Linear Tech 12-cell battery monitor/BMS

[methods]

So there are no output isolation FETs then?

Are you asking if I have lugs on my board to intercept the main power which very well could be 150V 750A running on 000 gauge wire (about an inch thick)?
The answer to that my friend would be an emphatic - no

Only small time BMS's go inline with the power. If I were to control the main power it would be with a single fet that can break the 12V signal that drives the 2000A contactor.

Hey - thanks for the idea

-methods

 

[amberwolf]

Although I'd use a regular bipolar power transistor for that, in open-collector mode as a pulldown for the "ground" end of the contactor coil. If it fries for any reason it's a lot less likely to fry shorted than a FET, and is probably less sensitive to RF noise spikes induced in the coil than a FET would be, requiring less external stuff to protect it.

EDIT (ADDED): Also, an open-collector output can be used in several ways, meaning when you are not directly driving the contactor but instead telling the controller to do it (like the Soliton1 with integrated precharge and contactor).

 

[grindz145]

haha yeah, I know... But a contactor close to the pack makes alot of sense. You can have a hell of a meltdown before a 1000A fuse goes...

EDIT: P.S. Believe it or not I've seen it done with that much current before....

 

[methods]

EDIT: P.S. Believe it or not I've seen it done with that much current before....

I was just teasing you. I actually had not really thought about active control much yet. I am just focused on taking accurate measurements, managing balance heat, and ensuring reliability / safety. Who knows - maybe the board will end up with 1" lugs on it

-methods

 

[liveforphysics]

Two thumbs up for being one of the few who is willing to roll like this. This is your finest work yet.

 

[grindz145]

Anyone read anything about the TI bq77PL15x of the bq77PL910? It looks like these are scheduled to be released 1st quarter. they offer fet drive (low and high side) and a balancing circuit on the PL910. They can both operate in stand-alone mode. I'm pretty excited... 18s stackable and 10s respectively.

 

[reagle]

It all depends on where their voltage limits are settable.
In the past you could not get PL900 to go low enough for LiFePo4. Maybe this time they got it

Anyone read anything about the TI bq77PL15x of the bq77PL910? It looks like these are scheduled to be released 1st quarter. they offer fet drive (low and high side) and a balancing circuit on the PL910. They can both operate in stand-alone mode. I'm pretty excited... 18s stackable and 10s respectively.

 

[grindz145]

It all depends on where their voltage limits are settable.
In the past you could not get PL900 to go low enough for LiFePo4. Maybe this time they got it

Anyone read anything about the TI bq77PL15x of the bq77PL910? It looks like these are scheduled to be released 1st quarter. they offer fet drive (low and high side) and a balancing circuit on the PL910. They can both operate in stand-alone mode. I'm pretty excited... 18s stackable and 10s respectively.

According to the datasheets they will both be able to do LiFePO4. They claim "any chemistry" which should probably be any lithium chemistry and even then its pushing it...but the UV and OV settable limits look like they should work just fine with LiFePO4 (WooHoo!). Of course we never know for sure until we actually get to evaluate it, but I'm cautiously optimisitic

 

[reagle]

bq77PL910 looks interesting for standalone tool/bike packs up to 10 cells.
Ability to set protection limits and leave it be is nice and requires no processor/firmware. And they did indeed make the ranges wide enough to handle pretty much anything out there. Now, when can we get these...

According to the datasheets they will both be able to do LiFePO4. They claim "any chemistry" which should probably be any lithium chemistry and even then its pushing it...but the UV and OV settable limits look like they should work just fine with LiFePO4 (WooHoo!). Of course we never know for sure until we actually get to evaluate it, but I'm cautiously optimisitic

 

[methods]

The LTC can do all of those things.
It has a very low power stand alone monitoring mode with output.
12 channel.

Any off the shelf RC equipment can do monitoring for $20 tho right?
For monitoring it does not matter if you stack 6S units - only when you want to balance does it become important to have them all networked.

btw: got my board working 100% and shipped it to the customer.
Looking forward to see how it works out.

V2.0 starts tonight

* 1A low cost surface mount shunts spread over a large area
* Charge throttling based on cell temperature, shunt temperature, and cell voltage

I am setting it up to run 36S @ 20A charge
That is more than you can get out of a wall so. . .

-methods

 

[grindz145]

Time to install a 25A receptacle... dedicated bike line

 

[reagle]

Looks like EV Components is now carrying Manzanita Micro's version of LTC6802 boards. They look pretty nicely designed:
http://evcomponents.com/cscart/index.php?dispatch=products.view&product_id=29878
An LTC chip plus micro plus 4/8 load resistors and optoisolated link.
Damn, those guys are quick- i am still getting around to designing mine..
Hmm, on a second thought and being lazy bastard I am, I could just use one as a hardware platform and make whatever firmware I want- the micro looks to be Atmel AVR I think

 

[methods]

At that price I could ask $1080 for my 36ch version :x
Maybe I am in the wrong business

Looks like he is only running it in supervisory mode. Seems like he would want to utilize the full functionality.
I like a few of the ideas I see there on his board - I will steal them

We finished the balancing section. It is laid out but needs to be fabbed. We went with a second board that connects via a header. Both boards slide into an extruded aluminum housing that is fan cooled. 1.2A shunt currents, 36 channels. 4 x IRFB4115 fets that can handle 150V 50A charge current. LVC, HVC, CHRGKILL, and AUX optical outputs.

I will post up soon.

-methods

 

[grindz145]

That makes alot of sense to limit input current with Fets, even though the output will be way too high.

You still using that arduino for control of have you integrated something else?

 

[methods]

Since I wanted to use my old code and I dont have any boot-strapped chips I am still using the nano boards.
Eventually I will drop something else down - but for now - the Arduino is amazingly quick and easy for prototyping.

-methods

 

[methods]

Added features:

* RF Digital isolation (instead of opto)
* Isolated DC-DC draws power from USB, does not center tap pack
* Opto-outs for LVC, HVC, Charger Shutdown, AUX
* 1.2A programmable balance currents
* LED shunt indicators
* 3 fans
* Extruded aluminum housing
* 4 x IRFB 4115 150V fets for throttling up to 50A charge current
* Boards stack via header connections then slide into box
* Fans mounted on lid - slides on and off
* isolated grounds: USB ground, Battery Ground, Charger Ground (switched in via 4115)

Features:

* Programmable charge throttling
* Programmable set points for everything (LVC, HVC, shunting, etc)
* Programmable behavior
* Ultra low battery draw
* Cool housing (housing stays only 10C over ambient at 125W balance power)
* Full GUI interface -> Datalogging, settings, etc.
* WDT enabled -> cuts balance and charge in case of fault

View attachment _019_Charger.jpg
View attachment _020_Charger.jpg
View attachment _023_Charger.jpg
View attachment 7
View attachment _002_Charger_2.jpg
View attachment _003_Charger.jpg
View attachment _003_Charger_2.jpg
View attachment 3
View attachment _014_Charger.jpg
View attachment _015_Charger.jpg


V3.0 will be modular in 12S chunks aimed more toward the ebike crowd but still able to service the full size EV market.

-methods

 

[liveforphysics]

Omg bro!!

I'm so proud of you!

 

[steveo]

Me like bms porn

will this bms be capable of any chemisty? adjustable LVC & HVC

good work methods .. cooling was my main issue when i begain using bms ... good to see you solved this problem!

-steveo

 

[methods]

will this bms be capable of any chemisty?

Absolutely.
Every aspect is programmable in every way.
Similar to how we program the infineon boxes but without the bullchit and driver issues. Only quality software :wink:

-methods

 

[mwkeefer]

Methods,

2Questions mate.

Will these auto cell detect and ignore cells which aren't present - for instance if I run a 15-20s I would need 2 units but couldn't populate but a few channels of the second unit?

When will you have them available for sale? I and the rest of us are quite interested!!! (personally I'm dying here for accurate and high current balancing - and it will be nice to be able to use those 3 SP32024s I got from ya a while back !!! LOL

Great work and excellent GUI!

-Mike

 

[liveforphysics]

That "max cell delta" gauge you've got in the middle there is so nice. That is like the most critical thing to know on a pack.

If I were doing a dashboard for an E-bike, I would just want two things displayed. Max cell delta, and average cell SOC. Maybe a pack peak-temp sensor reading would be nice as well, but not all the time, only to appear on the display if it crosses over a certian user-defined threshold.

 

[Hyena]

very very cool methy! I can't wait until you've got boards ready to ship out

 

[methods]

Actually that GUI was just a quick-n-dirty to display some of the data for testing.
The "real" GUI has tabs with different levels of data showing on each screen. My wife calls them "geek mode" and "retard mode".

Methods,

2Questions mate.

ok

Will these auto cell detect and ignore cells which aren't present - for instance if I run a 15-20s I would need 2 units but couldn't populate but a few channels of the second unit?

Yea, that is trivial. An open wire test is built into the LTC. There is a minimum cell count though - I believe it is 4 cells per chip, so if you want to run less than 12S you have to split it up in such a way that there are at least 4 cells on each chip - so 4S minimum, 12S max. You short out the unused channels.

When will you have them available for sale? I and the rest of us are quite interested!!! (personally I'm dying here for accurate and high current balancing - and it will be nice to be able to use those 3 SP32024s I got from ya a while back !!! LOL

Great work and excellent GUI!

-Mike

Not any time soon. These boards are designed for a very specific application and that will keep me tied up for a long time.
When I get things worked out I will then (on my own time) come up with a V3.0 that is modular, cheap, and suited for Ebikes.

-methods

 

[qli029]

well, this is a good forum, learning a lots

 

[CamLight]

V3.0 will be modular in 12S chunks aimed more toward the ebike crowd but still able to service the full size EV market.

-methods

Nicely done methods!