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

[kadir]

I never had the issue of the BMS blowing.After Charge I always disconnect the BMS. and only Dicharge them my 9Ah, or till 72V (3V for each cell). I have seen the weakest cell to be 2.94V and the supercell to be at 3.11V, others were at (~3V) when I was exactly at 72.1V resting voltage.

One thing I have noticed is that, lets say one cell gets disconnected.

On my voltage meter, it shows cell 1 is 3.8V, while cell 2 is 2.94V; instead of the 3.4V each. Another thing to notice was, the sum of the two normal voltage was the same as the sum of weird voltages.

Thats how my BMS is, and thats how I check the voltages on the fly

 

[GGoodrum]

I've not had a problem with the Molex connectors at all. The pins also are crimped on, and I can use the same tool I use with the VAL-U-LOK connectors, and they securely lock into the housing.

The most important part is that you can still use 18-gauge wire with these. It's not about how much current the pins can handle, it is how big a wire you can use, in order to minimize the resistance, and the subsequent voltage drop. This becomes more important when you have higher balance/shunt currents, like 1A and above, and if you have wire lengths longer than 6-8".

-- Gary

 

[fechter]

When you connect the BMS board to the cells, the voltage across the cell circuits all start out at zero, then it suddenly gets battery voltage. The divider resistors will instantly divide the voltage, but each cell circuit has to go from zero to something, and in the process will pass through a zone where the LVC can be activated. Due to component variations, if some cells hit LVC and activate before the others, it could pull those down, causing increased voltage on others. If the voltage spike is fast enough, the cell shunt may not be fast enough to clamp it in time. The TC54 is not very over voltage tolerant, and will blow up somewhere around 12-15v.

If you have a large multi-pin connector, it is likely that the pins won't all make contact at exactly the same time, so you have the potential for uneven voltage distribution on the circuit.

I think normally all the cells will all come up fast enough that none of them should trigger LVC, but all of this happens so fast it is hard to test. You could also get spikes any time a wire is disconnected with current running through it (like a charger).

To prevent this kind of failure, I'm looking at adding TVS diodes or zener diodes across each cell circuit that can clamp at a voltage below what it takes to blow up a TC54. The TC54 is rated for 10v max. TVS diodes have a significant amount of leakage current at the lower voltage ratings for some reason. One that clamps around 10v has much less leakage than one that clamps at 5v.

 

[CamLight]

The most important part is that you can still use 18-gauge wire with these. It's not about how much current the pins can handle, it is how big a wire you can use, in order to minimize the resistance, and the subsequent voltage drop. This becomes more important when you have higher balance/shunt currents, like 1A and above, and if you have wire lengths longer than 6-8".

-- Gary

And larger gauge wires are surprisingly good heat sinks and can significantly increase the current handling ability of a terminal/pin over its nominal rating (which is often for just a 10C or, at most, a 30C temp rise).

 

[steveo]

When you connect the BMS board to the cells, the voltage across the cell circuits all start out at zero, then it suddenly gets battery voltage. The divider resistors will instantly divide the voltage, but each cell circuit has to go from zero to something, and in the process will pass through a zone where the LVC can be activated. Due to component variations, if some cells hit LVC and activate before the others, it could pull those down, causing increased voltage on others. If the voltage spike is fast enough, the cell shunt may not be fast enough to clamp it in time. The TC54 is not very over voltage tolerant, and will blow up somewhere around 12-15v.

If you have a large multi-pin connector, it is likely that the pins won't all make contact at exactly the same time, so you have the potential for uneven voltage distribution on the circuit.

I think normally all the cells will all come up fast enough that none of them should trigger LVC, but all of this happens so fast it is hard to test. You could also get spikes any time a wire is disconnected with current running through it (like a charger).

To prevent this kind of failure, I'm looking at adding TVS diodes or zener diodes across each cell circuit that can clamp at a voltage below what it takes to blow up a TC54. The TC54 is rated for 10v max. TVS diodes have a significant amount of leakage current at the lower voltage ratings for some reason. One that clamps around 10v has much less leakage than one that clamps at 5v.

Hey Fechter,

this sounds like the exact problem i had several times when i had my bms...

-steveo

 

[fechter]

I think some super-cheap 8-10v zener diodes across each cell circuit should protect against this.

 

[sandman]

Gary,

sandman wrote:Is it possible to connect one of the 4-cell breakoff sections from V 2.6c to the much older V1.5 (16-cell) such that I can have LVC/HVC for 20 cells? I am using LiFePO4's.

Thanks,
Dick



Sorry, I missed this earlier.

The cell circuits of the v1.5 board and the v2.6c are basically identical, except for the shunt transistors. The older v1.5 boards used KSA931s and the v.6c used BD136s, but that doesn't mean they couldn't also use the KSA931s. For the 4-channel v2.6c section, just use the same shunt resistors (6.8 ohm/5W), and the same divider resistors (75k/180k...), and it should be the same.




Another question about adding a 4 cell v 2.6 section to the v 1.5 16cell board. I am delaying putting parts on my new charge control board waiting for all the changes to occur. I want to charge my 20 cells using the 16 cell board. I must need to add a connection from the most positive cell on the v2.6 section to the main board. Where would that go? The positive charger lead will probably connect to the positive of the 20th cell??? Do I need to open any traces on the 16 cell board?

I thank you for your help,

Dick

 

[amberwolf]

I think some super-cheap 8-10v zener diodes across each cell circuit should protect against this.

Directly across the cell connection itself, reverse-biased? If so, what about 5V zeners? The cell voltage should never rise anywhere near that, so I'd expect that to be safe?

I ask because I have a bunch of 1W or 5W 5V zeners I could install in my v2.6 boards that I'm building up, before I ever hook them to the pack. I also have 12V zeners, but those probably wouldn't help.

 

[fechter]

5v regular zener diodes should be good. The leakage is around 10uA for those, so not much. TVS diodes have much higher leakage. A zener will blow up if it gets exposed to a swapped cell wire connection, but it should be obvious. For normal connect/disconnect, a zener should be fine to protect the TC54s from overvoltage.

 

[GGoodrum]

I must need to add a connection from the most positive cell on the v2.6 section to the main board. Where would that go? The positive charger lead will probably connect to the positive of the 20th cell??? Do I need to open any traces on the 16 cell board?

I thank you for your help,

Dick

I would run one wire from the positive of the 4th channel in the v2.6 section, to the negative connection for the 1st channel on the v1.5, like this:

View attachment CC-v1.5-v2.6c.png

The charger connections only need to go to the Charge Controller, and then directly to the pack. The cell circuits have their own connections.

-- Gary

 

[fechter]

Unless I'm misunderstanding something, the add-on section should go on the positive end of the main board. Maybe it doesn't matter if the charger is isolated, but I would normally have the charger (-) connection at one end since that's where the throttling takes place and all the cell circuits go up from there.

 

[GGoodrum]

If you are building up a v2.x board, one thing Richard just found is that you can totally eliminate the nasty HF noise/oscillations on these circuits by adding a small cap between pins 1 and 3 of the LM/FAN431. The value can be something between .002uF and .01uF. this was les of a problem on the v2.x boards, as they had half the shunt currents, but it is still enough to cause all kinds of weirdness. Anyway, we've finally got the original LM431-driven shunts to work the way they were supposed to, and now we are down to tweaking the HVC set point to optimize the throttling. Speaking of that, Richard has found some new PWM chips that could allow us to simplify the Charge Controller a bit.

Now that we have a rock-solid performing shunt, I'm also testing what will become a standalone balancer. It will use the case, fans and special end plates and the automatic fan controls, used on the BM Booster. What it will do is simply use the 1A+ shunts to quickly pull down each cell to 4.10V. This is pretty much what I do with my boosted Battery Medic. I set it to the Discharge mode, and set the voltage to 4.10V. What I do is charge the packs first, and when I start to see a bit of a variance, I use the BM/Booster combo to bring all the cells down to 4.10V, and then do a "top-off" recharge, to 4.15V. I tried before to use a BMS shunt board as a standalone balancer, but because of all the noise/oscillation problems, the cell variances were just too much, on the order of around 60-80mV. Now, with the "capped" LM431s, the deltas are under 20mV. I can get this much closer, on the order of 4-6mV, by using hand-matched divider resistors, on the front-end. Anyway, this should work pretty well. You just plug it into a 6s section, after a charge, and all six LEDs should lit up. When each cell hit the cutoff, its LED will go out. When all the LEDs are out, the cells are balanced. The fans will come on, as needed, and will also shutoff, when they are not needed. This allows you to just hook up the unit and walk away. Nothing will drain the cells down, once they hit the cutoff.

 

[GGoodrum]

Unless I'm misunderstanding something, the add-on section should go on the positive end of the main board. Maybe it doesn't matter if the charger is isolated, but I would normally have the charger (-) connection at one end since that's where the throttling takes place and all the cell circuits go up from there.

What I think you are missing is that he's not using the old v1.5 control section, but instead is using a Charge Controller board to do the throttling.

 

[sandman]

Gary,

I am sorry I led you astray, I plan to use the vers. 1.5 charge controller on the old 16 cell board. Basically I am adding 4 cells to my former 16 cell pack and adding a 4 cell break off section (v 2.6) to the old 16 cell board. I plan to continue to charge the same way I have since first using the (vers 1.5) 16 cell board. I realize I will need to take the charger voltage up to about 77 volts.

If I connect a wire from the add on section to the main board negative of the first cell (blue wire as your diagram indicates) it seems like I would be exposing the 20th shunt circuit to the full pack voltage or shorting the pack. Isn't this wire unnecessary? Each cell circuit is independent from its neighbor whether on the same board or adjacent??? The two boards have a common 12v and ground connection shown in your earlier diagram. I think I need a wire from the negative side of the 17th cell circuit to the most positive connection on the 16 cell board.

My best guess to connect the charger would keep the negative lead on the same place on the 16 cell board and the positive lead to the most positive cell.

Dick

 

[GGoodrum]

Gary,

I am sorry I led you astray, I plan to use the vers. 1.5 charge controller on the old 16 cell board. Basically I am adding 4 cells to my former 16 cell pack and adding a 4 cell break off section (v 2.6) to the old 16 cell board. I plan to continue to charge the same way I have since first using the (vers 1.5) 16 cell board. I realize I will need to take the charger voltage up to about 77 volts.

If I connect a wire from the add on section to the main board negative of the first cell (blue wire as your diagram indicates) it seems like I would be exposing the 20th shunt circuit to the full pack voltage or shorting the pack. Isn't this wire unnecessary? Each cell circuit is independent from its neighbor whether on the same board or adjacent??? The two boards have a common 12v and ground connection shown in your earlier diagram. I think I need a wire from the negative side of the 17th cell circuit to the most positive connection on the 16 cell board.

My best guess to connect the charger would keep the negative lead on the same place on the 16 cell board and the positive lead to the most positive cell.

Dick

In that case, you definitely don't want to hook it up like the diagram. You would do like you surmised, and connect the negative side of the 17th cell to where it says "Pack +".

 

[PJD]

Gary and Richard,

I'm returning her from a long absence.

I fortuitously blundered into your conversation with Dick since I may want to do the same expansion of my old v 1.5 boards as part of a 60V upgrade to the scooters - which continue to work just fine by the way.

So, are any of the breakaway V2.6 boards still available? Can you provide a very brief summary of what you all have been doing over the past 18 months or so?

Thanks a lot,

Paul D.

 

[ejonesss]

i was wondering if powering the bms via a ups transformer would work?

i took apart an old 48 volt ups (uses 4 12 volt batteries in series.

the transformer if rewired to use the output stage as the input can then put out 34 volts ac.

34 volts is not enough to charge a 48 volt lifepo4 pack.

i can build a doubler using diodes and capacitors to make 68 volts.

while 68 volts is 4.25 per cell does the bms throttler reduce the voltage to prevent taking the cells that high?

i would like to use a ups transformer to charge the batteries since it can charge the battery faster than the soneil charger.

judging by the wires the transformer can handle 30 to 4 a .

 

[MitchJi]

Hi Gary and Richard,

Gary and Richard,

I'm returning here from a long absence.

So, are any of the breakaway V2.6 boards still available? Can you provide a very brief summary of what you all have been doing over the past 18 months or so?

It would be nice on the V4 thread, if a couple of years from now, people could easily (without bothering you) and without reading 125 pages of posts get a good idea of the current board features and status.

When you start the new V4 thread would you please follow your initial post(s) with one or two empty posts (reserve one or two places at the top of the list of replies) and use those for significant changes, upgrades and any important current status information?

I think it would also be a good idea to reserve one or two spots for a FAQ which you could keep updated with any repeat questions or general interest questions.

Thanks!

Mitch

 

[GGoodrum]

Hi Gary and Richard,

It would be nice on the V4 thread, if a couple of years from now, people could easily (without bothering you) and without reading 125 pages of posts get a good idea of the current board features and status.

When you start the new V4 thread would you please follow your initial post(s) with one or two empty posts (reserve one or two places at the top of the list of replies) and use those for significant changes, upgrades and any important current status information?

I think it would also be a good idea to reserve one or two spots for a FAQ which you could keep updated with any repeat questions or general interest questions.

Thanks!

Mitch

Yep, good ideas, we'll definitely do all those things, hopefully before this thread hits 200 pages...

Pat, unfortunately, you haven't missed much. We, on the other hand, have missed deadline after deadline with the v4.x BMS. :x

After god nows how many iterations, I think we have the most stable design yet. No more oscillations and no more voltage drop issues and no more weirdness. We are finally down to just adding some input protection, to prevent knuckleheads like Thud, or Dogman, from blowing TC54s when they mis-wire their packs. All kidding aside, I've probably popped the most myself, especially back when I was still building a123 packs. With LiPos, there's less chance of miswiring anything because all the "sub-packs" come pre-wired with balance plug pigtails. Anyway, at a minimum, we'll add 5.1V zeners across each cell, but we're also going to test some PTC resettable fuses.

Here's what the latest layout looks like:

View attachment 6s BMS-v4.1.1m.png

It still needs whatever we end up with for protection, but other than that it is pretty complete. This initial version is still designed to either be installed in Hammond extruded aluminum boxes (like the Battery Medic Booster...), or stacked, on on top of each other. Each board has an automatic control section for two 5V or 12V fans. These will automatically come on whenever the shunts start operating, and will automatically shut off, when the shunts are not operating.

In order to minimize any voltage drop issues, with 1A+ shunts, the cell connection mini-terminal blocks are moved to the middle of the board, between the shunts and the LVC/HVC circuits. The shunt traces are extra wide (.120"...), and these can be beefed up with solder.

As soon as we get the PTC/zener issue sorted, I think I will finally be ready to to get some boards made.

-- Gary

 

[ejonesss]

@ggoodrum

the design looks good.

since the PTC/zener is the only issue could you post a parts bom list so we can get started by ordering the parts (we hate it when the parts stores are out of parts).

 

[LI-ghtcycle]

Sounds Great GGoodrum!

I am very happy to have you and the rest working on this project to make such a fine BMS available to the "unwashed" masses! (ok, so I do shower, but you know what I mean! :wink: )

I second the motion for a BOM at your earliest convenience , and I also would love any details such as dimensions of the boards themselves, and cooling requirements. (i.e. suggested fans for cooling and configuration of the PCB(s) for something in the order of a 8p18s battery pack for optimal cooling in a ebike triangle pack of A123 cells)

Do I understand correctly that the PCB in your latest post is a 6 cell group, and if so what denominations will they be available in? I seem to remember groups of 4 as being the standard?

Thanks!

 

[GGoodrum]

I haven't had time to do the BOM yet, but I'll try to get to it tonight, maybe. It is not too far off a combination of a LVC/HVC board, and the Booster board. You don't need the 2n3904 transistors from the Booster, and you will need to add some .0047uF ceramic caps. The ones I picked are 140-50Z5-472-RC. Also, in place of the ILD2s, I'm now using the cheaper Fairchild 512-MCT6 optos, which have the same pinout. For shunt resistors I'm still using 283-6.2-rc, which are 6.2 ohm/3W metal oxide resistors. It looks like we will use the 5.1V zeners and the PTC fuses. the zeners I picked are 512-bzx79c5v1, and I'd get extras, as they will essentially act as TC54 protectors, and will pop, if needed. The PTC part we picked is 652-mf-rx110/72-0 but we don't even know if these will work yet, so you might hold off on them until we finish testing them.

The charge controller section is going to be basically the same as the CC BOM, plus a few parts, like the CD4060 tier chip.

Like I said, I will try and get the BOM done as soon as I can.

Regarding the number of channels per section, I'm still playing with it. I will definitely do 6, and that will be the first one released, but I may also do an 8-channel section layout as well. One other idea I'm looking at are 12 and 16/18-channel dedicated/single board layouts. I can make these fit the larger Hammond-type boxes, which I really like. For these, the LEDs will poke up through holes in a custom lid, one that will also have mounts for three or four fans.

-- Gary

 

[LI-ghtcycle]

I haven't had time to do the BOM yet, but I'll try to get to it tonight, maybe. It is not too far off a combination of a LVC/HVC board, and the Booster board. You don't need the 2n3904 transistors from the Booster, and you will need to add some .0047uF ceramic caps. The ones I picked are 140-50Z5-472-RC. Also, in place of the ILD2s, I'm now using the cheaper Fairchild 512-MCT6 optos, which have the same pinout. For shunt resistors I'm still using 283-6.2-rc, which are 6.2 ohm/3W metal oxide resistors. It looks like we will use the 5.1V zeners and the PTC fuses. the zeners I picked are 512-bzx79c5v1, and I'd get extras, as they will essentially act as TC54 protectors, and will pop, if needed. The PTC part we picked is 652-mf-rx110/72-0 but we don't even know if these will work yet, so you might hold off on them until we finish testing them.

The charge controller section is going to be basically the same as the CC BOM, plus a few parts, like the CD4060 tier chip.

Like I said, I will try and get the BOM done as soon as I can.

Regarding the number of channels per section, I'm still playing with it. I will definitely do 6, and that will be the first one released, but I may also do an 8-channel section layout as well. One other idea I'm looking at are 12 and 16/18-channel dedicated/single board layouts. I can make these fit the larger Hammond-type boxes, which I really like. For these, the LEDs will poke up through holes in a custom lid, one that will also have mounts for three or four fans.

-- Gary

DROOOOOOL!!!!!!!

Sounds great! I'm not familiar with the hammond boxes? I'm just curious the approximate size that say a 18 channel single board might be in, I'm assuming that it's best to have the leads from the BMS to battery should be short as practical, and should be part of the battery pack?

 

[PJD]

Pat, unfortunately, you haven't missed much. We, on the other hand, have missed deadline after deadline with the v4.x BMS. :x
After god nows how many iterations, I think we have the most stable design yet. No more oscillations and no more voltage drop issues and no more weirdness.

Well, the old, unsophisticated V1.5 boards continue to work well for me (or is there something I need to know?). Even when charging at 16A by doubling-up the chargers the masonite cover over the board only gets a bit warm over the charger control FET. They may not be capable of shutting charging off upon completion of charging, but they have held up to at least 4 hours of forgetting to unplug the charger when fully charged - basically putting the packs of "float" over this period, so I haven't worried too much.

What new features will the new boards have?

The only oddity is that the last cell at the positive end of the pack seems to get to a higher voltage (2.72 to 2.73) compared to the others (2.67 to 2.69), and this happens on both boards I assembled, so it isn't just some random variation in the LM431 or resistor values. This is still within charging tolerances.

At 3 years of calender age, (about 2 years of service age) the Thundersky cells themselves are degrading in the same manner in the scooter with 2500 miles on the pack as the one with 6000 miles on the pack - so calender life, not cycle life seems to be the governing factor. At 2.3C discharge, they now sag to about 2.9 volts at a temperature of 60F. Still plenty usable for another season, but probably not very good for winter use anymore. But I'm wandering off topic...

 

[ejonesss]

@ggoodrum when you select the cell connection protectors you may want to select ones with a high enough amp rating to act as a short to forcibly blow fuses in line with the cell connection or do the bose indestructible speaker effect and just light up a small light bulb indicating improper connection.

back in the 80's bose came out with a speaker that claims to be indestructible you could connect directly to a 110 outlet and it would buzz but you would never blow it because i think it used a fuse shaped light bulb similar to a dome light bulb for a car.

if the speaker was to be overloaded it would just light up the bulb limiting the current or something.


something similar could be put in line with the cell taps so if someone shorts the output of the battery they dont blow parts on the bms.

i should know i had to replaced a transistor and the lm431 and solder up a trace blown on the most + cell because i accidently touched the bullet connectors wrong while connecting ( another reason to possibly use anderson connections).