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

[ejonesss]

things seems to be quiet here could it be http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 changing the direction on the up coming bms 4.0?

(standard obligatory question) not to sound impatient but it has been a couple weeks since it was said that the bms 4.0 would be coming out

oh well maybe a blessing in disguise with the advent of http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 if it works out better then i will go towards that.

 

[fechter]

things seems to be quiet here could it be http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 changing the direction on the up coming bms 4.0?

(standard obligatory question) not to sound impatient but it has been a couple weeks since it was said that the bms 4.0 would be coming out

oh well maybe a blessing in disguise with the advent of http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 if it works out better then i will go towards that.

Err, uh, I guess that one is going to be Ver.5.

Based on previous experience, there is a good chance that "Ver.5" will have some fatal flaw that prevents it from becoming practical. Even without fatal flaws, it will take a long time to validate the design and get it to a stage where it can be sold. Meanwhile Ver.4 has a lot of work put into it and should work well until something better comes out.

 

[GGoodrum]

I'm sorry for the delay, but I needed a break. I spent the long holiday weekend in San Diego, but I'm heading home today. I still need to finish up the layout for what will become v4.1. I'll try and get this finished, and into a new run of boards in the next day, or so.

-- Gary

 

[The Mighty Volt]

Is there a reliable and proven 18s version that can handle high currents and voltages of around 55-60v?

Also, do the owners/supplier of these boards provide any kind of after-sales support viz wiring etc?

Thanks and good luck with the new board.

 

[ejonesss]

the only fatal flaw i can se possible is that capacitors in general can not handle massive amount of current through them in series for extended periods and will short out and void out all isolation or even short a cell and smoke the entire unit

i do know capacitors can short .

case in point i was preparing an old window air conditioner to recycle for the metals and powered it up to see what was wrong if anything and the compressor would turn over and then trip the overload protection.

it turned out that the herm side shorted so the compressor would not start.

in the end i had to disconnect the start terminal and use screw driver to short the start and run terminals briefly to start it manually.

things seems to be quiet here could it be http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 changing the direction on the up coming bms 4.0?

(standard obligatory question) not to sound impatient but it has been a couple weeks since it was said that the bms 4.0 would be coming out

oh well maybe a blessing in disguise with the advent of http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=18392 if it works out better then i will go towards that.

Err, uh, I guess that one is going to be Ver.5.

Based on previous experience, there is a good chance that "Ver.5" will have some fatal flaw that prevents it from becoming practical. Even without fatal flaws, it will take a long time to validate the design and get it to a stage where it can be sold. Meanwhile Ver.4 has a lot of work put into it and should work well until something better comes out.

 

[bandaro]

oooh, this is a more specific thread for my question:

when charging with a bms, what does a charger like this: http://www.e-mtb.com.au/36v-5a-lifepo4-charger-w-alloy-case-p-239.html do? its a headway brand, apart from that i know nothing about it...

does it just give out X volts and Y amps, and the bms does the balancing and cc/cv phase? so basically, is the charger a simple transformer-converter.

 

[EVan]

This sounds like exactly what I'm looking for, but to save me reading 124 pages: is some version of it available and in use?

 

[eRev]

Hi all,

I'm having issues with my V2.6 BMS circuit when charging. I have (22) SkyEnergy 60Ah cells and lab grade power supply able to provide 10-amps of charge current. Terminal voltage set to 3.7V x 22. Cell monitoring leads are rather long at 32-35" and bundled together - charge leads are suitably sized and not very long, say 10AWG / 3ft.

As one would expect, lower capacity cells will reach the HVC causing the BMS to kick into throttle mode. Once throttling is stable at 500mA (or there-abouts) with only one or two cells at HVC (3.65V) I start getting several other cell HVC cutoff LEDS illuminating, but the corresponding cells are stil at 3.4V or there-abouts.

I suspect noise / ocillations as talked about in this thread is causing interference with these select HVC circuits. I need a recommendation for a fix - a bandaid fix if you will. My EV motorcycle layout is highly constrained, so shortening the cell monitoring leads is not an option.

To date, I haven't allowed the BMS to completely cycle (balance the pack) for fear that such a condition as mentioned above will cause damge to the BMS.

I appreciate any help moving forward.

/Steve

 

[fechter]

Yep, that could happen with those long wires. What gauge are they?

I would suggest installing 1uf MLCC capacitors across each cell circuit, located close to the LM431s. No guarantee that will stop the oscillations but it will help.

You might get away with any old capacitors across each cell circuit located on the board, like attached to the legs of the power resistors. For electrolytics, use at least 10uf.

About the only other thing to try is use heavier wire for the cell taps to reduce the resistance.

 

[eRev]

Thanks fetcher for the prompt reply.

Cell monitoring wires are 16AWG with a 5Amp fuse at the cell terminal.

I will try the recommended 1uF MLCC capacitor fix as suggested and report back.

Thanks again.

/Steve

 

[kadir]

Anticipating the Arrival of the V4 boards

 

[GGoodrum]

Still working on layout changes, as I'm trying a new variant that replaces the LM431 voltage detector with another TC54, which will drive a logic level power FET in the digital mode directly, to turn on the shunts. The small amount of hysteresis that the TC54 has should eliminate the oscillation problem altogether. I'm testing this today, and have already done the layout changes. I just need to see how it affects the throttling, but I think it will be fine, based on some previous tests.

I'm also going to try this 6s test board as a standalone balancer, which will mimic the Battery Medic/booster combination. In this case, I will charge the cells to about 4.16V per cell, and have the balancer pull all the cells down to about 4.12V.

-- Gary

 

[ejonesss]

good idea!!! by having 2 tc54's it will furtherly reduce the costs of the bms because we order double of a given part.

 

[steveo]

Hey Everyone,

I'm just gona jump in on this thread in ask .. I havn't following the 125 page bible in a while ...

All I want to know is will gary be selling the 24cell bms pcbs again; i noticed tppack no longer has them? any reason why he doesn't have them in stock any more?...

I'm looking for a bms system to manage 25ah 32s a123 battery pack!

thanks
-steveo

 

[ejonesss]

i think he has stopped selling the bms boards because he is developing the next version.

Hey Everyone,

I'm just gona jump in on this thread in ask .. I havn't following the 125 page bible in a while ...

All I want to know is will gary be selling the 24cell bms pcbs again; i noticed tppack no longer has them? any reason why he doesn't have them in stock any more?...

I'm looking for a bms system to manage 25ah 32s a123 battery pack!

thanks
-steveo

 

[steveo]

i think he has stopped selling the bms boards because he is developing the next version.

Hey Everyone,

I'm just gona jump in on this thread in ask .. I havn't following the 125 page bible in a while ...

All I want to know is will gary be selling the 24cell bms pcbs again; i noticed tppack no longer has them? any reason why he doesn't have them in stock any more?...

I'm looking for a bms system to manage 25ah 32s a123 battery pack!

thanks
-steveo

Wohooo ... awsomeness!!!!

Gary please fill me in; Roughly .. when will you release them ... will we be able to put heat sinks on the discharge reisistors??? I think this is a definate cooling upgrade i recommend !!!!!

-steveo

 

[GGoodrum]

Yes, there will be a "replacement" for the v2.x series. We have been sorting through solutions to eliminate the oscillations that plagued our previous v4.x test variants. These were not really an issue with the v2.x boards, because with those the shunt currents were only 500mA, or so. The oscillations become more of a problem when the shunt currents go up significantly, to about 1A, or so. the solution is to use the shunts in the digital realm, with a bit of hysteresis. In essence we are "embracing" the oscillations, but in a controlled manner now. Anyway, we are testing replacements for the LM431 voltage detectors, and Richard is going blind, reading detector datasheets, but we are pretty close.

In any case, to answer your question Stevo, yes, there will be something available soon, and going up to 32s, or higher, will not be a problem.

As for the cooling, what I've found is that with the newer v4.x-style boards, which have pads on both sides of the boards and a bunch of vias between layers, it helps quite a bit. Also, using two 3W resistors, instead of one 5W, allows more surface area contact. The boards now also have automatic controls for fans. Active cooling is pretty much required, for higher shunt levels.

-- Gary

 

[eRev]

Well, I just let the smoke out of my V2.6 BMS - suspect carelessness on my part while soldering or was that unsoldering.

When the 1uF caps were added to suppress the shunt ckt oscillations, the HVC circuits seemed to behave themselves then after maybe 5-6 cells hit HVC a group of 7 or 8 would suddenly light-up too and latched-on even though the cells were below HVC. Disconnecting the charger and battery pack didn't unlatch the ckts. I futz with this for a while and found a couple cold solder connections on the caps just added - then about 10 seconds after I reconnected...

So, Gary - per my attempt to contact you thru TPpacks - can you make your current LVC/HVC pcb a 24-channel variant w/o cutting the boards up please? I'm really digging the effort you and Richard are putting into this project and I can't let what happened to my V2.6 BMS be a show stopper - I want to get away from shunt based balancing all-together. The HVC circuit will stop my bulk charger when the weak sister raises her hand and from there I will single cell balance w/ a 3.60V source, monitor only charge current and digitally sequence. /Steve

 

[GGoodrum]

The short answer is yes. The LVC/HVC boards right now are connected, six to a board. The v4.x boards will be connected as well.

Richard and I are zeroing in on a new front-end that switches the shunts on and off digitally. In addition to the already listed advantages, it also means that almost all of the heat generated will be from the shunt resistors now, so smaller parts can be used, freeing up some real estate.

The board sections will be connected, but they can also be cut apart and stacked, or put in an extruded aluminum box.

Sorry about the email, but I get too busy sometimes. Eventually, I do get to most, though, but lately I've been swamped.

-- Gary

 

[eRev]

No problem Gary. As mentioned many times in this thread - we appreciate the hard work that goes into making affordable BMS a reality. Kudos sir. /Steve

 

[steveo]

The short answer is yes. The LVC/HVC boards right now are connected, six to a board. The v4.x boards will be connected as well.

Richard and I are zeroing in on a new front-end that switches the shunts on and off digitally. In addition to the already listed advantages, it also means that almost all of the heat generated will be from the shunt resistors now, so smaller parts can be used, freeing up some real estate.

The board sections will be connected, but they can also be cut apart and stacked, or put in an extruded aluminum box.

Sorry about the email, but I get too busy sometimes. Eventually, I do get to most, though, but lately I've been swamped.

-- Gary

one more question for you gary ....

Will the newer boards be more robus in the case where if one of the cell channels gets unplugged will it blow out the channel?

this issue was very discouraging with the older boards..

-steveo

 

[GGoodrum]

one more question for you gary ....

Will the newer boards be more robus in the case where if one of the cell channels gets unplugged will it blow out the channel?

this issue was very discouraging with the older boards..

-steveo

Yes, the new circuits are more tolerant, but I don't believe unplugging a channel will kill it in the old design. Reversing the wires on a channel would definitely do it, though.

The best way to prevent mis-wires is to use solid connection locking plugs, between the pack and the BMS boards. That way you check that the plugs are wired properly first, before plugging them together.

-- Gary

 

[steveo]

one more question for you gary ....

Will the newer boards be more robus in the case where if one of the cell channels gets unplugged will it blow out the channel?

this issue was very discouraging with the older boards..

-steveo

Yes, the new circuits are more tolerant, but I don't believe unplugging a channel will kill it in the old design. Reversing the wires on a channel would definitely do it, though.

The best way to prevent mis-wires is to use solid connection locking plugs, between the pack and the BMS boards. That way you check that the plugs are wired properly first, before plugging them together.

-- Gary

Hey Gary

I've blow the bms i had several times if one of the channel wires gets unplugged or breaks while other wires are connected...

i have 3 plugs on a 24s board... if i unplug the channels in the incorrect order i've blown out a channel..

hard to explain ...

but just want to make sure you are aware

-steveo

 

[GGoodrum]

Hey Gary

I've blow the bms i had several times if one of the channel wires gets unplugged or breaks while other wires are connected...

i have 3 plugs on a 24s board... if i unplug the channels in the incorrect order i've blown out a channel..

hard to explain ...

but just want to make sure you are aware

-steveo

I'm pretty sure the new design can't fry anything if a wire becomes inadvertently unplugged, but I'll make sure it can't. Also, to cure the problem of having to plug the sections in a particular order, you can fix that by making each section independent, with its own connector. It would require one extra wire per section, and with the v2.x boards the sections would need to have the connections between the 4th cell in one section and the 1st cell in the 2nd section, broken. This would allow, however, any 4-channel section to be connected to any 4s block of cells in the pack.

The new layouts will have mini-terminal blocks for each section, which will either be 4-channels, 6-channels or 8-channels. I plan on doing at least two section variants, but I'm no sure if it will be 4 and 6 or 6 and 8. I'm leaning towards 6 and 8, to cover the most configurations, like two 6s boards for 12s, two 8s boards for 16s, three 6s boards for 18s and either three 8s or four 6s for 24s. Anyway, I'm using Molex Microfit 3.0mm plugs and connectors, with 18-gauge wires, and I like these a lot. They make a very solid connection and can carry up to about 3A per pin. They are also fairly small and compact, compared to the larger Amp/Tyco VAL-U-LOK 4.2mm PE series I've used in the past. I use the 7-pin version for 6s boards, and the 9-pin version for 8s.

-- Gary

 

[kadir]

I use 15 Pin Sub-D connectors. Among the 15 pins I only use 12 of them for each 12 Cells. This gives me an option of not replacing the whole connector if one connection pin is broken. These are slightly different connectors. Normally in the D-connectors for serial, the wires are exposed and normally soldered on. The ones that I used, have pins crimped to the wires, and then the pins lock into the housing, thus hiding the exposed wires.

I dont like molex because, sometimes, if the the male pin is at an angle and doesn't slide onto the female, the pin might push each other out from the other end. I have done that twice. And killed one cell, while damaging the other one.

Here
http://fayazkadir.com/blog/uncategorized/bms-wire/
is a slide of the picture of the connectors.

The wires can easily Handle 3Amps, they are rated for 2A. But I am sure, there is atleast one Amp of safety margin.
I have tested them at 3 Amps. I run a seperately excited DC motor, with the Stator current at a continuous 3 Amps. And its still running as we speak.

Or if someone wants a more solid connector, Ebay can provide large 32 pin Cannon conectors( ~2 cm in diameter). They are military grade connectors with pins capable of handling 5Amps (atleast).