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

[Patrick]

Gary and Fechter,

I've got a unique circumstance on my bikes. The charger is integrated with the controller, and uses the same battery cables to charge. So, the BMS has to be connected whilst motoring to use the LVC. And at least one of the battery (charger) cables has to disconnected during charging in order to flow the charge through the BMS. So I'm thinking I can just put in a (high-current) switch on one leg (negative?) of the battery cables to disconnect the batteries and enable the BMS during charge. Not ideal, but OK.

Here's my question: what happens if the switch isn't thrown? That is, the charger is connected across the batts, and they are also connected to the BMS? Will this let the smoke out?

Thanks!

Patrick

 

[PJD]

Fechter >>>>> Can you please give Gary a call and ask him to update the group? Thanks,
--Tom_D

This is getting uncomfortably similar to the last attempt at this...

 

[GGoodrum]

Sorry, for the lack of updates. My life right now sucks! My consulting job that I had to take on, has absolutely consumed me, and they would take more of my time, if they could. Yesterday was my 25th wedding anniversary and I got home at 9:30. Not good. That is going to REALLY cost me.

I did, however, finally got some of the kits completed tonight, and these will go out, first thing in the morning. I will do as many as I can tomorrow night, and then the rest the next night. Whatever is left will go out Saturday.

-- Gary

 

[Jeremy Harris]

Sorry to hear of the hassle you're having, Gary, but in some ways too much work is a better thing than too little..........

I, for one, really appreciate the effort you're putting in to get these BMS units to us, so will patiently wait until you have the time to sort them.

I suppose whatever money you may have made from the consulting job will now be forfeit as compensation for missing your 25th anniversary. Missing stuff like that ALWAYS seems to cost more time and money than might seem reasonable...................

Jeremy

 

[JB_EB]

While I'm waiting to order my BMS I'm pushing ahead with building my new battery box.
Gary or Fechter - Would one of you be able to provide me with the dimensions of the BMS PCB please? I would also like the approx height with the connector and the low and high current resistors mounted if possible. I'd hate to make up the box and find I can't fit the BMS when I get it....
Thanks,
JohnB

 

[fechter]

Gary and Fechter,
So I'm thinking I can just put in a (high-current) switch on one leg (negative?) of the battery cables to disconnect the batteries and enable the BMS during charge. Not ideal, but OK.

Here's my question: what happens if the switch isn't thrown? That is, the charger is connected across the batts, and they are also connected to the BMS? Will this let the smoke out?

Thanks!

Patrick

Hmm... that could be a problem if you forgot the switch and went to run the motor at high current. The current would be passing through the body diode in the charge control FET. I suppose if you had a few IRFB4110's in parallel and mounted on a heat sink, they would survive depending on the current. If the current is low enough so the voltage drop in the diode is not a problem, you could just run it like that. Perhaps one of those "ideal diode" chips that turns on the FET whenever the body diode is conducting would allow for less heat.

Conversely, if you forgot the switch in the run position and went to charge, there would be no charge current control, and you could fry a cell.

You could possibly use a big relay with a 120vac coil and drive the coil with the AC input to the charger. This way it would be hard to forget. The relay could have a second set of contacts to switch on the charge control circuit. The LVC would function regardless of the switch position.

Otherwise, it should be possible to somehow separate the charger line from the battery line inside the controller.

On the size, I don't have mine with me right now, but the layout file indicates 7" x 3.5" area.
The height of the resistors is less than the other parts, like the capacitors and shunt transistors. The connector is probably the tallest part. I'd leave at least 1" for thickness.
All that heat has to go somewhere too.

 

[AndyH]

The USPS says my BMS kit is in the mail - Thanks Gary!

[shuffle shuffle..which pile is my soldering iron hiding under?]

[edit] It's here! [/edit]

 

[chardog]

Gary, any chance you make a bms for valence cells?

I just finished making a pack 15s8p with valence 18650 cells (3.2v or something) totalling 48v 10ah.

Edit: not sure if I should arrange 15 or 16 series. Right now only have 15.

 

[fechter]

Valence cells should be the same as A123 or Thundersky as far as voltage goes.

A 16 cell board can be used with 15 cells no problem. I heard the stock Valence BMS sort of sucks for balancing.

 

[chardog]

fetcher, I've read a lot of this thread, but have some things that need to be clarified.

1. I've already wired my setup into 8 groups of 15 series (which I think I messed up). Do I wire 15 groups of 8 parallel instead?

2. What is the verdict on chargers? SLA or lithium?

 

[JB_EB]

Hi Gary (or Fechter), is there any update on when new kits will be available for purchase at TP Products?

Thanks,

JohnB

 

[fechter]

fetcher, I've read a lot of this thread, but have some things that need to be clarified.

1. I've already wired my setup into 8 groups of 15 series (which I think I messed up). Do I wire 15 groups of 8 parallel instead?

2. What is the verdict on chargers? SLA or lithium?

You want 15 groups of 8 parallel.

No verdict on chargers yet, but just about anything with the right voltage should work. A tweakable SLA charger should automatically terminate the charge when the pack is fully charged if the voltage is set right. We need some actual testing to verify this behavior.

Sorry, I have no info on when the units will be available

 

[GGoodrum]

fetcher, I've read a lot of this thread, but have some things that need to be clarified.

1. I've already wired my setup into 8 groups of 15 series (which I think I messed up). Do I wire 15 groups of 8 parallel instead?

2. What is the verdict on chargers? SLA or lithium?

In order to match the "standard" way of matching LiFePO4-based packs to existing SLA setups, you would need to use 16 cells. This really has more to do with charging. LiFePO4 cells can be charged with the same CC/CV protocol used by SLA chargers, and it just works out that for charging purposes, six SLA cells roughly equal 4 LiFePO4 cells. SLAs need 2.40-2.45V as the CC/CV cutoff voltage, so for a 48V setup, which has 24 SLA cells, the chargers are typically set to around 58-59V. LiFePO4 cells need 3.65-3.70V as the cutoff point, so a 16-cell LiFePO4 pack can use the same 58-59V SLA charger.

Where it gets confusing is that the "nominal" voltages for SLAs and LiFePO4s don't quite track the same way. SLAs have a bigger voltage drop, once a load is applied, so the cell's nominal voltage drops to around 2.0V per cell, so for a 24-cell setup, it will have 48V as the nominal voltage. LiFePO4 cells hold the voltage up much higher, and have a nominal voltage of around 3.3V per cell. That means a 16-cell pack will have a nominal voltage of 52.8V, or about 4V higher than the SLA "equivalent" setup. Since the electronics needs to be rated high enough that it can handle an SLA battery fresh off the charger (i.e. -- 58-59V...), getting this "extra" 4V is an added performance benefit.

So, your 15s configuration will come closer to matching a 48V SLA setup, from a "nominal" point-of-view, your charging options are going to be a lot more limited. You will really need to look at using a lab supply, or a custom charger, as I don't know of a single SLA-based charger that can be "dialed down" enough for a 15s setup, which needs about 55V.

Also, if you want to use our BMS board with this setup, you need to connect all the cell junctions together, and then connect these to the BMS board. The cells need to be paralleled first, and then put in series.

Regarding the availability of more kits, I'm still struggling to get the last few out from the first batch. As I said earlier, I won't make any more available until I have time to put the kits together ahead of time. I'm afraid that is not going to happen anytime soon. What I will do soon, however, is make the PCBS available, along with a detailed Mouser/DigiKey parts list, and people can just order their own parts.

-- Gary

 

[lepton]

Thanks for the update Gary. I'd definitely appreciate being able to buy the PCBs independently. Sourcing parts isn't a big deal.

 

[Patrick]

So, your 15s configuration will come closer to matching a 48V SLA setup, from a "nominal" point-of-view, your charging options are going to be a lot more limited. You will really need to look at using a lab supply, or a custom charger, as I don't know of a single SLA-based charger that can be "dialed down" enough for a 15s setup, which needs about 55V.

-- Gary

The Zivan NG1 can do this. On the digital board (contains the MCU) inside, there are two 10 or 15 turn trim pots, one for current, and one for voltage. The voltage can be increased about 10% above nominal, and down to almost anything. I would be happy to post pictures if there is an interest.

*** Be VERY CAREFUL when working on an open Zivan *** Everything is live, including the heat sinks which can be at mains voltage or above.

 

[Stevil_Knevil]

So, your 15s configuration will come closer to matching a 48V SLA setup, from a "nominal" point-of-view, your charging options are going to be a lot more limited. You will really need to look at using a lab supply, or a custom charger, as I don't know of a single SLA-based charger that can be "dialed down" enough for a 15s setup, which needs about 55V.

-- Gary

The Zivan NG1 can do this. On the digital board (contains the MCU) inside, there are two 10 or 15 turn trim pots, one for current, and one for voltage. The voltage can be increased about 10% above nominal, and down to almost anything. I would be happy to post pictures if there is an interest.

*** Be VERY CAREFUL when working on an open Zivan *** Everything is live, including the heat sinks which can be at mains voltage or above.

Hi Patrick,

YES please! Would you mind starting a new thread so it does not get buried here? Many thanks.

 

[jeffkay]

Hello Endlessly-Spherical folks,
I have received my BMS order, built it and tested it on the 16S12P A123 pack of "Deep Cycle", a full size motorcycle conversion. I am very happy with the results and it was worth the wait. Earlier in this thread Gary describes the operation which I found to agree with charging my pack. As you may have noticed, one change in this design that differs from the earlier Bob/Gary attempt, is that there is now no real "end of charging" turn-off of the system. You will have to devise some way to time your charging and turn off the shunting, otherwise wasting heat and some degree of battery cycles. Other than working that out, I was happy to see 16 parallel subpacks all withing a few hundreds of a volt from each other--easily the best balance my pack has ever seen to date!

Thanks Gary, Richard (Fechter) and Bob for all the work that went into this project made for the EV community. I will also be incorporating this in the new EV project, "EV of Destruction", already rolling around on its three wheels...

Jeff K.

P.S. Richard, please publish the bump voltage resistor sub table for different top-of-charge voltages like 4.0 or 4.2, etc.

 

[fechter]

Nice to hear the kits are getting out. Thanks for the report.

Not terminating the charge may or may not be an issue. If the charging source voltage is set just right, you can get the charge current to taper off at the end, but you would slow down the balancing. I don't think it will hurt anything to just let it run until you disconnect it. Something we can work on for the next version....

I'll see about the resistor values. There's like a zillion possible combinations, so the trick is to find commonly available values.

What are you using for a charger?

 

[jeffkay]

Richard, My charger is a CV Vicor power supply. I have both 4 and 8 amp types. I can adjust my charger and currently have it set at 59.5V for the A123 pack but I have LIMn pack that can go to 4.0 or 4.2. What pot value would you recommend that could encompass from 3.6 to 4.2v? I remember you had pots on the earlier models but that increased setup/checkout time and I understand why Gary took it out.

Regarding the CV type of supply. It does work, but charging slows down near the last third as there is less current differential. Not really a big deal. The shunts actually help keep the current coming!
Jeff K.

 

[fechter]

To make it adjustable, you could replace the 75k fixed resistor (R1) with a 100k trimmer. This would allow adjustment to over 4.5v. It would be better to place a trimmer between the two fixed divider resistors to make the adjustment less sensitive, but that would require a different board layout. With the right trimmer pot, you might be able to fit it into the same spot as R1 and not need a new layout.

 

[PJD]

Richard or Gary,

My kit arived and is being assembled now, no problems so far, but my soldering skills are novice-level, so I'm going slow.

My chinese chargers are set for 59.2 to 59.5 volts during the CV stage, and this does not appear to be easily adjustable. The charging terminates (or specifically, stops, waits for surface charge voltage to bleed down to 55 volts then switches to float) based on a charging current setpoint. This setpoint IS adjustable, from probably 1 amp down to 250 mA or so.

I still don't quite understand how the BMS's current regulation behaves. Will it be possible to adjust the charger's current setpoint to, say, the shunt current plus a desired residual charging current, and get the charger to stop normally?


Jeff K,

Are you using the high-current (6.2 ohm resistor) shunts?

 

[jeffkay]

Jeff K,

Are you using the high-current (6.2 ohm resistor) shunts?

Hi-Current resistors. I have a 20 amp charger but I have not tried it yet. With that, you need to send separate large wires to the pack.
Jeff K.

 

[Jeremy Harris]

Many thanks Gary, Richard and Bob.

My kit arrived from Gary today (having travelled half way around the world to reach me), so this evening will be build and test time!

Jeremy

 

[GGoodrum]

Paul -- Your charger voltage should be fine. The 180k/75k resistors used in the voltage dividers will set the shunt turnon point at just under 3.70V per cell, or about 59V even. If the float voltage is also right around 59.2-59.5V, it won't matter whether or not you leave it on. If the cells are full, they aren't going to take any more current, so the shunts will bypass the full 500mA, or so. You can set the current for the switch to the float mode to anything under 500mA. If you do, it shouldn't ever go into the float mode, as there will always be at least 500mA going through the BMS, if not the cells.

What the shunts do is simply let the next cell down the line have at least 500mA of current, if it needs to "catch up" with the cells that are fuller. Let's say that the first cell in series starts to get full, quite a bit sooner than the others, which are still in the CC mode. Without the shunts, but assuming it does have a limit on its voltage to 3.7V, it would start dropping the current it lets in. Since all the current HAS to go through all the cells, the less full cells will not ever finish. In reality, unless you got a really bad cell, what usually happens is that most of the cells are at about the same level, and as they hit around 3.65-3.70V, it suddenly becomes much harder to accept any more current, so the voltage starts to rise at a very quick rate. Depending on how many cells are at that point, you could have the voltage for these hit anywhere from about 3.75V to 3.90V, or higher. This will cause the collective voltage to hit the chargers CC/CV crossover point (59.2-59.5V, in your case...), which causes the current to start to drop. At some point, the current will drop to something close to zero. You could very well have the case where several cells never even reached the 3.65V point, so they only end up charged to the 80-85% level. With the shunt-based logic, each of these low cells can still have up to 500mA available, if they need it, to get to the same level as the rest.

-- Gary

 

[ErikK]

Hi y'all, Erik K, newbie here. Wow, I read this whole thread and the previous one and am very impressed. What a great project, and what a cool example of collaborative effort. I'll be very happy to buy a pair for my bike once the availability problems are solved.

Gary, if you are buried under too many things to do, just ask for help. I am sure there are people here who could help you. A very simple one is that somebody could sell parts kits, so you only need to sell bare boards. Very little coordination or financial commitment needed on either side. Or somebody could fulfill the orders for you, just like the outfit in China that builds and ships Xboxes for Microsoft. That would take a bit more coordination of course but it could be very good for you to get out from under the burdens of production and shipping and deadlines, and also good for the customers to get the BMS soon. There are ways to do this so that you retain the parts of it that you want to keep, while others do the parts you don't want to do.

I personally don't especially want this job. In fact if somebody wants to do this for my products I'd be happy to pay good money! Since (unlike my poor products which are not open source and not in the EV world) this BMS is a group effort, you may well be able to find a volunteer. Maybe an enthusiastic individual who has more time than you do, or maybe a company that is set up to do this easily.

Anyway, congratulations on a super project! You have made a cool contribution to this little corner of the world.