GGoodrum
1 MW
ryan_lirui said:
Actually, for 24s, two boxes can be used, and I have slightly different endplates for the extension box. I'm also doing endplates for the shorter Hammond 1" high box which can be used to hold one 6s shunt board.
Even Newer 4 to 24-cell Battery Management System (BMS)
- Thread starter GGoodrum
- Start date
Actually, for 24s, two boxes can be used, and I have slightly different endplates for the extension box. I'm also doing endplates for the shorter Hammond 1" high box which can be used to hold one 6s shunt board.
steveo
100 kW
Hey Gary,
Just wanted to ask, I had an issue with the older version of bms if i unpluged one of the battery channels while the rest where connected i would end up blowing out a channel in the bms, this happed several times for me because i only used the bms to charge my pack. Has this issue been resolved with the newer generation bms? I hope you understand what i mean with "blowing out a channel". This would probably never happend had i kept it permanatly mounted to the batter pack.
p.s. great work on the little box shape 18 channel bms, very well thought out!
-steveo
Just wanted to ask, I had an issue with the older version of bms if i unpluged one of the battery channels while the rest where connected i would end up blowing out a channel in the bms, this happed several times for me because i only used the bms to charge my pack. Has this issue been resolved with the newer generation bms? I hope you understand what i mean with "blowing out a channel". This would probably never happend had i kept it permanatly mounted to the batter pack.
p.s. great work on the little box shape 18 channel bms, very well thought out!
-steveo
GGoodrum
1 MW
steveo said:
Although we are looking into adding some active protection (PTCs and TVSs...) to keep wiring errors from popping TC54s, I found the best way is to not wire connections directly from the cells to the BMS channels, but use a separate set of connectors. This way, wiring errors can be tested for, and corrected, before connecting the BMS unit. With LiPos, this has been less of an issue because the 6s-5Ah packs come with 7-pin balancer pigtails, so it is much harder to screw up the connections.
By breaking the pack into smaller sub-sections, like 6s, each of these sections can be completely independent, in terms of connections to the BMS. Each of these 6s sections can be connected to a corresponding group of 6 cells, in any order. The only connection between these boards, and the control board is the two-wire, isolated opto output.
-- Gary
One question: Does the board use any particular connector for the battery sense cables and power cables? or is it assumed that only wires will be soldered in directly and in-line connectors will be used between BMS and batteries?
I need to buy some connectors and if there is already something planned then I'll just use that. Otherwise I'll buy some in-line connectors of my choice. (what do people favour for battery-sense connections, BTW?) (I use powerpoles for power connections)
I need to buy some connectors and if there is already something planned then I'll just use that. Otherwise I'll buy some in-line connectors of my choice. (what do people favour for battery-sense connections, BTW?) (I use powerpoles for power connections)
__Tango
1 W
j3tch1u said:
Great. Thanks!
__Tango
1 W
Another random question about wire gauges. I've seen a lot talked about the charging cable gauge. What about the wires from each cell to the BMS? What is the smallest gauge wire that can be used? 24 gauge? I have a want to use it because i have cat5e cable, connectors, and punchdown blocks coming out of my ears. 
__Tango said:Another random question about wire gauges. I've seen a lot talked about the charging cable gauge. What about the wires from each cell to the BMS? What is the smallest gauge wire that can be used? 24 gauge? I have a want to use it because i have cat5e cable, connectors, and punchdown blocks coming out of my ears.
The cell tap wires would only carry the shunt current at most, so 24ga might be OK if the wires aren't too long. If the wires are too long/small then the voltage drop in the wire may allow a cell to go slightly high/low compared to the set point.
If one tap wire is disconnected from the pack, that shouldn't cause a problem. If the BMS is connected to the cells and one of the cell interconnects goes bad or is disconnected (between cells), it will probably blow something.steveo said:
hello,
I have just finish a 20s bms (v2.4) but I have some problem with it:
when I test it with a 3.6v ps, all the module seem to work correctly (no current below or at 3.6, current upper or at 3.65)
connected to the batteries pack, the charge and balance seem also ok. all led will go on.
but
1/ some leakadge in random cell:
if I remove the charger plug: after some second the led shutdown but it remain some current leak from one cell (100ma-200ma)
the cell dropping the current is not always the same !
when i remove and put back the pack interconnection, the current stop and didn't come back all the time.
after some time, all cell remain near 3.6, execpt one at around 3.35v (but not always the same)
the current change also when moving cell wire.
I thing there is an oscillation. the cell wire is 1.5mm² and about 15cm. there is an 2Amp fuse near the cell.
in the silicium BMS, I saw a cap between in and output of the LM431. is there any change to solve the problem with this cap ?
2/ never switch the thyristor to shutdown the charge:
the "all shunt" decrease from 9v to 3v then begin to oscillate.
if I put a multimeter on the gate of SCR1, the end of charge works correctly.
have you a list of change from V2.4 to v2.6 ?
I already saw a change on R3 on pcb from 10k to 4.7k. this doesn't solve the problem.
sorry if the answerd is already posted, I didn't read all the threat.
thanks
JC
I have just finish a 20s bms (v2.4) but I have some problem with it:
when I test it with a 3.6v ps, all the module seem to work correctly (no current below or at 3.6, current upper or at 3.65)
connected to the batteries pack, the charge and balance seem also ok. all led will go on.
but
1/ some leakadge in random cell:
if I remove the charger plug: after some second the led shutdown but it remain some current leak from one cell (100ma-200ma)
the cell dropping the current is not always the same !
when i remove and put back the pack interconnection, the current stop and didn't come back all the time.
after some time, all cell remain near 3.6, execpt one at around 3.35v (but not always the same)
the current change also when moving cell wire.
I thing there is an oscillation. the cell wire is 1.5mm² and about 15cm. there is an 2Amp fuse near the cell.
in the silicium BMS, I saw a cap between in and output of the LM431. is there any change to solve the problem with this cap ?
2/ never switch the thyristor to shutdown the charge:
the "all shunt" decrease from 9v to 3v then begin to oscillate.
if I put a multimeter on the gate of SCR1, the end of charge works correctly.
have you a list of change from V2.4 to v2.6 ?
I already saw a change on R3 on pcb from 10k to 4.7k. this doesn't solve the problem.
sorry if the answerd is already posted, I didn't read all the threat.
thanks
JC
I've never heard of the 'leakage' problem you describe, or at least not one that randomly moves from cell to cell. You think this may be an oscillation in the cell circuit? Other than the leakage, it sounds like the cell circuits are working. An oscilloscope would be one way to see if something is oscillating.
One possibility would be soldering flux residue around the LM431 or the divider resistors causing a conductive path. Make sure it's clean.
A capacitor between input and output on the LM431 can't hurt and should stop any kind of oscillation. You could use .1uf caps tacked onto the bottom of the board.
The auto shutoff should definitely trigger when the All Shunts line is down to 3v. Make sure the resistor going to the SCR anode (I think it's R5) is 1k or less. If that one is too high, it may not latch. The change from 10k to 4.7k on the Q1 emitter will raise the voltage where the SCR should latch.
Check continutity between the SCR gate and the collector of Q1. A high resistance there could cause your problem.
The only other change from 2.4 to 2.6 was a divider on the input of the gate driver to ensure a low state is achieved during thottling. I don't think this is a problem with your board.
One possibility would be soldering flux residue around the LM431 or the divider resistors causing a conductive path. Make sure it's clean.
A capacitor between input and output on the LM431 can't hurt and should stop any kind of oscillation. You could use .1uf caps tacked onto the bottom of the board.
The auto shutoff should definitely trigger when the All Shunts line is down to 3v. Make sure the resistor going to the SCR anode (I think it's R5) is 1k or less. If that one is too high, it may not latch. The change from 10k to 4.7k on the Q1 emitter will raise the voltage where the SCR should latch.
Check continutity between the SCR gate and the collector of Q1. A high resistance there could cause your problem.
The only other change from 2.4 to 2.6 was a divider on the input of the gate driver to ensure a low state is achieved during thottling. I don't think this is a problem with your board.
hi fechter
thanks for your reply
I have made some progress but still not 100% working
I have replaced the 2A fuse on each cell balance wire by a 20A because the previous one was too inductive. there was also some acoustic noise coming out when balancing.
I have also added a 100nf between input and output on lm431. there is not more HF oscillation now.
still one problem:
the average balancing current is only 150mA. the all shunt remain at 11v. (before I made a wrong mesurement because the label is not right on pcb.)
if I short circuit the mosfet and use a 600mA power supply limitation, the all shunt decrease and the thyristor trig correctly.
Is it normal to have only 150ma during balancing ? I was expecting 500ma instead ? (I have put the 6.8Ohms/5W resistor)
if not there is something wrong with the gate driver current limitation. I don't fully understand this part of schematic.
thanks for your time
JC
thanks for your reply
I have made some progress but still not 100% working
I have replaced the 2A fuse on each cell balance wire by a 20A because the previous one was too inductive. there was also some acoustic noise coming out when balancing.
I have also added a 100nf between input and output on lm431. there is not more HF oscillation now.
still one problem:
the average balancing current is only 150mA. the all shunt remain at 11v. (before I made a wrong mesurement because the label is not right on pcb.)
if I short circuit the mosfet and use a 600mA power supply limitation, the all shunt decrease and the thyristor trig correctly.
Is it normal to have only 150ma during balancing ? I was expecting 500ma instead ? (I have put the 6.8Ohms/5W resistor)
if not there is something wrong with the gate driver current limitation. I don't fully understand this part of schematic.
thanks for your time
JC
Good deal on the capacitors.
It is normal for the current to drop below the shunt level during throttling when there are no cells attached. If you adjust your supply so that the shunt LEDs are just barely full, you should see close to 500ma. If the voltage is incresed slightly so that throttling begins, it is normal for the current to drop. When cells are attached, it will stay much closer to the target current.
This problem was addressed in the new design by using a fixed frequency PWM. I think in actual practice, the current reduction does not cause a problem.
It is normal for the current to drop below the shunt level during throttling when there are no cells attached. If you adjust your supply so that the shunt LEDs are just barely full, you should see close to 500ma. If the voltage is incresed slightly so that throttling begins, it is normal for the current to drop. When cells are attached, it will stay much closer to the target current.
This problem was addressed in the new design by using a fixed frequency PWM. I think in actual practice, the current reduction does not cause a problem.
damcard
100 W
Greetings,
I have 18 6s1p 5ah zippy flightmax packs coming in the mail. I want to make two packs for my next bike with nine of the flightmax packs in parallel and run them them in parallel. So, that will be 12s9p. I have tried reading through this thread to understand what I need to build a bms for it, but got lost it is so big. Can someone point me in the right direction? I want to use this design as a learning experience for now. I can start another thread for it if it brings this one off topic. -damcard
I have 18 6s1p 5ah zippy flightmax packs coming in the mail. I want to make two packs for my next bike with nine of the flightmax packs in parallel and run them them in parallel. So, that will be 12s9p. I have tried reading through this thread to understand what I need to build a bms for it, but got lost it is so big. Can someone point me in the right direction? I want to use this design as a learning experience for now. I can start another thread for it if it brings this one off topic. -damcard
fechter said:
well, the previous test was done with all the cell connected in normal condition.
the thyristor doesn't trig.
If I bypass the mosfet to increase the current, the trigger is ok.
so I thing there is a problem with the current too loo.
is it an easy way to I increase a little bit the throttling current ?
jc
Hmmm..
You could try reducing the charging voltage slightly. The shunts will increase in current as the cell voltage increases until the opto coupler triggers throttling, then if the charger voltage swings too high when the FET is off, you'll see a drop in shunt current. If you adjust the charging voltage, you should find a peak in current somewhere about 2v above the pack voltage. If the charging voltage is high, you'll see the throttling LED turn mostly green during throttling.
Another possibility is one of the shunt resistors is significantly higher than the others. One of the weaknesses of this design is if one cell triggers thottling at a lower current, it prevents the other cells from reaching the trigger point since they are all in series.
You could try measuring the voltage across all of the shunt resistors when it is running and compare the values. They should be around 3v.
Another thing that has been seen before is if one of the cell LEDs was inserted backward at any time, the LED can be partially damaged so that it will still light up, but the forward voltage will be much higher than normal. A good one is about 2.0v when lit. A bad one may be 2.4v or higher. You can try measuring the voltage across the LEDs when lit to test this.
You could try reducing the charging voltage slightly. The shunts will increase in current as the cell voltage increases until the opto coupler triggers throttling, then if the charger voltage swings too high when the FET is off, you'll see a drop in shunt current. If you adjust the charging voltage, you should find a peak in current somewhere about 2v above the pack voltage. If the charging voltage is high, you'll see the throttling LED turn mostly green during throttling.
Another possibility is one of the shunt resistors is significantly higher than the others. One of the weaknesses of this design is if one cell triggers thottling at a lower current, it prevents the other cells from reaching the trigger point since they are all in series.
You could try measuring the voltage across all of the shunt resistors when it is running and compare the values. They should be around 3v.
Another thing that has been seen before is if one of the cell LEDs was inserted backward at any time, the LED can be partially damaged so that it will still light up, but the forward voltage will be much higher than normal. A good one is about 2.0v when lit. A bad one may be 2.4v or higher. You can try measuring the voltage across the LEDs when lit to test this.
__Tango
1 W
fechter said:__Tango said:Another random question about wire gauges. I've seen a lot talked about the charging cable gauge. What about the wires from each cell to the BMS? What is the smallest gauge wire that can be used? 24 gauge? I have a want to use it because i have cat5e cable, connectors, and punchdown blocks coming out of my ears.
The cell tap wires would only carry the shunt current at most, so 24ga might be OK if the wires aren't too long. If the wires are too long/small then the voltage drop in the wire may allow a cell to go slightly high/low compared to the set point.
Great. Do you think 18inches is too long?
__Tango said:
That should be OK, but I would suggest measuring the cell voltages to see that the variation is not too much.
Kingfish
100 MW
Gents –
Thanks very kindly, KF
- Can you tell us what will be the sizes offered for the new 4.0 BMS will be please?
- What is the targeted cost of each sized unit?
- When you expect to have them ready for us to place our orders?
Thanks very kindly, KF
AndyH
10 kW
Kingfish said:
Each shunt board can handle up to six channels. They can be stacked as you've seen in earlier photos, but they can also be left to run 'long' and/or long and stacked. For LiFePO4, assuming a 3.7V finish and a 400V limit for the control board, it should be good with no control section mods for up to 108 cells in series. (I'm guessing on the cell limit - I don't know yet if the 400V point is a 'hard limit' or a 'pack nominal' level.)
I've recently completed a 25S LiPo unit (5 channels on each shunt board) - Gary has done an amazing job laying out the boards. First time thru with no itemized parts list took just over an hour to build a 5S shunt board. The Molex Micro-fit connectors on the shunt boards make it very easy to connect to the cells.
View attachment 25ch_39s.jpg
This unit is going to make it very easy to distribute the shunt boards (each shunt board only needs a 2-wire connection to the control board) for larger packs, and it will be easy to make the BMS a 'cube' or a longer/flatter configuration.
Best guesses? Cost for ready-built units should be close to the price of the V2.6 BMSs. Physical board size has changed a bit - Gary's covered that earlier. If the control section is powered from the pack, the smallest unit will be 5S. When? About two weeks after Gary and Richard finalize the torture testing the boards are going thru. :wink: (our mileage may vary, void where prohibited...)
Andy
alangsam1@me.com
10 mW
- Joined
- Jan 13, 2010
- Messages
- 26
really looks good. this is like waiting for 9 months to have a baby!
alangsam1@me.com
10 mW
- Joined
- Jan 13, 2010
- Messages
- 26
so im clear. i connect these BMS boards to the balance taps coming from my LVC boards. I connect one side of the BMS to my power supply and the output of the BMS to the battery pack
Kingfish
100 MW
Gents –
The photos rock!
I need to know the physical size of the boards for the enclosure that I am designing. 8)
The photos rock!
I need to know the physical size of the boards for the enclosure that I am designing. 8)
- What are the Length, Width, Height – and minimal clearance please?
- May I presume that the Master and Slave boards are the same size?
- What is the weight of each board?
The waiting is killing me....
I've fitted my cells to the moped and am going on inaugural ride (further than round the block) tomorrow. So I'm running BMSless for the time being, hoping everything comes together before I do any damage...
Like the man says, no pressure....
I've fitted my cells to the moped and am going on inaugural ride (further than round the block) tomorrow. So I'm running BMSless for the time being, hoping everything comes together before I do any damage...
Like the man says, no pressure....
Similar threads
