JB_EB said:
Correct. The connector to the charger needs two high current contacts, and two low current contacts (or an integrated switch on the BMS side).
Even Newer 4 to 24-cell Battery Management System (BMS)
- Thread starter GGoodrum
- Start date
Correct. The connector to the charger needs two high current contacts, and two low current contacts (or an integrated switch on the BMS side).
GGoodrum
1 MW
JB_EB said:
That is correct.
JB_EB said:
Actually, the charger/supply just needs to be a few millivolts above the sum of the individual cutoff voltages. The cutoff voltages are a function of the voltage divider resistors, so you can adjust the number by changing these values, if you needed to. With the 180k/75k 1% resistors supplied, the number ends up being somewhere between 3.65-3.69V per cell. To go lower than this, the 75K value can be lowered.
-- Gary
slayer
1 kW
So are you saying for around 90.00 kit to be aasembled you get both the cms and lvc combined in this bms kit
GGoodrum
1 MW
slayer said:
Hopefully less than $90, but yes, that is correct.
deVries
100 kW
GGoodrum said:
Four questions:
1) Anyone have links to the best ideas for chargers or dc converters for this?
For example: I want to use a total of 56 a123 DeWalt cells in a 4 parallel configuration. So it's a 14s4p set-up at the theoretical voltage of 14x3.65 = 51.1v So, I guess I would need at least a 50v DC converter that should work fine. Correct? Best Buy Suggestions are: ?
2) Also, will the above example work with the circuit board even if I'm only using 14s4p of the 16s connection slots that are available? Correct? In other words, I don't have to use all 16s cell slots on the circuit board for any lesser number of Xs lower voltage set-ups I use? I can use any number available. Correct?
3) Are these kits already pre-assembled w/soldering complete? What does the buyer have to do to complete the kit... like the old Heath Kits?
4) It's my understanding that one circuit board will work with several identical parallel packs of x# up to 16 cell packs in x# of parallel packs. As I think Gary put it simply, the bucket (each parallel pack) simply overflows from charging one parallel pack to the next until the entire parallel set is fully charged. Correct?
Thanks!
GGoodrum
1 MW
deVries said:
Since you are going for a "non-standard" configuration (i.e. -- not a multiple of 4 cells in series...), you are cutting out the possibility of using an existing CC/CV-type charger, so I would look on ebay for a variable supply. Awhile back, I picked up an HP 0-60V/0-15A supply on ebay for $125.
deVries said:
Yes, you can use this with any number of cells/channels, 4 or over. You just jumper across the unused positions.
deVries said:
Assembly is not included. You get the board, separate parts bags and a detailed set of instructions. If you can't do basic soldering, this kit is probably not for you. I suspect, however, that some here will be able to offer their assembly services. That said, it is not terribly difficult to put together. Testing is very simple as well.
deVries said:
Not quite. You need to connect cells in parallel first, and then to the board. You don't connect multiple series packs in parallel.
-- Gary
I have my eye on some surplus switching power supplies. Since they have isolated outputs, I can put units in series to get the desired voltage. If I take about 3.7v x the number of cells, I get about 59.2v for 16 cells.
I should look for a power supply that runs at about 60v. Most have a trim adjustment for the voltage, which I could turn down a bit. I could use a 48v supply in series with a 12v supply also.
I should look for a power supply that runs at about 60v. Most have a trim adjustment for the voltage, which I could turn down a bit. I could use a 48v supply in series with a 12v supply also.
deVries
100 kW
From your website: Unique "stick pack" construction technique groups four blocks of four paralleled cells, one block in front of the other, with the LVC board on the top of the stack. Between blocks are G10 plates with nickel-plated springs, which are used to make connections with the cells. A threaded rod down the middle holds everything snugly in place. Should cells ever need replacing, this can be accomplished with a simple partial dis-assembly of the stack. No messy soldering is required at all. Multiple 12V/9.2Ah "stick packs" can be connected in series in order to be used in 24V, 36V, 48V and/or 72V setups.GGoodrum said:
Can you do a special ES member combo deal
for a "limited time" where one can order the battery assembly stick-packs for 4sx4p x #needed for voltage choice (minus its LVC board) as an add-on for your new LVC-charging BMS system with the necessary mods to make it work? In other words, for example, 4 stick packs to assemble 48+v PLUS your new board?OR
Does anyone know of good sources for battery assembly using solderless battery kits that can be purchased? I'm going to use 26650 batteries Dimensions: 26mm x 65mm for mine, and I have to avoid soldering tabs to these babies. Not possible for me to do that, spot weld tabs, so I need a solderless solution.
Thoughts and suggestions?
Thanks!
shinyballs
1 kW
For this integrated LVC/BMS board, is it BAD to position it far from the battery? In my setup, all the batteries are mounted in the center frame and controller in the rear rack. And this board will be inside a box clamped in the seatpost.
btw, great work guys!
btw, great work guys!
GGoodrum
1 MW
I had a bit of a setupback today. It turns out the KSA931 PNP transistor that we picked for driving the shunt resistor, has the collector and base leads reversed, from the normal E-B-C configuration. I didn't know this before installing 16 of these on the new board. This caused the orange LEDs to come on too early, at about 2.6V. It took me awhile to find the problem, which was bad enough, but then it took me the next 10-1/2 hours to cut out the bad ones, clean out the holes, bend the leads around on 16 new ones, and then install/solder them. The good news is that the shunts come on at about 3.7V now, but the bad news is that six of the orange LEDs aren't coming on.
The bottomline is that the kits are not going to be ready this week, like I had hoped. I'm going to have to either pick another part, with the standard E-B-C piniout, or scrap these boards, and do a new one. In any case, Richard has all the parts and a board, so he will continue on with more testing. One thing we still need to get is a baseline on the amount of bypass current and heat that is generated. I did see that when I got the shunts working, the current it passed was close to an amp, and the big 5W resistor never got above about 150F. If that holds true with all 16 channels operating, that would be good news indeed.
-- Gary
The bottomline is that the kits are not going to be ready this week, like I had hoped. I'm going to have to either pick another part, with the standard E-B-C piniout, or scrap these boards, and do a new one. In any case, Richard has all the parts and a board, so he will continue on with more testing. One thing we still need to get is a baseline on the amount of bypass current and heat that is generated. I did see that when I got the shunts working, the current it passed was close to an amp, and the big 5W resistor never got above about 150F. If that holds true with all 16 channels operating, that would be good news indeed.
-- Gary
shinyballs
1 kW
GGoodrum Is it critical that this board be very close to the batteries or it still works fine if farther by more than 4 feet of wire?
GGoodrum
1 MW
shinyballs said:
I think four feet is too much. I wouldn't have the leads longer than about 18 inches. I know someone here is using 30" leads, because he had me make them, but that's the longest I've seen..
-- Gary
GGoodrum said:
Oops, yep, I screwed up.
There are hundreds of transistors that are similar and I was paying more attention to the ratings than the pin configuration. This was real bad for Gary, since it resulted in a crapload of needless work and a big pile of transistors that won't work with the board. One the one hand we're trying to hurry to get these out to people who need them, but on the other hand we need to make sure eveything works properly. I just got the newest version board in the mail, and I'll try building and testing it as soon as I can. At least I know ahead of time to reconfigure the leads on the transistors.
JB_EB
10 mW
Bad luck about the transistors. The joys of prototyping. Don't you just love the "standard pinouts"!
I have a low-tech question. I will be charging an 8s5p string so need ~ 30VDC.
I have a 240VAC to 24VAC 3A secondary transformer (240V, 50Hz mains here in Australia) and I was wondering if I added a bridge rectifier and a few thousand uF of capacitance, if I could use this unregulated linear power supply to drive the BMS? The output would be ~34V unregulated DC with some 100Hz ripple.
Or should we only use a regulated supply.
I would be happy to take 4-5 hours to charge overnight, so don't really need a high current power supply.
(Forgive me - it is my Scottish ancestry coming to the fore and looking at alternative cheap solutions.
)
JohnB
The joys of prototyping. Don't you just love the "standard pinouts"!I have a low-tech question. I will be charging an 8s5p string so need ~ 30VDC.
I have a 240VAC to 24VAC 3A secondary transformer (240V, 50Hz mains here in Australia) and I was wondering if I added a bridge rectifier and a few thousand uF of capacitance, if I could use this unregulated linear power supply to drive the BMS? The output would be ~34V unregulated DC with some 100Hz ripple.
Or should we only use a regulated supply.
I would be happy to take 4-5 hours to charge overnight, so don't really need a high current power supply.
(Forgive me - it is my Scottish ancestry coming to the fore and looking at alternative cheap solutions.
)JohnB
Jeremy Harris
100 MW
I've been caught by the dreaded TO92 transistor pin out problem before now, so you have my sympathy - it's a dead easy mistake to make. I've always thought it crazy to use different pin out designations for the same package. The one that got me years ago was the BC182 - the damn thing is available in two different pin outs in the same package, with similar part numbers (the BC182L is BCE, the BC182A is CBE). When the smoke cleared it was an easy fix for me though, just go and buy the other version...................
Jeremy
Jeremy
JB_EB said:I have a low-tech question. I will be charging an 8s5p string so need ~ 30VDC.
I have a 240VAC to 24VAC 3A secondary transformer (240V, 50Hz mains here in Australia) and I was wondering if I added a bridge rectifier and a few thousand uF of capacitance, if I could use this unregulated linear power supply to drive the BMS? The output would be ~34V unregulated DC with some 100Hz ripple.
Or should we only use a regulated supply.
I would be happy to take 4-5 hours to charge overnight, so don't really need a high current power supply.
(Forgive me - it is my Scottish ancestry coming to the fore and looking at alternative cheap solutions.
JohnB
Yes, I think you should be able to get away with that as long as the voltage stays in a good range.
One problem will be limiting the current. With a low pack, the thing may try to draw big amps, so you may need a big honkin resistor or just long charger wires to limit the max current.
deVries
100 kW
deVries said:
GGoodrum said:
-- Thanks Gary & everyone else for helping out too. This is my first build, and I'm not even an electronics hobbyist, yet! But I'm trying. I can see upon further reflection my thinking was/is confused to begin with. So, I will rephrase, hoping this will be ontrack???...
I'm going to make 16 4p cell-packs that I need to eventually connect in series. With your board I should connect each 4cell p-pack to each of your 16 LVC/Balancing connections. With my set-up I would use all 16 of your slots/ports in 4p-4cell configuration x 16 times. Assuming this is correct, then how should I connect the 4cell p-packs in series keeping that "separated" or "isolated" from the 4cellp connections? Won't that change all the voltages getting to each 16 port slot too once everything is connected in series? Sorry, I know I'm just talking newbie... maybe nonsense...
I'm confused about how the circuit can just isolate and deal with the 16 4cell p-packs each connected to the 16 slot/port connections, when these 4cell p-packs also have to be in series too. (Sorry I'm such a rank newbie in understanding all this!)
Hmm, Once you have connected the 4 individual cells into a parallel group you now need to think of this 4parallel group as a single cell. You then connect 16 of these 4parallel groups in series.
When you measure the voltage from end to end on the 16 cells in series you will get a nominal 48v (closer to 60v hot off the charger)
The BMS has a connection to each junction between adjacent 4p blocks. If you were to measure the voltage from one side of a 4p block to the other you will see a nominal 3.2v. It does not matter which 4p battery you measure in the series string, it will still be 3.2v nominal. Remember that you are measuring the voltage with respect to whatever reference point you happen to place your negative voltmeter lead on, not the combined series voltage.
When you measure the voltage from end to end on the 16 cells in series you will get a nominal 48v (closer to 60v hot off the charger)
The BMS has a connection to each junction between adjacent 4p blocks. If you were to measure the voltage from one side of a 4p block to the other you will see a nominal 3.2v. It does not matter which 4p battery you measure in the series string, it will still be 3.2v nominal. Remember that you are measuring the voltage with respect to whatever reference point you happen to place your negative voltmeter lead on, not the combined series voltage.
deVries
100 kW
Ok, seems I needed to modify my configuration anyway for packaging and mounting purposes. :wink: :lol:GGoodrum said:
Now, I'm going with 4p packs connected to the 16 slot/ports on the board. When these 16 4p packs are connected in series it will top off at a fraction over 59v.
For set-up to use existing CC/CV-type charger what batt-series configuration should I use? What chargers would you recommend to match your best ideas and suggestions?
Anyone?
TIA.
slayer
1 kW
Soneil are good for 16 series
http://soneil.com/48_volt.html
http://soneil.com/48_volt.html
deVries
100 kW
fechter said:
From my point of view I want to charge as fast as possible without risking any damage to the cycle lifespan or reducing capacity... Is there an ideal AMP charge range to stay within?
Also, the battery specs I have A123racing suggest overcharging the batteries slightly to 3.9v will help keep the batteries balanced and not really affect the lifespan of the batts. Will it be easy to mod your BMS to cut-off at this higher voltage instead? Cost to do it? Also, I'm told the batts can be discharged down to 2.4 volts without problems. Again, will it be easy to mod your LVC to cut-off at this lower voltage instead? Cost to do it?
TIA.
The cutoff voltage is a function of the resistors installed on the voltage divider for each channel. With the values we're using presently, I'm getting 3.68 volts. To change the voltage would require changing the resistors. You could pick resistor values to get any cutoff voltage you want (within reason).
3.9v sounds too high, but they're in the business of selling batteries...
3.9v sounds too high, but they're in the business of selling batteries...
deVries
100 kW
Three more questions for you (please&thankyou)...
2) Also, I'm told the batts can be discharged down to 2.4 volts without problems. (This is the Racing RC version of the batt.) Again, will it be easy to mod your LVC to cut-off at this lower voltage instead? Just another 16 resistor change-out to go from 2.7 to 2.4 or ?
3) Will the corrected board be 'final tested' this week if all goes according to soldering plans? :wink: (Sorry Gary had that set-back... bummer.)
for you (please&thankyou)......fechter said:
1) Any thoughts about this? ...Is there an ideal AMP charge range to stay within for max. lifespan & cycles? (I will have 4p packs connected to all 16 slots/ports on your BMS.) Example: if I had a 20ah charger it will fill each 4p-cell pack at 5ah per cell? 20ah/4cells=5ah juice going into each cell in the 4pcell pack? Correct?deVries said:
2) Also, I'm told the batts can be discharged down to 2.4 volts without problems. (This is the Racing RC version of the batt.) Again, will it be easy to mod your LVC to cut-off at this lower voltage instead? Just another 16 resistor change-out to go from 2.7 to 2.4 or ?
3) Will the corrected board be 'final tested' this week if all goes according to soldering plans? :wink: (Sorry Gary had that set-back... bummer.)
GGoodrum
1 MW
a123 cells are good for 20A charge rates, and with 4 cells in parallel, each would only see 5A. In anycase, with a 20A charger, you can recharge a 4p/9.2Ah in about a half-hour.
You can discharge a123 cells to about 2V, and not damage them. However, once the capacity is used, the voltage drops very quickly. If you do't catch this, the voltage can go all the way to zero, permanently killing the cell. What the LVC portion of the BMS does is to monitor each cell's voltage, and if any one cell, or block of paralleled cells, has its voltage drop below 2.1V, under load, it will trip a signal that can be connected into a controller's brake input, which will cause it to cut the throttle, removing the load. The voltage will recover to something above 3V, which reenables the the throttle. If the throttle is still full-on, the voltage will drop again, cutting the throttle again. Usually, if you back off the throttle about halfway, you can go another mile, or two, before it cuts out for good.
Richard's found a few things, and has been tweaking some resistor values. He's also testing adding a small cap on the gate driver input, in order to make the regulation a bit more uniform. As soon as he's done, I'll see whether or not we can move forward with the existing boards, or if a new run is required.
-- Gary
You can discharge a123 cells to about 2V, and not damage them. However, once the capacity is used, the voltage drops very quickly. If you do't catch this, the voltage can go all the way to zero, permanently killing the cell. What the LVC portion of the BMS does is to monitor each cell's voltage, and if any one cell, or block of paralleled cells, has its voltage drop below 2.1V, under load, it will trip a signal that can be connected into a controller's brake input, which will cause it to cut the throttle, removing the load. The voltage will recover to something above 3V, which reenables the the throttle. If the throttle is still full-on, the voltage will drop again, cutting the throttle again. Usually, if you back off the throttle about halfway, you can go another mile, or two, before it cuts out for good.
Richard's found a few things, and has been tweaking some resistor values. He's also testing adding a small cap on the gate driver input, in order to make the regulation a bit more uniform. As soon as he's done, I'll see whether or not we can move forward with the existing boards, or if a new run is required.
-- Gary
I'm getting really close. I have a few more values to tweak.
Similar threads
