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

[GGoodrum]

The only questions I'll have when the boards are ready, is...

1. How much for/where to order all the parts?

2. And, how to wire it to the battery.


Other than that. I'll be good to go.

Laaa-Deee-Daaa-Daaa-Daaa... New battery, on the way soon.

The parts for a 16-channel version will be about $60. I will provide a parts list text file in a format that can be submitted as a "Bill of Material" (BOM) to Mouser (http://www.mouser.com), which will save having to enter each part number, entering quantities, etc.

I'm working on an update to the illustrated instructions that I did for the previous version. In it will be the diagrams and instructions on how to connect the cells. The connections are numbered on the board, so it is pretty bullet-proof, but I highly recommend having a separate connector, or connectors, that go to the cell junctions, with a matching plug, or plugs, that go to the board. That way, you can do two things, test the connections, to make sure you got them right; and ensure that the cells get connected to the channels on the board in sequence.

I'm also a big fan of the .100-style RC connectors, and use a lot of them. I also like the AMP 4.2mm PE series, Molex-styled, connectors, mainly because you can have a single plug connections, but also because the pins in these are rated to 9A. The come in a variety of sizes, as well. What you need is a connector with one pin more than the number of cells, if you are going to use a single plug. For the previous 16-channel board, I used the 18-pin plug and connector, as they only come in even numbers.

I will also have diagrams illustrating how to break up a 24-cell setup into multiple boards that can still all be used together. An example of this would be to have three 8-cell sub-packs, that are connected in series for discharging, but in parallel for charging. In this case, the board would be split into three 8-cell sections, with the first one having the charger control portion, in addition to the first 8 channels. The "slave" boards are then connected to the other two 8-cell packs. The bus lines from each of the slave boards can then be daisychained to the "master" board. The sup-pack + and - lines are brought out separately and are connected in series for discharging, and in parallel, for charging. Each cell's shunt logic is separate, and self-contained, and so it matters not how the packs are configured as long as each channel can see the + and - connections for the cell it is monitoring, and as long as all the output bus lines are still all connected.

-- Gary

 

[jeffkay]

Gary, What will this board cost? How many channels in what break-offable chunks? How many wires exist between chunks? I could not find the orange LEDs on mouser using your numbers. Can you check it? Are they diffused or clear?
Jeff K.

 

[GGoodrum]

Gary, What will this board cost? How many channels in what break-offable chunks? How many wires exist between chunks? I could not find the orange LEDs on mouser using your numbers. Can you check it? Are they diffused or clear?
Jeff K.

I don't know what the board will cost until I know how many will get ordered. I haven't made that decision yet. I sent one of the new ones to Richard, which he might have today. Once he's happy that the new board works like his "blenderized" prototype, I will place the order. I need to be sure, this time, that the design is reasonably stable, and that we don't need to make any more board changes, before committing to a large quantity of boards. The only way to keep the costs reasonable is to order them in sufficient quantities. What I need to avoid is ending up getting stuck with another pile of unusable boards. All that said, I'm hoping to keep the board cost under $50.

There are 24 channels total, with "breaks" every 4 channels. These breaks are implemented using a line of closely-spaced, small holes. It is pretty easy to cut along the "perforations", and end up with a fairly clean edge.

There are six bus lines that need to be connected. I have provided holes for this purpose, which are circled in yellow, in the example below:

View attachment BMS Bus Line Connections.jpg


The LEDs are not critical. Almost any red or orange LED will do. The orange ones I used ae not difused, but the difused ones will work just as well. I don't have the part numbers with me, right now, but I will test all the parts when I make up the BOM file.

-- Gary

 

[Fishmasterdan]

Patiently waiting. I would like to get my hands on one these

 

[Daniel828]

Thanks Gary
--Dan

 

[Doctorbass]

put me on the list Gary!.. ...for the 24s version for sure!

Could you update us on how many order you have and how many you expect to have?

Doc

 

[methods]

Read back like 3 posts

 

[Patriot]

Question.

If I decide to take my charger with me, mounted on the bike, can I leave the charger connected to the battery while discharging when riding?

I ask, because I'd like to mount the charger in a spot that is kind of buried, with only the 100vac cord exposed for charging when I get home or at work.

 

[Doctorbass]

Read back like 3 posts

Thanks Methods :wink: , i just hope we will be enough to place an order.

Doc

 

[fechter]

Question.

If I decide to take my charger with me, mounted on the bike, can I leave the charger connected to the battery while discharging when riding?

I ask, because I'd like to mount the charger in a spot that is kind of buried, with only the 100vac cord exposed for charging when I get home or at work.

It depends a bit on your charger. Some chargers will put a significant drain on your batteries when they are not powered. The other issue is the jumper for the control circuit. You can leave the charger connections intact and run a pair of small wires out to a switch or relay to turn the BMS control circuit on. With a switch, you just have to remember to turn it off when you finish charging or the BMS circuit will drain the pack (around 25ma). At this rate, it would take weeks to deplete a pack, but it would be best to turn it off. This drain is only on the first 4 cells, which over a long time will seriously imbalance those cells. This would require a longer balancing time during recharge.

You could also use a small relay with a 120vac coil. The coil would be wired in parallel with the charger input so it activates anytime the charger is plugged into the wall.

I've been working on a solution to make the BMS control circuit power switch automatically, but this is turning out to be very difficult. It can be done easily with a single diode, but the diode needs to handle the full charge current, which would require a heat sink. A fairly typical schottkey would need to dissipate around 15W at 20 amps. For charge currents below around 5 amps, you could use a diode with a relatively small heat sink. The diode could be mounted off the board somewhere extended by wires. The voltage drop across the diode would require that the charger voltage be increased by about .5v to compensate.

 

[Patriot]

So... in other words...


Unplug it.

 

[Tom_D]

You could also use a small relay with a 120vac coil. The coil would be wired in parallel with the charger input so it activates anytime the charger is plugged into the wall.

Richard, could you suggest a relay from Mouser that would do, like one of these: http://www.mouser.com/catalog/635/1717.pdf? Then, I can add to my order along with the other parts.

So, my understanding is that with this relay in parallel to the charger's 120v line, the BMS control circuit would be switched on (via the board jumper) when the charger was turned on, CORRECT ??... and that if I used a Soneil 4808 SRF charger, I could leave it connected all the time to the battery pack when not charging, if you know, CORRECT ??


Thanks, Just trying to finalize all elements that go into the battery box.
Tom_D

 

[fechter]

So, my understanding is that with this relay in parallel to the charger's 120v line, the BMS control circuit would be switched on (via the board jumper) when the charger was turned on, CORRECT ??... and that if I used a Soneil 4808 SRF charger, I could leave it connected all the time to the battery pack when not charging, if you know, CORRECT ??


Thanks, Just trying to finalize all elements that go into the battery box.
Tom_D

Yes on connecting the relay.
Here are the 3 cheapest ones I found on Mouser that would work for the control signal:

655-RT424615
655-RTB14615
893-SCLBWDPDTC120VAC

I believe most of these are PCB mount, but you can wrap wires around the pins and solder directly. Make sure to carefully insulate all the connections! Fat heat shrink or something involving hot melt glue might work well.

Most of these have multiple sets of contacts. You only need one set for the control circuit.

I'm not sure if the Soniel will put a significant drain on the pack when it is off. I don't think it will be a problem, but to be sure, you could measure it sometime.

To be absolutely sure, you could use a DPDT relay and use one side for the control circuit and the other side to disconnect one leg of the charger output. In this case the contacts need to be rated higher than the charger output.

 

[Tom_D]

To be absolutely sure, you could use a DPDT relay and use one side for the control circuit and the other side to disconnect one leg of the charger output. In this case the contacts need to be rated higher than the charger output.

Richard, thanks for all your tips, and especially this one! I like the idea of totally isolating the charger from both the BMS and battery pack when not activated.
Now added to the design.
--Tom_D

 

[manfred59]

I have been finished building my 16 cell board. I have tested with a Meanwell PSU max. 56V/6,5A. With 59V all orange leds were on, even the main led, so everything should be ok.
But I have some questions:
1. I only have 15 cells in my pack and when I tried to connect the 18pin connectors to my battery, then the last of the 16 leds was on (because I have only 15 cells I think).
The current on my WattsUp showed 0,3A at 59,3V.
2. I need the BMS only for charging, because my controller has a LVC and I never will drain my battery to a dangerous level.
So, how could I change my board for charging at full power with 15 cells. I tried to desolder the 5W resistor, but this does not match!
3. What voltage do I need for charging 15 cells. Should this be exactly 55V(3,65V/cell) or could it be some Volts more?
Manfred

 

[ejonesss]

for 2 you need to have a bms because if the cells get out of balance then they could be ruined.

correct me if i am wrong but the new bms is designed to detect if any 1 or more cells goes too low and prevents over discharging otherwise you could ruin or shorten the life of the cells.

I have been finished building my 16 cell board. I have tested with a Meanwell PSU max. 56V/6,5A. With 60V all orange leds were on, even the main led, so everything should be ok.
But I have some questions:
1. I only have 15 cells in my pack and when I tried to connect the 18pin connectors to my battery, then the last of the 16 leds was on (because I have only 15 cells I think).
The current on my WattsUp showed 0,3A at 56V. But after 10 minutes the orange led (16th) no longer shined and the main led changed to more red colour.
2. I need the BMS only for charging, because my controller has a LVC and I never will drain my battery to a dangerous level.
So, how could I change my board for charging at full power with 15 cells. I tried to desolder the 5W resistor, but this does not match!
3. What voltage do I need for charging 15 cells. Should this be exactly 55V(3,65V/cell) or could it be some Volts more?
Manfred

 

[methods]

If he uses the BMS for charging then the cells will be balanced every time.
The BMS does not balance the cells when not charging
He stated that he has an LVC on his controller and wants to use that.

A BMS system is a nice feature to have.
It is not by any means necessary and some may prefer simplicity over added safety.

-methods

 

[fechter]

I have been finished building my 16 cell board. I have tested with a Meanwell PSU max. 56V/6,5A. With 59V all orange leds were on, even the main led, so everything should be ok.
But I have some questions:
1. I only have 15 cells in my pack and when I tried to connect the 18pin connectors to my battery, then the last of the 16 leds was on (because I have only 15 cells I think).
The current on my WattsUp showed 0,3A at 59,3V.
2. I need the BMS only for charging, because my controller has a LVC and I never will drain my battery to a dangerous level.
So, how could I change my board for charging at full power with 15 cells. I tried to desolder the 5W resistor, but this does not match!
3. What voltage do I need for charging 15 cells. Should this be exactly 55V(3,65V/cell) or could it be some Volts more?
Manfred

To use fewer cells, you do not need to disconnect any components.
The positive-most channel on the BMS that goes to pack (+) should be jumpered so that pack (+) now goes to the 15th channel instead of the 16th channel (hope that makes sense).

With the PSU attached to charger (-) and to cell 15 (+), and no cells attached, you should have 15 LEDs lit and #16 not lit.

It sounds like you bypassed the wrong end of the string and put 2 cells worth of voltage into the last shunt (bet the resistor gets really hot).

I can work on a diagram of this. I know it can be confusing.

The BMS measures each cell. A controller LVC measures the pack. It is possible for a cell to go low before the LVC on the controller is reached. It is safer to use the LVC line on the BMS to short out the throttle or ebrake line. Alternately, you could wire it to a 9v battery and an audio alarm. If your pack is in good shape and you never run it too low, the BMS based LVC should never trip.

 

[GGoodrum]

I have been finished building my 16 cell board. I have tested with a Meanwell PSU max. 56V/6,5A. With 59V all orange leds were on, even the main led, so everything should be ok.
But I have some questions:
1. I only have 15 cells in my pack and when I tried to connect the 18pin connectors to my battery, then the last of the 16 leds was on (because I have only 15 cells I think).
The current on my WattsUp showed 0,3A at 59,3V.
2. I need the BMS only for charging, because my controller has a LVC and I never will drain my battery to a dangerous level.
So, how could I change my board for charging at full power with 15 cells. I tried to desolder the 5W resistor, but this does not match!
3. What voltage do I need for charging 15 cells. Should this be exactly 55V(3,65V/cell) or could it be some Volts more?
Manfred

1. Yes, if you only have 15 cells, you would only want to populate 15 channels.
2. Not draining your pack too far is okay, but it is not good to soley rely on the controller's LVC, unless you know for a fact that all cells are healthy. One way to know when you might be having a problem with a cell is to keep an eye on when its orange LED comes on, in relation to the rest. With healthy cells, the LEDs will all come on within a few minutes, from the first one to the last one. If a cell is weak, and is losing capacity, its LED will come on a lot sooner than the rest.
3. Ideally, the charger voltage should be about 3.7V per cell, so in your case, 55.5V. The shunts kick in at about 3.68V, with the 180.0k/75.0k 1% resistors we've selected, and you want the charger voltage to be just a bit above the sum of the shunt voltages.

-- Gary

 

[manfred59]

Thanks for your answers, but the important question is not answered untill now!
Untill now I have ~100 cycles on my a123 battery pack and the cells are good in balance!

1. Yes, if you only have 15 cells, you would only want to populate 15 channels.

That means I have to loosen all components for channel #16 or could I disconnect a trace and set a jumper or only one component (which one)?
And later if I want to charge 16 cells with this board it should be as easy to restore the BMS for use with 16 cells. A schematic or picture would be fine for me for better understanding.

 

[GGoodrum]

Thanks for your answers, but the important question is not answered untill now!
Untill now I have ~100 cycles on my a123 battery pack and the cells are good in balance!

1. Yes, if you only have 15 cells, you would only want to populate 15 channels.

That means I have to loosen all components for channel #16 or could I disconnect a trace and set a jumper or only one component (which one)?
And later if I want to charge 16 cells with this board it should be as easy to restore the BMS for use with 16 cells. A schematic or picture would be fine for me for better understanding.

I believe all you need to do is short pins 17 and 16 together, as shown below, and that should disable/bypass the 16th channel.

View attachment 15-Cell BMS Mod.png


If you were starting out from scratch, and only ever wanted 15 channels, you wouldn't populate the last channel, and you'd still do the jumper I've shown above.

-- Gary

 

[fechter]

Right, it won't hurt to leave the unused channels populated with parts in this version.

 

[dnmun]

manfred, it wouldn't hurt to buy another cell and add another 7% capacity, since you already have BMS populated for it, and you have a charger which will get to 59V. that's gotta be simpler.

 

[manfred59]

I believe all you need to do is short pins 17 and 16 together, as shown below, and that should disable/bypass the 16th channel.

Thanks to Gary, it works fine!
There are other problems for me, and I think someone will have a solution for me.
I am using a MeanWell SP-320-48/321W PSU and if I connect my BMS and 15s battery at 56V the current rises to 7,8A (436W!!) and the PSU starts pulsing also the fan of the PSU.
1. I think it would be necessary to limit the current to 5A or 6A. Any solutions for the MeanWell? There is only a Voltage regulation. Features: http://www.meanwelldirect.co.uk/public/ranges/pdfs\r376\r376_3.pdf
2. Could I terminate the charging process (BMS v.1.5) with a timer to one or two hours without disconnecting the battery, if I want to charge over night in the garage?

 

[ZapPat]

Manfred - Check this thread out for some DC power supply current limiting ideas: Making a CC/CV charger