BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries • Hacker News Board

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#131

Post by clouddrover » Fri, Oct 22, 2021, 2:52 AM UTC

gremloni wrote: post #47

anonymousiam wrote: post #17

Tesla said that they will use LFP batteries from now on in all "standard range" cars. https://www.cnbc.com/2021/10/20/tesla-switching-to-lfp-batte... As I understand it, the "standard range" cars are no different from the long range cars, other than a software flag that allows a greater battery discharge. So this will complicate the Tesla fast battery swap stations if they ever actually build any of those. https://ww…

Battery swap stations are never going to happen. They make up too much of the cost of the vehicle. Maybe in government public transport fleets or something.

Battery swap stations have happened.

NIO has done over 4 million battery swaps: https://www.carscoops.com/2021/09/nio-customers-have-perform...

NIO is expanding their battery swap stations to Europe: https://www.carscoops.com/2021/10/nio-brings-battery-swappin...

Gogoro has done over 200 million battery swaps: https://www.gogoro.com/news/400k-gogoro-network-subscribers/

Gogoro is expanding their battery swap network: https://thedriven.io/2021/10/12/taiwans-gogoro-launches-batt...

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#132

Post by sudosysgen » Fri, Oct 22, 2021, 2:54 AM UTC

jacquesm wrote: post #126

sudosysgen wrote: post #121

Earlier quoted context omitted.

Hey, I've done something similar with a Lenovo battery. If you're willing to scrap a battery pack, it's definitely doable by reusing the electronics. You should add your own BMS though. It's fine to run them in parallel for the balance wires, but connect the phase wires to your own BMS. It should work :) I recommend Daly BMS's, they're not too expensive and they kept my 1.5kWh pack in balance and protected it from a…

That's exactly the plan, with one very small modification, I'll be adding a 1A balancer that gets triggered on the 'charge' signal, and I'll replace the balance wires to the original BMS with a laddernetwork of resistors to fool the old BMS into thinking it is still seeing a nicely balanced pack. This is the balancer (not a BMS, just an active balancer): https://www.aliexpress.com/item/32945565819.html What's neat ab…

Your Aliexpress link sadly doesn't work :/

It's a good idea to take care to balance the cells. That said, 1A of balance current is very overkill.

I actually ran tests as I was curious. I cycled my battery 10 times at 1C. The parallels only went 0.03v out of balance, over 10 hours of discharge and 30 hours of charge. That's at the top end, so really I had what, 0.5Ah of imbalance, over 40 hours of non-stop use in the very worst scenrio. So the 40mA of balance current from my BMS is actually enough even in the worst situation.

In reality you'll be doing at most one battery cycle a day, maximum two, so you'll actually be needing only 4 mA of balance current!

Because of that I decided that the 40mA was all that I needed and I didn't add any really heavy duty balancing circuit. The one that came with my BMS was more than enough, even though like you I was quite skeptical at first.

So I think you should keep it simple. If I were you I'd solder the balance wires of both BMS's as parallel and go on with my day knowing that whatever I'll throw at it, it will stay balanced. Plus there is less that can go wrong :). But then again, I'm not the one not to enjoy some overengineering

The thing is that when you have as many cells in parallel as we do, for you probably 15?, the imperfections between each cell really do balance out and even with the pretty knackered cells I went with it will stay in balance.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#133

Post by sudosysgen » Fri, Oct 22, 2021, 2:58 AM UTC

jacquesm wrote: post #128

sudosysgen wrote: post #124

Earlier quoted context omitted.

Hmm, it's better than others, but I can still find Li-ion for a better price. Besides, I'm outside of the EU, I can't imagine the shipping costs haha

Ah yes, that complicates matters considerably. And true, Li-Ion is still a better price (and higher density to boot). But it's worth monitoring this because at some point there will be a crossover.

Indeed there will be. In North America, batteryhookup.com has A123, Headway and CATL cells at a pretty okay price. The CATL cells still aren't worth it IMO, but the Headway LiFePO4 cells can push a ridiculous amount of current, something like 200A from a single cell.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#134

Post by sudosysgen » Fri, Oct 22, 2021, 2:59 AM UTC

LgWoodenBadger wrote: post #95

danans wrote: post #8

Earlier quoted context omitted.

Adding to this info - LiFePO4 batteries are far more chemically stable and therefore need no cooling system. I have LiFePO4 home backup batteries and they are completely passively cooled.

Weren’t LiFePO4 the batteries that need to be charged in fireproof bags because they have a tendency to spontaneously combust? Those were the horror stories I read about back when I had some for my Walkera rc helicopter.

You're thinking of Li-Po, not LiFePO4.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#135

Post by jacquesm » Fri, Oct 22, 2021, 3:03 AM UTC

sudosysgen wrote: post #130

jacquesm wrote: post #122

Earlier quoted context omitted.

That's one very mean bike. I think I have enough steel in one leg to keep me off rides like that but consider me impressed that you are taking that into traffic. Is this street legal where you live? You'll be doing wheelies if you are not careful :) How much does the whole thing weigh? Mine is at 29 Kg, and that's with only a 500Wh pack, I figure about 6 Kg/KWh so by the time I'm done building the new pack it will be…

Well, any bike will do wheelies if you have the knack for it :) But if you're wondering, no it doesn't do power wheelies - I can definitely feel the weight shifting back though. I haven't weighed it yet, but: The bike itself is 15kg. The battery weights 48g per cell, so at 140 cells (14s10p) that would be 6.7kg, but I'll round up to 7.5kg for the wiring, mounting, etc... The motor and controller and screen are around…

Very neat, and nice to have a fairly high voltage setup like that, mine is only 36V nominal so it needs pretty fat wiring to not lose a bunch of power on the way to the motor (that was the first thing I fixed, the difference was noticable, both in range and in power).

35 Kg is doable, as long as that battery doesn't run dry!

Mine really isn't pleasant if you end up without assist, it's doable but I really hate it, especially uphill. Can't wait until I have that pack ready, but I'm not going to hurry it, I'd hate for it to fail halfway or so, and that's a lot of tinkering to be done still (and a lot of welds...). But the batteries are already here, the balancer as well, I have a BMS from a scrapped pack and a good idea of what the housing should look like, if my test pack works I'm off to the races, then it is mostly straight up assembly and otherwise I'll need to re-think the way the BMS is hooked up.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#136

Post by elihu » Fri, Oct 22, 2021, 3:10 AM UTC

sudosysgen wrote: post #61

elihu wrote: post #44

Earlier quoted context omitted.

> This is somewhat mitigated by the fact that LiFePO4 handles 100% charges and 100% discharge I don't think that's quite true. LFP cells have a minimum voltage; they may survive being discharged to zero, but it probably isn't a good idea for long-term longevity. (I'm currently working on an EV conversion that uses LFP cells, but I haven't actually done durability testing or anything like that, I'm just going by what…

That's not quite what 100% discharge usually refers. 100% discharge means from the maximum charge voltage to the minimum discharge voltage. That corresponds to, normally, using up the rated capacity of the cell. On lithium batteries you'd normally only use 80% of the rated capacity if you want to maintain durability.

Ah, that makes sense. I guess it's kind of arbitrary (I mean, you could just call 80% full "100%" if that's all you'd use in practice) but it does seem quite useful if the "full" voltage is generally agreed to be in a narrow range and not have a wide range of maximum values depending on whether you care more about absolute capacity or longevity.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#137

Post by jacquesm » Fri, Oct 22, 2021, 3:12 AM UTC

sudosysgen wrote: post #132

jacquesm wrote: post #126

Earlier quoted context omitted.

That's exactly the plan, with one very small modification, I'll be adding a 1A balancer that gets triggered on the 'charge' signal, and I'll replace the balance wires to the original BMS with a laddernetwork of resistors to fool the old BMS into thinking it is still seeing a nicely balanced pack. This is the balancer (not a BMS, just an active balancer): https://www.aliexpress.com/item/32945565819.html What's neat ab…

Your Aliexpress link sadly doesn't work :/ It's a good idea to take care to balance the cells. That said, 1A of balance current is very overkill . I actually ran tests as I was curious. I cycled my battery 10 times at 1C. The parallels only went 0.03v out of balance, over 10 hours of discharge and 30 hours of charge. That's at the top end, so really I had what, 0.5Ah of imbalance, over 40 hours of non-stop use in the…

Link is now fixed.

Yes, 1A is overkill, 100 mA would be plenty (and likely still too much), but the Bosch BMS has maybe 1 mA balancing capability and that's too little for a pack that size so I decided to throw in something beefier and this was the smallest one that had bluetooth out so I can monitor it, a pack that gets unbalanced too much is a good indication that something is wrong, for instance a weld that's not done well, or a cell that is on the blink. Those things are best caught early with homebrew packs of this size, you really don't want to have a battery fire.

A good, new pack stays within a few mV over a whole charge/discharge cycle. The reason is that if you start off with a balanced pack the laggards during discharge will be the first ones to be topped off so it all balances out in the end modulo the difference in internal resistance, and this leads to the various cells dissipating a bit more or less heat (as well as the cells on the inside of the pack being better isolated). That's cumulative and over a longer number of cycles you'll see the pack drift more and more out of balance.

So it's not so much a matter of me being skeptical as it is knowing the exact limits of the Bosch BMS and the fact that if the pack ever goes out of balance more than that that it will brick the BMS, and that's a nuisance, coupled with the fact that this particular balancer had some more useful features.

16 cells in parallel, and you're right that the cells tend to self balance to a large extent. But it's also a really good way to monitor the pack so it's dual duty for me, wear monitoring (which has safety implications) and balancing all in one for less than $100, which on the total cost of this project is quite doable. The cells I'm using are $6 a piece (Samsung E35s, 3450 mAh) and 160 of them is already quite the outlay.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#138

Post by jacquesm » Fri, Oct 22, 2021, 3:17 AM UTC

sudosysgen wrote: post #133

jacquesm wrote: post #128

Earlier quoted context omitted.

Ah yes, that complicates matters considerably. And true, Li-Ion is still a better price (and higher density to boot). But it's worth monitoring this because at some point there will be a crossover.

Indeed there will be. In North America, batteryhookup.com has A123, Headway and CATL cells at a pretty okay price. The CATL cells still aren't worth it IMO, but the Headway LiFePO4 cells can push a ridiculous amount of current, something like 200A from a single cell.

Yes, they're crazy. Batteryhookup is an interesting company, they're very well known in the homebrew powerwall scene.

Here's to hoping that the next battery revolution will come in on time and that we'll see some real economies of scale here.

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#139

Post by bsder » Fri, Oct 22, 2021, 3:20 AM UTC

CamperBob2 wrote: post #65

bsder wrote: post #60

Earlier quoted context omitted.

> There was hysteria in the mainstream media that Japan was going to overtake the US to become the biggest economy in the world. People who see the world in zero-sum terms made idiots of themselves.* To be fair, the Japanese, through MITI subsidies, slaughtered the US semiconductor industry through the early 80s. It took a lot of US government funding through VHSIC and the VLSI Project to prevent semiconductors from…

It took a lot of US government funding through VHSIC and the VLSI Project to prevent semiconductors from collapsing completely. What are some specific examples? I don't remember anything like that happening. Japan had Hitachi and Matsushita and Toshiba, but we had Motorola and Intel and NatSemi and TI and countless other dominant players. There was plenty of room in the market for both.

Please remember that the only thing which had any real volume in that time (1985 and prior) period was dynamic RAM. DRAM drove semiconductor process development until microprocessors eventually took over as defining the process capabilities (about 1990-ish).

The US DRAM manufacturers got absolutely destroyed by the Japanese ones. So much so that getting out of DRAM was an existential crisis moment for Intel.

One big point of the crux was scan lithography vs step and repeat lithography. You needed step and repeat lithography to get to the next level, but it was way more complicated than scan lithography. The US companies made the jump--the Japanese ones didn't. The Japanese companies clobbered the US companies in terms of profitability and production.

This is part of the origin of "never be first" in the semiconductor industry.

The US government then stepped in and pumped a massive amount of money behind semiconductor tech to stabilize that. Of course, once step and repeat (and some other related technologies) went online, the Japanese DRAM companies now got slaughtered and basically driven out of business.

Things don't always "just work out" and sometimes the game really does create "winners and losers".

Re: BYD has reportedly received 10 GWh of orders from Tesla for LFP batteries

#140

Post by sudosysgen » Fri, Oct 22, 2021, 3:31 AM UTC

jacquesm wrote: post #137

sudosysgen wrote: post #132

Earlier quoted context omitted.

Your Aliexpress link sadly doesn't work :/ It's a good idea to take care to balance the cells. That said, 1A of balance current is very overkill . I actually ran tests as I was curious. I cycled my battery 10 times at 1C. The parallels only went 0.03v out of balance, over 10 hours of discharge and 30 hours of charge. That's at the top end, so really I had what, 0.5Ah of imbalance, over 40 hours of non-stop use in the…

Link is now fixed. Yes, 1A is overkill, 100 mA would be plenty (and likely still too much), but the Bosch BMS has maybe 1 mA balancing capability and that's too little for a pack that size so I decided to throw in something beefier and this was the smallest one that had bluetooth out so I can monitor it, a pack that gets unbalanced too much is a good indication that something is wrong, for instance a weld that's not…

Ah, if the cells going out of balance even slightly will brick the BMS, that is something else. In that case "lying" to the BMS is a good idea and at that point you might as well add your own balancer.

AS for the battery fire, that is why I suggest using a second BMS and running the phase wires from there - a good BMS such as the one I use will cut off the power when there is an even slight imbalance, which is my cue to investigate it and avoid a battery fire :) - while that second BMS would handle the balancing.

But paying for 6$ cells, my oh my, that does make it worth it :) I paid all in all 70 cents a cell and even then I was quite skittish :P