[asymptonic]

At any rate, if I were a manufacturer, you just have a slot on your assembly line attaching the BMS slave modules to batteries out of a bucket of preconfigured ones with the node number on them, slap a sticker with the module #, and send it down to be dropped into the pack. This would save oodles of time.

Edit: scratch that, just label the preconfigured bms slaves and attach them to a pack with the modules already in them. Still...

Now they just need to invent wireless bus bars.

 

[joekitch]

Did a lil chat with Ti support. Looks like it's coming soon (tm)

 

[reiderM]

Presumably yes. Although more sophisticated systems are possible, it appears that BMS balancing is normally based on discharging individual cells through bypass resistors, so there's no need for current to flow between the cell and a central BMS device. It is common for modules to be equipped with "slave" BMS units which handle this themselves, only using their connections to the BMS master for coordinating communication; that makes a wireless network a straightforward (although unnecessary) variation on the usual wired network between BMS master and slaves.

My thoughts exactly. What do you gain from wireless? There's really not a lot of wires in a BMS master-slave network, not sure what GM is on about. This seems to be more of a "use big technological words to impress customer" type video.

 

[asymptonic]

My thoughts exactly. What do you gain from wireless? There's really not a lot of wires in a BMS master-slave network, not sure what GM is on about. This seems to be more of a "use big technological words to impress customer" type video.

No I think you missed the point entirely. There is no need for current flow between the cell and a central BMS. You only need the coordinating BMS master to tell something at the cells to discharge, and read the voltage at the cell itself. How that communication occurs (as long as it's reliable) is irrelevant. So what you gain from wireless is a gain of some amount of time saved assembling the battery if you're making 20 thousand vehicles, and X modules * Y cars * Z length of copper wiring which ain't nothing for a company which for which cost optimizing is a way of life. For the DIYers it may be a couple dozen bucks, but multiplied times all the vehicles and multiplied times the other areas they are engineering to save a few adds up real quick.

For us DIYers it'd be pure convenience in wiring, again, if reliable and not complicated for us to interface to. Personally, given my experience in the rest of my career, I'll trade money for wires but that's not to say this is a bad idea. I think it's a great one for the industry.

 

[brian_]

It's interesting that the company that builds dozens of different models with a huge number of variations of each of them - often in equipment which communicates with the rest of the vehicle - hasn't resorted to wireless for anything else. Perhaps the issue in general is interference (that would certainly concern me), and having all of the BMS components inside the same metal box presented an opportunity.

 

[reiderM]

No, we are on the same page. I'm familiar with how it works in most OEM vehicles currently. Obviously wireless cell taps would be very much impossible (wireless electricity)! I just don't think there would really be any cost savings. 22awg wires for CAN communication with balance boards is a cheap and simple solution. Wireless is not as simple, and I can't imagine it's as cheap (although I'm not sure).

I need to learn more about it though for sure.

 

[asymptonic]

No, we are on the same page. I'm familiar with how it works in most OEM vehicles currently. Obviously wireless cell taps would be very much impossible (wireless electricity)! I just don't think there would really be any cost savings. 22awg wires for CAN communication with balance boards is a cheap and simple solution. Wireless is not as simple, and I can't imagine it's as cheap (although I'm not sure).

I need to learn more about it though for sure.

Most definitely cheaper. 22awg is cheap, sure, but a battery pack is ~4x5 ft? Just guessing, but if you have say and I'm just doing back of the napkin here, 96 cells, then you still have to run 1 22awg wire between each cell tap and a perfectly centered BMS about 32in, with copper 22awg copper at 514 ft/lb means about 2lb of copper, at 4.3$ a lb at the time of this post, so ~$9 of copper before you price in the creation of the small wiring, sheathing, and upcharge for creating the wiring. Then add the cost of connectors per module, tacking on per supplier profits, etc, and again, adding up. Then the connectors on the BMS master and the many many additional traces on the PCB and overall size increase that result.

The components on the PCB, even looking at it, are a handful of surface mount components and a small cheap IC. The wireless bits are cheap too. And no long distance high cost copper wiring. My guess is that even in material cost they're 2x ahead of a wired BMS setup at scale.

But this is all back of the napkin material cost. The real cost is that they're saving some fraction of human time on an assembly line. Each of these connections to each battery module takes time * wage cost, plus perhaps quality control, which I'm guessing is at or above the material cost.

 

[remy_martian]

They may be cheap, but what's cheap when there's no stock and you need a letter from the Pope to get samples and more than 1-1/2 pages of the datasheet, not even showing the g-damned pinout?

And $1200 for an eval board for a $30 part? 😂😂😂 I chewed my guys out for that kind of idiocy for our semiconductor products. Build enough boards to go around AT COST...

This is the f*ckery of designing and building anything these days. No stock, 52 week lead times, and attitude. And the lack of components ripples through the rest of the car's decisions all the way to the battery stack 🤬

I'd swear China is buying up every bit of semiconductor stock to screw the West, but my tinfoil hat is out for powder coat.

/rant_off

 

[joekitch]

No tinfoil hat needed, the chip crunch has well known causes and exacerbations

Also, GM's solution isn't the same as Ti's, I imagine that $1200 is a projected cost of a consumer kit or something wich would roll in dev cost. Then again maybe it ends up way cheaper in the end.

My build was going to use the master/slave BMS setup for aem's VCUs so I'm already not too concerned about material cost or wiring complexity as such, this just seems like the way to go for the future and wanna get in on it

 

[remy_martian]

Yes. I stand corrected. GM is allegedly using ADI. Wireless BMS is going to be locked down, making it a total PITA and will need factory boards replaced.

$1200 is the cost of an engineering evaluation board. These are not intended to be sold to consumers and are sold in very limited quantities.

Nobody can explain the chip shortage other than game console instead of car chips blah blah blah which is BS. Somebody has bought up and is hoarding everything. Every corner is cleaned out.

Watch what happens when the hoarders open the flood gates after the industry has "caught up".

 

[cricketo]

folks discussed balancing resistors in the thread. Are they actually using resistors instead of some form of active balancing ?

 

[joekitch]

Its more like factory shutdowns early covid sent a shockwave through the entire industry, at basically the same time the demand for electronics went through the roof as people stayed at home much more. Those two things combined had numerous downstream effects.

What we're experiencing is a reckoning for the just in time manufacturing model, nobody stockpiles anything, everything is outsourced, makes for brittle supply chains indeed

 

[remy_martian]

^^ Switched...

 

[asymptonic]

reckoning for the just in time manufacturing model

Bingo. JIT only works when everything is running smoothly and demand is predictable. The last couple of years have been the exact opposite.

Edit: it also boggles my mind to think about how many steps there are for any given product, including electronic components. And a fault in any one of them piles up like an icy Florida highway.