BMS Bypass OK?

BlueSeas

100 W
Joined
Sep 4, 2019
Messages
238
Location
St Petersburg
The new BMS I'm testing as part of my current build thread requires a charger as part of the process to get it to turn on. Wire it all up, and it's dead to the world until it senses reverse current from a charger.

I'm not keen on this feature. I quess it's OK if the firmware is rock solid. But if the damn thing turns itself off, say due to the low voltage shutdown setting, I don't want to pedal home this 120 lb monster.

Does bypassing the BMS B- and BMS Load B-, with a switch, essentially shorting these connections hurt the electronic switch in the BMS?

Thanks!
 
that damn bms feature

causes many unhappy users


for my herd

often bypass bms power load path

no switch though

load power straight from tabs

only charge power through bms

much smaller/cheaper bms charge only


outdoor discharge

using conservative controller lvc

often safe enough
 
Kcuf,

Think I understand your philosophy. I do want though for the BMS to report cell voltage, preferably by Bluetooth, so I get a fancier BMS.

I suppose though...I could put the load directly to the shunt, and power the shunt from the batteries, leaving the CA3 as the "pack" level monitor, and the BMS only really for charging and cell level reporting. Connect only the charger B- to the load terminals on the output from the BMS. Have to think on that....

Thanks, Bob
 
BlueSeas said:
The new BMS I'm testing as part of my current build thread requires a charger as part of the process to get it to turn on. Wire it all up, and it's dead to the world until it senses reverse current from a charger.

I'm not keen on this feature. I quess it's OK if the firmware is rock solid. But if the damn thing turns itself off, say due to the low voltage shutdown setting, I don't want to pedal home this 120 lb monster.
Well, it *is* a good safety feature, in that it helps prevent damage to the cells that some people have done by riding a pack to "empty" and then waiting till it recovers enough to turn back on, then doing it again and again until the cells pass out from exhaustion. ;)

Does bypassing the BMS B- and BMS Load B-, with a switch, essentially shorting these connections hurt the electronic switch in the BMS?
Assuming B- is the actual battery negative, and Load B- is the actual discharge output negative, then you could short across those (switch, breaker, jumper cable, etc) to bypass the BMS, without hurting the BMS itself. However, you then risk damaging the cells from overdischarge (or whatever it was the BMS was trying to protect the pack from), and then whatever future risks that damage then incurs.

However, if I were to need to do this sort of thing, what I would probably do is not to directly jumper across the FETs, but rather to put a "bypass" switch into the logic lines that turn on the gate signal to the FETs, so that I could just turn the FETs back on. It might require cutting a trace on the PCB to put the switch in there, depending on the design.

If the logic basically grounds a pullup that turns on the gate drive, then the switch just has to disconnect the logic from the gate drive so the pullup does the work of re-enabling the FETs, so you just need a simple SPST toggle (or slide or push) switch).

If the logic goes to 5v (or whatever) to turn on the gate drive, and grounds it to turn it off, then the switch has to disconnect the logic *and* provide a 5v (or whatever) to the gate drive, so you can use an SPDT that uses the common to conect to the gate drive, one pole to connect ot the 5v (or whatever), and one pole to connect to the logic output.

Etc.

If you use a relay to do this, then you cna put the switch anywhere you like (even on the handlebars) and the relay inside the battery.
 
Amberwolf,

I understand all your suggestions. Especially overriding the BMS to Battery destruction!

But without a schematic, and a description of the logic flow in the BMS itself, not sure I'm smart enough to modify the BMS itself to assert control over the FET's. Not sure I can get that from the Chinese designer, although I asked the same generic question to them via email. Will post a response if I get one.

Getting into the internals blindly worries me more than the external switch...I don't see that really blowing up anything. But really don't know....

Even overriding using the "switch" I propose has implications for the voltage and current requirements. But I know those! Even a 15A switch good for 72V gets me home at 15mph using 14 AWG wire.
 
BlueSeas said:
But without a schematic, and a description of the logic flow in the BMS itself, not sure I'm smart enough to modify the BMS itself to assert control over the FET's.
This is really just a matter of finding the FET gate pin, then tracing the PCB routing to the gate driver (chip or transistor), and then finding the input to the gate driver. Then you use a multimeter to test that pin to see what voltage it is at when the output is enabled, and again when it is disabled.

Even overriding using the "switch" I propose has implications for the voltage and current requirements. But I know those! Even a 15A switch good for 72V gets me home at 15mph using 14 AWG wire.

True.

If you need full current, you can just use a wire the same gauge as the discharge wires, attach it to the same point as the battery negative, and then when needed, just connect it directly to the controller negative in place of the BMS negative.
 
amberwolf said:
If you need full current, you can just use a wire the same gauge as the discharge wires, attach it to the same point as the battery negative, and then when needed, just connect it directly to the controller negative in place of the BMS negative.

Thanks! I think I'm doing the equal of this. My original wiring had the BMS first, followed by the CA3 shunt. I'm reversing this, the pack B- to the CA3 shunt first, followed by the BMS. This creates a "new" B- power post in an accessible location without removing the pack. So in an emergency, I just remove the BMS output, and place the battery B- on the power post and get home. Neither the CA3 or the BMS will be able to see current. But I will still have a voltage readout. That works for this situation. And in this case is easy to implement. The issue in this install, the pack B- isn't accessible after installation, without removing the battery.

I still don't like the charge requirement to reactivate the BMS.....
 
BlueSeas said:
My original wiring had the BMS first, followed by the CA3 shunt. I'm reversing this, the pack B- to the CA3 shunt first, followed by the BMS. This creates a "new" B- power post in an accessible location without removing the pack. So in an emergency, I just remove the BMS output, and place the battery B- on the power post and get home. Neither the CA3 or the BMS will be able to see current. But I will still have a voltage readout. That works for this situation.
If you're doing this level of rewiring, why not make the output wire of the shunt (to the BMS) be the connection point for the movable cable from controller negative, rather than the input wire of the shunt (from the battery)?

That way you still get to monitor current, see how many wh and ah you're using, etc., so you can also monitor how much you're pushing the pack past that shutoff point, etc. ?

I still don't like the charge requirement to reactivate the BMS.....
For users that "know what they're doing", it's not really a necessary requirement. But for the average person, it's probably safer to do this than to auto-reset...because as you can see in numerous battery / bike / system troubleshooting threads over the years, lots of people will ride their battery to death by pushing it until it shuts off, then backing off a bit till it lets them use it more, then pushing it to shutoff, and repeating that until it simply doesnt' recover, and *then* they recharge it (because they have no choice at that point...if they did, they'd just keep pushing it). (in fact, some of them actually bypass the BMS at that point to continue using the cells until they are *really really* dead and really damage them good...because they don't understand how this stuff works).


There is likely a reset available somewhere in there, even if it's under potting/etc., or as I noted in another thread to someone else, you could look up the chip they use, find it's spec sheet, locate the reset pin, and make a switch to ground that pin momentarily, which would then reboot the BMS, presumably leaving it in a normal state at that point (depends on how they wrote their software).
 
amberwolf said:
If you're doing this level of rewiring, why not make the output wire of the shunt (to the BMS) be the connection point for the movable cable from controller negative, rather than the input wire of the shunt (from the battery)?

That way you still get to monitor current, see how many wh and ah you're using, etc., so you can also monitor how much you're pushing the pack past that shutoff point, etc. ?

For users that "know what they're doing", it's not really a necessary requirement. But for the average person, it's probably safer to do this than to auto-reset..).

This is a good suggestion, this build in process, so no actual re-wiring yet. Just the wiring notes in a design file I keep. Take a look:

AF9A32BF-EDE2-4D67-B599-91E69F31DB73.jpeg

I can see where some eBikers might not realize they "have" to pedal home, or call for transport, or end up needing a new battery....

I believe this BMS has 2 adjustable levels of low voltage cutoff. The first level recovers after the voltage recovers to a set recovery point, the second level turns off the BMS. This second one is important too, more than one pack has also been destroyed by the parasitic load of a BMS while in storage. And I can see where a charger is a logical requirement to restore BMS function if it shuts down on an extended low voltage event.

Ideally if it could "remember" why it shutdown, treating the B+ protection board power lead OFF, different than the low voltage OFF.

The problem maybe hardware. Once the BMS CPU is shutdown, which you want in either case, there isn't anything alive to see the voltage restored to the protection board. Although, this BMS does not have a separate input for the charger, I'm not exactly sure how it is completely dead while waiting for a reverse current flow event to reactivate. Maybe a momentary connection across the FET switch would wake it up? It might also let the smoke out....
 
I had a thought (quick, get the fire extinguisher! ;) ):

This *should* work, if the BMS shutdown is the reason I think it is (it turning off it's own internal power converter, which is what keeps itself "on", kinda like a latching relay setup).

Wire a NO SPST momentary switch from main cell battery positive to the charge port positive. You can install a large-value (10kohm+) resistor in series with the switch if you like; this will ensure there are no significant currents flowing. Probably just a 1kohm would be sufficient for that, possibly even less.

It is possible (but unlikely) that you may *also* need a switched wire from main cell battery negative to the charge port negative, in which case you'd use a NO momentary 2PST (DPST) switch. Depends on whether or not the BMS has any internal ground from it's own internal power converter to the main cell battery negative or if it only goes thru the charge / discharge FETs.
 
amberwolf said:
I had a thought (quick, get the fire extinguisher! ;) ):

This *should* work, if the BMS shutdown is the reason I think it is (it turning off it's own internal power converter, which is what keeps itself "on", kinda like a latching relay setup).

Wire a NO SPST momentary switch from main cell battery positive to the charge port positive. You can install a large-value (10kohm+) resistor in series with the switch if you like; this will ensure there are no significant currents flowing. Probably just a 1kohm would be sufficient for that, possibly even less.

It is possible (but unlikely) that you may *also* need a switched wire from main cell battery negative to the charge port negative, in which case you'd use a NO momentary 2PST (DPST) switch. Depends on whether or not the BMS has any internal ground from it's own internal power converter to the main cell battery negative or if it only goes thru the charge / discharge FETs.

There is no Charge Port Positive, only a negative, which is shared with the Load Port Negative.

In the picture, the 4 small black wires on the left are thermistors, followed by the B+ power lead, then the 25 cell taps from C24+ to C1-. The final two connections, B-, also to C1- of the pack, and P- which is both load and charge negative. Those last two a generous pair of AWG 7, that required a little trimming to fit inside a single AWG 4 lug for connecting.

I like the resistor idea, I bet that momentary connection between P- and B- will work. That circuit would be incredibly fragile if a momentary 10K resistor there hurts anything. For sure, the fire extinguisher won't be needed. Might need a slightly lower value.

For bonus credit, find the fault in the silk screen printing on the BMS cover. Hint, you don't need to be able to read Chinese.

E6DAA397-5755-453E-A7E3-387F84E95746.jpeg
 
no b22
 
BlueSeas said:
There is no Charge Port Positive, only a negative, which is shared with the Load Port Negative.
Then you can do what Fechter did:
https://endless-sphere.com/forums/viewtopic.php?p=1613516#p1613520
which is the same idea, just different wiring:
fechter said:
I've had a couple of BMS boards that would hang when triggered and you had to charge through the discharge port to get it turned on again. Completely unplugging it and reconnecting would also reset it.

Most BMS boards have the P+, C+ and B+ all connected permanently together. Switching all happens on the negative side.

I did something like Amberwolf suggested, where I installed a push button switch and a 1K resistor from B- to P-. When the BMS trips, now I can just push the button to reset it. Saved me more than once.
 
I tried 5K, 1K ohms and finally just a piece of wire across B- and P-. That will not wake up the BMS. It takes an external power source inserted between P- and B+, load side of the BMS, that is around 1V more than what's on B- and B+, the battery pack side of the BMS. So Basically a charger.

When testing, at 1V greater than the protected pack, the BMS wakes up, essentially no load on the new power source for a couple seconds, then after the BMS boots, charging begins.

If it was AC voltage, you could use the pack to power a transformer wound to boost the voltage 2%. This also becomes an isolated source. Then use that to kickstart the BMS. But I can't think of a simple way to do that with DC power?
 
why i prefer bypass these bms

too goofy for operational reliability
 
BlueSeas said:
If it was AC voltage, you could use the pack to power a transformer wound to boost the voltage 2%. This also becomes an isolated source. Then use that to kickstart the BMS. But I can't think of a simple way to do that with DC power?
Maybe one of those little low-current DC-DC boards that can start off even your "dead" battery voltage, wired thru a switch to the cell pack + and - so you can turn it on when needed to reboot the BMS. Output of the board can be connected thru diodes (to prevent leakage back into it's output draining anything from the pack when just left for a while) and that 10k resistor (to prevent pulling too much current out of the DC-DC) to the charge port of the BMS. Mmmm...not sure this will work because the output will end up connected thru the BMS back to it's own input. :(

Ah! but wait...it's not as simple, but if you put some big capacitors on the input of the DC-DC, with a precharge resistor on the input to them from that switch (so it never damages th switch), and *also* have a switch on the output of the DC-DC, so it isn't connected to the charge port until you want it to...then first you switch on the input (momentary contacts, so you can't accidentally leave it on), holding the button down for however long it takes to charge those caps up. Then let go of taht button, then press the output-to-charge-port button until the BMS reboots.

Some of these DC-DC boards are really small, and some of them even have adjustable output voltages (and autoadapting inputs within a reasonable range). I don't have a link to any, but poeple have posted about some of them in various project threads, mostly found on aliexpress and the like.


Or just stick a little 3v coin cell lithium battery in there to be switched in series with the battery's own voltage to the charge port wiht a momentary pushbutton (and one of those resistors). (dunno if this would actually work...but it sounds like it should).
 
amberwolf said:
Or just stick a little 3v coin cell lithium battery in there to be switched in series with the battery's own voltage to the charge port wiht a momentary pushbutton (and one of those resistors). (dunno if this would actually work...but it sounds like it should).

I thought about using an AA battery. Or even a 9V battery. So I stack that above the regular B+, so B- to B+ plus AA, is 1.5V higher. B+ to B- is regular pack voltage. Now injecting the AA battery, P- to B+ is pack voltage plus 1.5V. With the BMS output switch OFF, all is copacetic. Hold this too long, the BMS turns on, and connects B- to P-, then is there is a dead short between B+, and B+ plus the AA battery?

If you remove the BMS, and connect the B+ normal to (B+ + AA battery), there is a direct short across the AA battery.

There is potential here. Need to think more.

The other factor...there is actually a load on P- to B+ at all times. The motor controller capacitors are always present. The BMS leaks. Even with BMS discharge OFF, P- to B+ reads about 24V, not zero. So those caps have a partial charge. Not sure how this impacts things.

On your other idea, I did a brief search on DC to DC converters. There are downconverters galore, taking 72+V to 12/24V. However, there aren't many taking 72V to a higher voltage. The only one I saw went to like 120V.
 
amberwolf said:
Or just stick a little 3v coin cell lithium battery in there to be switched in series with the battery's own voltage to the charge port wiht a momentary pushbutton (and one of those resistors). (dunno if this would actually work...but it sounds like it should).

I thought about using an AA battery. Or even a 9V battery. So I stack that above the regular B+, so B- to B+ plus AA, is 1.5V higher. B+ to B- is regular pack voltage. Now injecting the AA battery, P- to B+ is pack voltage plus 1.5V. With the BMS output switch OFF, all is copacetic. Hold this too long, the BMS turns on, and connects B- to P-, then is there is a dead short between B+, and B+ plus the AA battery?

If you remove the BMS, and connect the B+ normal to (B+ + AA battery), there is a direct short across the AA battery.

There is potential here. Need to think more.

The other factor...there is actually a load on P- to B+ at all times. The motor controller capacitors are always present. The BMS leaks. Even with BMS discharge OFF, P- to B+ reads about 24V, not zero. So those caps have a partial charge. Not sure how this impacts things.

On your other idea, I did a brief search on DC to DC converters. There are downconverters galore, taking 72+V to 12/24V. However, there aren't many taking 72V to a higher voltage. The only one I saw went to like 120V.
 
Back
Top