Even Newer 4 to 24-cell Battery Management System (BMS)

Hi Bobmcree:

I purchased and received your Variable Power Supply Kit some time ago. At that time you promised to provide complete assembly
details complete with photos on bobmcree.com. I searched your website but could not find these instructions and photos. I now have an urgent need for this variable power supply as my Ping v1.0 24v10ah batt konked out on me. If you still have these instructions and photos could you supply them please???
 
Nate said:
Hi Fletcher. Sorry, I thought I posted the following but it was 3 or 4 in the morning and I rewrote the post four times. I forgot to add this:
I can get the main LED to go green using each opto, works correctly. The all shunts line latches correctly and a reset is required.

It is my understanding that without the battery hooked up the LEDs should flash then go off and the main goes green. All my LEDs stay on and the main is very red, no green tone at all.

I'd guess either your power supply voltage is just a tiny bit too low or the current is limited to less than the shunt current.
 
The current is 3.5a but the BMS is only taking .5 - .6. Wouldn't the Main LED be yellow at least? It is completely red and each channel is 3.7v and some are 3.8v... Not sure if I can up the voltage any more. This is a ping charger.
 
I'd guess the Ping charger voltage is set to max out just below where the circuit can trip. Just for a test, try putting a small battery (1.5v or 3v) in series with the charger output to boost the voltage.

There might be a voltage adjusting pot inside the charger.
 
I tested it up to 63v, no difference... Any other ideas? Thanks.

*Edit: I just tested with the battery connected. I got it up to 60v before turning it off. channels 1,3,5,7 and 9 stay at or below 3.5 so the LEDs won't light up. Most other channels are at 3.9 and channel 15 gets up to 4v! This thing is an oven! I could get a serious burn on it if not careful. I know this will get better after the battery balances out. Now here is the weird thing. I disconnect the charger and the main LED goes off but the other LEDs stay on. Is that supposed to happen? I'm stumped...
 
That's a problem. The shunts are coming on, but not sending the signal to cut off the charging current when they are maxed out, which allows the cells to over charge a bit. The shunts will stay on to drain off the excess voltage, then the cell LEDs will go out.

Yet, when you short the pins on the output side, it seems to work properly and the main LED goes green.

Any chance the optocouplers are installed backwards?

Edit: I can see from your picture that you have them installed properly.
 
Just rechecked the optocouplers again to be sure, they are fine. I left the battery on until it drained and the LEDs turned off. Should I be getting 3.75 - 3.8v per channel without the battery connected? Hoping someone can test this with theirs. Thanks.

Edit: Still need help... Is there another way to test the optocoulpers? What else can I do? Please, I need to get this working. Thanks.
 
M Goodrum or M Fechter,

Connected them on the transistor of opto " ebrake signal " and " controller ground " to manage the LVC on the BMS, are to plug on the GND and the BK of the controller Infineon for to activate the regen batterie when tension reached the doorstep of 2.1V ?
Exact or not exact ?
Thanks,
Maxime
 
mimichris said:
M Goodrum or M Fechter,

Connected them on the transistor of opto " ebrake signal " and " controller ground " to manage the LVC on the BMS, are to plug on the GND and the BK of the controller Infineon for to activate the regen batterie when tension reached the doorstep of 2.1V ?
Exact or not exact ?
Thanks,
Maxime

Yes I think you have it right. When any cell drops to 2.1v, the opto comes on and activates the brake line, which stops the controller output.
 
Nate said:
Just rechecked the optocouplers again to be sure, they are fine. I left the battery on until it drained and the LEDs turned off. Should I be getting 3.75 - 3.8v per channel without the battery connected? Hoping someone can test this with theirs. Thanks.

Edit: Still need help... Is there another way to test the optocoulpers? What else can I do? Please, I need to get this working. Thanks.

You have me pretty much baffled by that one.
Yes, you should have around 3.7v per channel with no cells connected. If the circuit goes into throttling (main LED yellowish), the voltage readings may be off by quite a bit due to the switching.

With the circuit powered, try measuring the voltage from the control ground to the 12v bus line and verify you have 12v. Then try shorting pins 7-8 on any opto and see if the 12v remains at 12v while the pins are shorted (seems like it would have to if the main one switches to green).

Another test would be to measure the voltage on the any shunts line against the control ground when none of the channel LEDs are lit, then increase the voltage until at least one channel LED is fully lit and see what happens to the voltage. It should drop from around 12v to less than 1v. You can measure across R6 for this voltage.

When a channel LED is fully lit, also try measuring the voltage across pins 1-2 on the opto for that channel. The opto will turn on at around 1v.
 
fechter said:
mimichris said:
M Goodrum or M Fechter,

Connected them on the transistor of opto " ebrake signal " and " controller ground " to manage the LVC on the BMS, are to plug on the GND and the BK of the controller Infineon for to activate the regen batterie when tension reached the doorstep of 2.1V ?
Exact or not exact ?
Thanks,
Maxime

Yes I think you have it right. When any cell drops to 2.1v, the opto comes on and activates the brake line, which stops the controller output.

Hi Goodrum,

I can connect the transistor of the opto ILD2 directly, BK and GND :

1) to the yellow wire of the brake handle,
2) to GND

without particular precaution?

When do I brake, the switch of the brakes handle consignment a signal to cut the controller and to stop the motor, doesn't the transistor of the opto risk to receive an inverse tension to its terminal ? Because I don't know tension sent by the switch of the brake handle, it is can be the GND.....

Thanks,
Maxime
 
fechter said:
You have me pretty much baffled by that one.
Yes, you should have around 3.7v per channel with no cells connected. If the circuit goes into throttling (main LED yellowish), the voltage readings may be off by quite a bit due to the switching.

With the circuit powered, try measuring the voltage from the control ground to the 12v bus line and verify you have 12v. Then try shorting pins 7-8 on any opto and see if the 12v remains at 12v while the pins are shorted (seems like it would have to if the main one switches to green).

Another test would be to measure the voltage on the any shunts line against the control ground when none of the channel LEDs are lit, then increase the voltage until at least one channel LED is fully lit and see what happens to the voltage. It should drop from around 12v to less than 1v. You can measure across R6 for this voltage.

When a channel LED is fully lit, also try measuring the voltage across pins 1-2 on the opto for that channel. The opto will turn on at around 1v.

It has not done any "throttling" yet, always Red or Green.

The 12v bus is only 11.4v? Could this be the problem? I short the opto and it stays at about 11.35v. Latched it is about 11.35v.

I don't have a variable power supply so I can't test the any shunts as described above... I did test the voltage a different way: No battery all LEDs on 5v, opto shorted .65v and latched 1.4v. I opened up the ping charger and there are no pots. My Soneil 4808srf is down for the count at the moment.

I can measure the voltage across pins 1 and 2 of the opto. Without battery connected all the LEDs are on and main is solid red, some are above 1v some are as low as .85v.
 
mimichris said:
I can connect the transistor of the opto ILD2 directly, BK and GND :

1) to the yellow wire of the brake handle,
2) to GND

without particular precaution?

When do I brake, the switch of the brakes handle consignment a signal to cut the controller and to stop the motor, doesn't the transistor of the opto risk to receive an inverse tension to its terminal ? Because I don't know tension sent by the switch of the brake handle, it is can be the GND.....

Thanks,
Maxime

You need to observe polarity. Yellow goes to BK, black goes to GND.
Both the opto transistor and the brake switch are doing the same thing by essentially shorting the brake wire to ground. You can connect them in parallel with no problem. The open circuit voltage should be 5v.

In some systems, the brake switch is used to power a rear brake light. In this case, you would not want to connect directly to the brake switch.
 
Nate said:
It has not done any "throttling" yet, always Red or Green.

The 12v bus is only 11.4v? Could this be the problem? I short the opto and it stays at about 11.35v. Latched it is about 11.35v.

I don't have a variable power supply so I can't test the any shunts as described above... I did test the voltage a different way: No battery all LEDs on 5v, opto shorted .65v and latched 1.4v. I opened up the ping charger and there are no pots. My Soneil 4808srf is down for the count at the moment.

I can measure the voltage across pins 1 and 2 of the opto. Without battery connected all the LEDs are on and main is solid red, some are above 1v some are as low as .85v.

11.4v is close enough.

Boy, that doesn't make much sense. When the LEDs are on, the voltage across R6 (input to the gate driver) should drop to around 1v. When you short pins 7-8, the voltage looks correct.

Check the board very carefully around U2 and R6 to make sure there is not a solder bridge somewhere. Hold the board up to a light and look through it from the bottom side. Measure the value of R6.

*edit*

If you run it with no cells and the LEDs lit for a while, does U2 get warm?
 
GGoodrum said:
<in a January post>
With 3.0V versions of the through-hole TC54s available at Digikey now, we really don't need to do a complicated set of voltage divider resistors/jumpers in order to implement a version that will work with LiMn/LiCo-based packs. All you need to do is substitute the 3.0V TC54s for the 2.1V versions and use different value resistors for the voltage dividor used by the LM431s

I don't mean to be dense, but I'm about to buy a bunch of Konions and I want to make sure I get it right. To work with LiMn cells, all I need to do is order a set of different TC54's and adjust the resistor divider used byt the LM431? In another post, I thought Gary had said that the new board design would have a simple trace cut feature for moving from LiFePo4 to LiMn, but I just want to make sure I order the right stuff from Mouser.

wanders
 
Nate said:
It has not done any "throttling" yet, always Red or Green.

The 12v bus is only 11.4v? Could this be the problem? I short the opto and it stays at about 11.35v. Latched it is about 11.35v.

I don't have a variable power supply so I can't test the any shunts as described above... I did test the voltage a different way: No battery all LEDs on 5v, opto shorted .65v and latched 1.4v. I opened up the ping charger and there are no pots. My Soneil 4808srf is down for the count at the moment.

I can measure the voltage across pins 1 and 2 of the opto. Without battery connected all the LEDs are on and main is solid red, some are above 1v some are as low as .85v.

OK, I had another few thoughts:
Test with no cells until it behaves properly.

The 5v measurement is too high if the cell LEDs are lit. It should be more like 1.4v or less.
This might happen if there is a short between the all shunts line and the any shunts line.
With power off and no cells, try measuring the resistance between the any shunts and all shunts lines. It should measure something greater than zero.

I found it real easy to get a solder bridge on the diode connections due to the close spacing. Inspect the bottom of the board very carefully around all the diode connections. Hold it up to a light and look through.
 
wanders said:
I don't mean to be dense, but I'm about to buy a bunch of Konions and I want to make sure I get it right. To work with LiMn cells, all I need to do is order a set of different TC54's and adjust the resistor divider used byt the LM431? In another post, I thought Gary had said that the new board design would have a simple trace cut feature for moving from LiFePo4 to LiMn, but I just want to make sure I order the right stuff from Mouser.

wanders

Yes, that is correct. We did not work out the divider values for LiMn cells, however. Do you know what the recommended charging voltage for those is?
 
fechter said:
OK, I had another few thoughts:
Test with no cells until it behaves properly.

The 5v measurement is too high if the cell LEDs are lit. It should be more like 1.4v or less.
This might happen if there is a short between the all shunts line and the any shunts line.
With power off and no cells, try measuring the resistance between the any shunts and all shunts lines. It should measure something greater than zero.

I found it real easy to get a solder bridge on the diode connections due to the close spacing. Inspect the bottom of the board very carefully around all the diode connections. Hold it up to a light and look through.

R6 is 4.68k negative lead on right and 3.77k the other way... U2 does not seem to get warm, only tested a for a min...
There does not seem to be a short. The resistance is: 63k All to Any, 66k Ground to All and 12k Ground to Any. I really hoped it would be something simple like a solder bridge, but I can't find one... I have a great headset magnifying glass... View attachment Headset Small.jpg Held up to light too... Here is the bottom:View attachment LiFePo4 BMS V2.3 Bottom 1s.jpgView attachment 2View attachment LiFePo4 BMS V2.3 Bottom 3s.jpgView attachment LiFePo4 BMS V2.3 Bottom 4s.jpg
 
Nate said:
R6 is 4.68k negative lead on right and 3.77k the other way... U2 does not seem to get warm, only tested a for a min...
There does not seem to be a short. The resistance is: 63k All to Any, 66k Ground to All and 12k Ground to Any. I really hoped it would be something simple like a solder bridge, but I can't find one... I have a great headset magnifying glass...


Looks clean alright. Nice work.

I hate to suggest removing parts for troubleshooting since the board is not rework-friendly and the through hole plating rips out easily, BUT, at this point I would suggest lifting R6 (you can just lift one side) to see how this affects the voltage on the any shunts line.

Without R6, the any shunts line should stay low and keep the main LED green. It appears that something is holding that line high somewhere.

If the LED stays green without R6 installed, then it would indicate that none of the optocouplers are getting enough juice to turn on.

Maybe use a meter and try tesing all the diodes. Diode check mode works best if your meter has that, otherwise use the resistance mode and swap the leads to measure both ways. A shorted diode could possibly cause that problem. Try that first.
 
Nate said:
It has not done any "throttling" yet, always Red or Green.
Not sure if this is relevant to your problem Nate, but I found that the main LED switched very suddenly from reddish orange to green once the voltage on pins 7 and 8 of all the optocouplers had dropped to around 2V. From what others had said, I expected a gradual change to green as each of the orange LEDs came on, but that doesn't happen, at least in my case. The main LED just turns gradually from red to slightly orangeish, then suddenly green as it cuts off. Maybe this is due to some variation in LEDs (or a problem with my eyesight).

I've tweaked my charger to 44.3V (for 12s), but think I may need to turn it up just a touch more to speed up the balancing stage.
 
fechter said:
Looks clean alright. Nice work.

I hate to suggest removing parts for troubleshooting since the board is not rework-friendly and the through hole plating rips out easily, BUT, at this point I would suggest lifting R6 (you can just lift one side) to see how this affects the voltage on the any shunts line.

Without R6, the any shunts line should stay low and keep the main LED green. It appears that something is holding that line high somewhere.

If the LED stays green without R6 installed, then it would indicate that none of the optocouplers are getting enough juice to turn on.

Maybe use a meter and try tesing all the diodes. Diode check mode works best if your meter has that, otherwise use the resistance mode and swap the leads to measure both ways. A shorted diode could possibly cause that problem. Try that first.
Thanks. Lots of cleaning done. I had already checked all the diodes and all are good except D2. I pulled up one side of that one to check and it was ok.
With R6 up the light is Green and all LEDs flash if I touch the ground, any and all shunts under the board. It flashes sometimes when taking the readings too. Here are the readings: (All tested with R6 up on one side)
Plugged in, light goes green, not latched: All to Any = 10.3v, All to ground = 12v, Any to ground = 1.6v and drops/fluctuates.
Latched: All to Any = 10.4v, All to ground = 11.3v, Any to ground = 1v and drops/fluctuates.
Not powered: All to Any = 181.2k, All to Ground = 72.5k, Any to ground starts at 40M? and steadily drops. 5M in about a minute...?

Malcolm said:
Not sure if this is relevant to your problem Nate, but I found that the main LED switched very suddenly from reddish orange to green once the voltage on pins 7 and 8 of all the optocouplers had dropped to around 2V. From what others had said, I expected a gradual change to green as each of the orange LEDs came on, but that doesn't happen, at least in my case. The main LED just turns gradually from red to slightly orangeish, then suddenly green as it cuts off. Maybe this is due to some variation in LEDs (or a problem with my eyesight).

I've tweaked my charger to 44.3V (for 12s), but think I may need to turn it up just a touch more to speed up the balancing stage.
Thanks for the info. I have yet to see any color other than Green or Red... No Yellow or Orange. Good to know what to expect when I get it working.
My charger is at a little over 60v for 16s so about 3.75v per channel.
 
Something is still not quite making sense.
The 1.6v any shunts to ground measurement still seems too high. With the main LED green, it should be more like zero. Something is backfeeding this line.

You might try replacing R6 with something higher, like 20k and see what happens.

At this point it almost seems like the gate driver chip is bad somehow (seems unlikely).

Another possibility is a leaky diode somewhere. This is also very unlikely, but possible. It would be real hard to locate without lifting them (PITA). Diodes usually short when they fail, but cheap, made you-know-where ones could be defective in a leaky manner. It would be one going to the any shunts line.
 
I tried replacing R6 with a 20k but it won't latch, the LEDs stay on.

I retested every diode with my Mastech MS2138R and the only one that "leaks" is D2. I once again pulled it up to check and it is fine. Looking at the schematic this looks to be correct but, since this is my first circuit board build and the first time I have looked at an electronics schematic I could be totally wrong. Why would I have to lift the others if I only get about .638v one way? Sorry for my ignorance...

I did take a reading of pins 7 and 8 on the optocouplers with all the LEDs on and found some strange results: Channel 1 = 10.82v. Channels 2, 6, 7, 8, 11, 14 and 16 = 4.3v - 4.4v. Channels 3, 4, 5, 9, 10, 12, 13 and 15 = 11.2v - 11.3v. Don't know if this applies but since Malcom mentioned it I thought I would test.

Another strange reading was from the ground to charging pad after channel 4. It was almost 16v. I tested between the other banks of 4 channels and they were each 14.75v. If you add it up it equals about the 60.3v my charger is set at. Do the first 4 channels take the extra volts from the charger? Does this indicate that my problem is in the first 4 channels? Once again, sorry for my ignorance...

I have another Gate Driver on the way. It should be here this Thursday. Thank you for all the help Mr. Fechter!
 
Don't replace the gate driver quite yet.

Your measurements on pins 7-8 indicate that none of the optocouplers are fully on enough to trip the charging circuit.

Some of the charging current goes through the control circuit and bypasses cells 1-4, but I would expect that to make the reading across them a bit lower. Let me think about that.

You can test diodes in circuit as long as the rest of the circuit does not affect the readings.

The 16v across the first 4 channels sounds too high. Try measuring the voltage across the shunt resistors on the first 4 channels when it's running. Also try measuring across each cell circuit at the battery connection points. Maybe one shunt is not coming on all the way and is preventing the current from getting high enough to trigger the optos. You need about 500ma to pass before the optos will fire.

If all else fails, you can ship the board to me and I can look at it.
 
Back
Top