kicks in it will shut off power to all of those, so they wont' show you anything useful at that point.

THe BMS data is only useful to compare to a secondary independent measurement of the *same* source (all of the individual cell groups). You already know what it has to tell you by itself--a "random" cell group drops below LVC and the pack shuts down. ;)


If I find I need to replace/upgrade the battery, I am going all in. I've been looking at Headway cells, for a while.
I would recommend something better than Headway; even the used EV-class large-format cells will generally be better (depending on their condition). Headway are "easy to assemble", but I'm not sure they have any other advantage for your application. They are not high-output or high-performance cells, and they will have to have multiple parallel cells (whcih will make for a large heavy pack) to get the kind of output you want/need.

Thanks for taking the time for all of the advice! Let me know if you want that motor. I'd be happy to send it to you.

I could certainly use it once I get that far in the trike project (it is always evolving); we can work out the details via PM.
 
Brent_C said:
That's why I have the total output limited to 100A. It isn't clear to me if it will cutoff immediately at >100A or if there is a "grace period" whereby you can briefly exceed the current cutoff.
Are you using the BMS to do the limiting? If so, it's probably a fairly hard limit before it cuts all power output.

If you use the Cycle Analyst to do the limiting, then it will just adjust the throttle so you don't exceed the limit, though there are gain parameters that affect the system that you may need to play with to avoid surging right around the limit itself.

Similarly, your controllers also have current limiting, and they also will simply adjust their output to the motor to avoid exceeding the limits set in them.

The original pack has a 40A BMS which cuts out pretty readily, even under relatively modest load, when using both motors. In fact, when I had just one motor (the DD45_Std), I had a problem with the 40A BMS cutting out if I rapidly accelerated up a hill, so I had to dial the controller back a bit to limit performance to avoid the cutout.

But, I have plugged in the old pack and turned off one of the motors (I can turn them on/off independently). It performed, as expected for either motor.
If you connect htem in parallel (assuming they are both 14s packs of the same chemistry, and match full-charge voltages) then the smaller pack will supply some of the load, while the bigger one will supply most of it. It may be enough to fix the problem you're having, at least until the packs age enough to sag more.

If the original pack is old enough it doesn't charge to full matchign the new one, you can use large (schottky) diodes between the packs so one can't discharge into the other, if necessary, but generallly it shouldn't be a problem.
 
OK, for anyone who cares ... here is what I did/learned.

I did some tweaking. I set the CA to limit power to 80A (and 4000 watts) and made a few other tweaks that I don't think matter, for this issue. I did some test riding/logging and found everything looked pretty reasonable and I wasn't experiencing the problem anymore. What's more, I couldn't really detect any performance decrease. So, I considered this a win.

Today, I made my regular commute and, right about at the same place, the whole thing dies ... again! But, this time I was (sort of) prepared. The battery was actually fine and balanced - this wasn't the problem. I got out my multimeter and started tracing the problem.

Turns out, I had a 45A connector where I needed something about twice that size. Each controller runs through its own 45A Anderson Powerpole connector. And, I have two main cable pairs coming right off of the battery (well, via a 120A fuse). The intent was for one cable pair to power one controller, and the other to have a split to power the other controller and the 12V converter.

But, I had inadvertently split one cable pair to power both controllers, and the other to power the 12V converter. So, both controllers were being run through a single 45A connector.

The connector had failed, but looked perfectly intact. I had just melted the plastic inside and distorted the contacts. So, it was just sporadically working/failing as it heated up and cooled. It wasn't until it completely failed that I even noticed it. So, that was a good thing.

I expect that replacing the blown connector and correctly wiring my splits will fix this. Parts ordered.
 
:bigthumb:
got to love an easy conclusion and fix :D

You’d want to check that out of balance cell group. If it really is 0.5V out then that is way too much and will cause premature bms cutouts and battery wear. 0.05V is more adequate cell group difference.
 
Brent_C said:
Turns out, I had a 45A connector where I needed something about twice that size. Each controller runs through its own 45A Anderson Powerpole connector. And, I have two main cable pairs coming right off of the battery (well, via a 120A fuse). The intent was for one cable pair to power one controller, and the other to have a split to power the other controller and the 12V converter.

But, I had inadvertently split one cable pair to power both controllers, and the other to power the 12V converter. So, both controllers were being run through a single 45A connector.

The connector had failed, but looked perfectly intact. I had just melted the plastic inside and distorted the contacts. So, it was just sporadically working/failing as it heated up and cooled. It wasn't until it completely failed that I even noticed it. So, that was a good thing.

That's the one serious problem with Anderson powerpole connectors: the (somewhat delicate at PP45 size) plastic case holds the spring which keeps the contacts aligned and pressed together.

A secondary problem that doesn't help is that a portion, probably significant, of the PP45s out there are not genuine, and the plastic is easier to distort under heat or mechanical stress. In my experience the genuine version shells are pretty hard to "pop" apart when they're slid together (forcing them this way requires a flat screwdriver/etc), and can only be slid apart (the way they should be), but the others are pretty easy and don't even require a tool to do it the wrong way (prying apart instead of sliding). The springs are also often not as springy and don't force teh contacts flat against each other against twisting forces applied to the wires.


Regarding mixing up the connections: Not only can the shells be connected to each other in multiple orientations (keying connector sets so they can't be plugged together incorrectly), they also make a number of colors of shells; if you like you can do what I did and use different color pairs for different connections. For example, my lighting system used yellow/green instead of the red/black for my traction system, *and* I rotated the shells 90 degrees on the lighting vs the traction as well, to ensure I could never plug them into each other in poor light. ;)
 
larsb said:
:bigthumb:
You’d want to check that out of balance cell group. If it really is 0.5V out then that is way too much and will cause premature bms cutouts and battery wear. 0.05V is more adequate cell group difference.

What I forgot to mention was that I ordered a new BMS. All indications are that the cells are fine and the BMS is lying. Even if there is a problem with the cells, I'm pretty confident I can't trust the BMS to tell me the truth. Of course, it is entirely possible it isn't the BMS at all, and I just have a poor connection or two, but rewiring a new BMS will force me to revisit all of that too.

I hate to replace the BMS just as a diagnostic measure, but I don't think I will be confident until I do - just because of some seemingly random numbers reported by the BMS.
 
amberwolf said:
A secondary problem that doesn't help is that a portion, probably significant, of the PP45s out there are not genuine, and the plastic is easier to distort under heat or mechanical stress.

Yeah, I'm familiar with that and these are definitely not genuine. I've read, many times, that knock-offs are pretty unreliable, but I guess I had to learn for myself. In any case, I guess I can't complain that the 45A connectors failed under something probably approaching 80A for over a minute or two.

amberwolf said:
Regarding mixing up the connections: Not only can the shells be connected to each other in multiple orientations (keying connector sets so they can't be plugged together incorrectly), they also make a number of colors of shells; if you like you can do what I did and use different color pairs for different connections.

What I decided to do was to use chunkier 50A connectors for the actual controllers, and just use the lighter 45A connectors to supply my 12V converter. That way, even I can't mess it up.
 
Brent_C said:
What I decided to do was to use chunkier 50A connectors for the actual controllers, and just use the lighter 45A connectors to supply my 12V converter. That way, even I can't mess it up.
:)

I use the SB50s on SBCruiser for all the traction stuff that is connectorized (some of it is hardwired). I don't usually recommend them to people because they always complain how large they are...(but they are tougher!).

(The lighting has a separate battery and wiring system)

BTW, the SB shells come in different colors and also different keyings between the housings, if you ever need that.

Oh, and even the clone versions of the SB shells are pretty good, of the different ones I've run across (most of mine are salvaged from old powerchair stuff, and some server UPS units, etc; when I need to I crimp new contacts or lugs to wiring, but I have often been able to just reuse the original wires too).
 
Brent_C said:
With a single motor and a 48V, 16Ah battery, I was able to JUST make my 22 mile commute that ended with that same hill. Granted, most of this was out of town, cruising at a constant 16ish mph.
That's because that 13S 48V 16Ah battery was most likely 5P (3,200mAh cells) powering just one BLDC motor (not 2) and cruising at only 16ish mph the majority of the time. Plus you may have been using more power pedal assist up the hill(s).

If more hills in your other commute you could've brought your charger with you to work and recharged the battery again (in a safe place) to 90% before returning home. That way you would've only needed to charge it to 90% before going to work.

Brent_C said:
OK, for anyone who cares ... here is what I did/learned.

I did some tweaking. I set the CA to limit power to 80A (and 4000 watts) and made a few other tweaks that I don't think matter, for this issue. I did some test riding/logging and found everything looked pretty reasonable and I wasn't experiencing the problem anymore. What's more, I couldn't really detect any performance decrease. So, I considered this a win.

Today, I made my regular commute and, right about at the same place, the whole thing dies ... again! But, this time I was (sort of) prepared. The battery was actually fine and balanced - this wasn't the problem. I got out my multimeter and started tracing the problem.

Turns out, I had a 45A connector where I needed something about twice that size. Each controller runs through its own 45A Anderson Powerpole connector. And, I have two main cable pairs coming right off of the battery (well, via a 120A fuse). The intent was for one cable pair to power one controller, and the other to have a split to power the other controller and the 12V converter.

But, I had inadvertently split one cable pair to power both controllers, and the other to power the 12V converter. So, both controllers were being run through a single 45A connector.

The connector had failed, but looked perfectly intact. I had just melted the plastic inside and distorted the contacts. So, it was just sporadically working/failing as it heated up and cooled. It wasn't until it completely failed that I even noticed it. So, that was a good thing.

I expect that replacing the blown connector and correctly wiring my splits will fix this. Parts ordered.
So, no need to get a different battery pack (or Headway cells) with the correct amperage connector(s). In affect your mistake acted as a safety fuse to prevent what could've been battery abuse, which could still happen with the new (greater) amperage connector(s) :(

Can you use more pedal power and/or a different gear ratio going up hills so as not to be so demanding of your 14S4P P42A pack both going to work and on your return back home ... OR ... recharge the pack to 90% at work before your return commute back home ... :thumb:
Brent_C said:
My current commute is about 3 miles and is all in town with traffic.
So then there is no problem commuting 3 miles (or 6 miles round trip?) with your current 14S4P P42A power-plant with correct connector(s)? You shouldn't need more than a 90% charge (57.4V) and discharge no lower than 43.4V unless your present 14S4P P42A pack has been worked too hard (abuse?) so as to shortened its cycle life. A 14S6P P42A pack should've--would've been a better choice for that 22 mile commute with hilly terrain (if room permitted). Plus recharging at work before heading back home.

If only it could, would, should have been a 14S6P P42A battery ... easy for me to say ... :wink:
 
Back
Top