Not economic for making too big LiFePO4 batteries

[jaja]

Someone asked me to make big LiFePO4 batteries such as 72V40AH...
I consider it uneconomical because of:
1. It takes too many single cells, cause it too heavy. And it is more subject to damage during using.
2. Lack of experience for the manufacturer to make so big BMS...
Suggest using two small batteries series...compared to one big battery,these two are durable...

Can anybody give me better suggest ? Thanks!!!

 

[Link]

Probably they wanted it for a small motorcycle, because no ebike can handle a pack that large.

The BMS thing, however, does make it easier to put two 36V 40Ah packs in series. Though I do believe that Hi-Power or whatever makes 72V balancing chargers...

 

[dogman dan]

That sise of battery needs to be made from single cells of 10 ah sise. If using 1 ah cells, 40 of em paralleled in each cell would tend to lead to some individual cells dying without the bms even noticing it. Series connecting with 36v bms to get 72v would fry the 36v bms wouldn't it? So you are back to discharging without the bms, pretty risky.

 

[Link]

Dead cells go to 0V, so they'd take the rest of the p group with them. The overall failure rate would be higher than with single 40Ah or four 10Ah cells, but, meh. Not really all that high to begin with.

Fechter's diode trick takes care of the BMS issues.

 

[dogman dan]

I try to read and understand yours, fetchers, ect, posts but often get lost. Doesn't series through the bms involve the second one having 72v running through it?

 

[Link]

It's not the 72V running through it (only a little potential across the FETs) that's the problem so much as what happens when one pack cuts out. Then there's the full 72V of potential against the LVC FETs. The diode trick lets that power bypass the BMS when it shuts off.

 

[fechter]

You could do it either way, that is two 36v packs in series (with protection diodes) or combine enough cells in parallel to get 40ah and use a single 72v BMS. You could also put two 72v, 20ah packs in parallel, but this would require diodes in series with each pack, which would have to dissipate quite a bit of heat (not so good).

 

[dogman dan]

Someday I may understand some of this more fully, but for now I think I understand it enough. I had this mental picture of 36v bms fets with 72v going through them getting hot. I can replace a bad plug or resolder something that worked loose, but that is about it for me inside the black boxes. I went to agriculture colledge back when a calculator that did a square root was a months pay. A fortran class turned me off computers for 20 years. Thanks guys!

 

[fechter]

... A fortran class turned me off computers for 20 years. Thanks guys!

Same thing happened to me. I still hate computers.
I'm working on building a 'how to' sticky for lithium pack construction that will hopefully explaing things in a simpler manner. It might take a while before I get to it...

 

[pwbset]

I'm working on building a 'how to' sticky for lithium pack construction that will hopefully explaing things in a simpler manner.

As someone who is building 2 packs of 10s6p 37v/9ah konion to be run at 74v/9ah please include layman's terms when dealing with series/parallel charging discharging etc. etc. I just am not getting my head wrapped around the necessary diode/Schottkey setup I need. I would like the 2 packs connected all the time in series, but charge them with my 2 37v/3a lipo chargers seperately.

Thanks for the effort sir! It will be much appreciated!!

 

[dirtdad]

You could do it either way, that is two 36v packs in series (with protection diodes) or combine enough cells in parallel to get 40ah and use a single 72v BMS. You could also put two 72v, 20ah packs in parallel, but this would require diodes in series with each pack, which would have to dissipate quite a bit of heat (not so good).

Richard,

Could you send me a link to the thread about doing this? You have may posts to wade through, I appreciate the pointer.

My recollection of the topic was that Ping BMSs could not handle being hooked up in series.

I am using 2 Yesa 36V 20AH in series, Yesa claims this is not problem, and no mention of diodes, but maybe it is implicit. Could the BMS be increased in capacity to handle a series pack connection without the diodes?

I suspect I need to do it anyway, and need help with the wiring diagram and sizing the diodes.

 

[fechter]

See the last post in this thread: http://endless-sphere.com/forums/viewtopic.php?f=16&t=1312&p=70490#p70490

If the FETs in the BMS have a voltage rating that is higher than the entire pack, then you don't need diodes. We know this is not the case with DeWalts and several other duct-tape packs.

The diodes only protect the BMS in the event that it trips and normally aren't doing anything. They should be rated for the full current and the voltage of the battery they are across. Since we hope you back off the throttle if a BMS trips (you'll notice a huge voltage drop and hopefully the LVC in the controller trips), the diodes will only have to take the full current for a very short time, so heating should not be an issue. You could use either a regular or Schottkey diode, but you don't need the lower voltage drop of the Schottkey. Use whatever is cheap and easy to get.

 

[swbluto]

... A fortran class turned me off computers for 20 years. Thanks guys!

Same thing happened to me. I still hate computers.

Ahhh... that's too bad. Computers are your friends, obviously. :lol:

Fortran probably isn't the best language to learn programming from. Neither is assembly. I would have voted "basic"(without using the goto construct that djiistrika(however you spell his last name) so hates) and then graduating to "c" and then "c++". Well... that's what I did.

Now that I'm in college, they're infatuated with Java so that's what I'm using at current. I like it's cross-platform portability ideal that actually seems to work IF you do things the "right way" from the start.

 

[fechter]

Back when I was in school, we wrote the programs on punched cards and ran them on a Univac 1108 that occupied an entire room. Things have improved since then.

 

[swbluto]

Wow. I didn't know you've lived such a long time. I reasoned that 30 years ago was 1978 and even basic existed(Traf-o-Data was created in 1974 I believe), although c was just coming into formation at Bell laboratories but I guess punch-cards would've been the standard college-way I suppose as they seem to rely on technologies that are usually at least half a decade old(Unless you go to MIT, *sigh*).

 

[dogman dan]

I think it was 1977 or 78 when I took the punch card fortran class. After spending about two weeks trying to get a simple program to run, each attempt taking an hour and a half for the cards to get thier chance to read, I found out that in fortran = is spelled EQ. Finally after two weeks, an actually smart lab assistant spotted and corrected my mistake. All the computer would say was " mistake somewhere in the stack of cards" . After that, I quit trying, took my F in the lab section of the class, and got a B in the class.

Years later, like as late as 88, my wife used to try to inventory the chemistry building on a pre windows IBM. She would come home and start drinking like a fish and cussing like a sailor. I didn't touch a computer again till, 1999. Now I'm online allday, and think nothing of opening er up and swapping out sound cards, or formatting a hard drive.

 

[swbluto]

:lol:

Wow, it sounds like you guys had horrible compiling times AND debuggers. Worst case scenario for development, I'd imagine.

Now-a-days, development programs will automatically identify and error in your line right after you type it in so you can correct right then and there. Others just tell you after you compile, and sometimes they're not too clear, but you tend to adjust to their quirks.

 

[64ragtop]

Back when I was in school, we wrote the programs on punched cards and ran them on a Univac 1108 that occupied an entire room. Things have improved since then.

And here I was thinking Fechter was maybe Links older brother.....

I haven't lived this long without learning something(s)!!!

Problem is I've forgotten a bunch of it too!


ATB

BC

 

[dogman dan]

In my case it was the 70's and early 80's at New Mexico State University. Most likely the computer mainframe was something hand me down from Sandia Labs or Los Alamos. So the mainframe was surely from the 60's. A lot of stuff in NM is allways 20 years behind the times. All better thanks to Mr Gates now of course.

 

[64ragtop]

In my .... the mainframe was surely from the 60's.....All better thanks to Mr Gates now of course.

My sarcasm alert tripped when I read that last line, dog. Surely you're NOT a subscriber to the motto:

"No Matter HOW Cynical I get, I Just CANNOT keep up!!!

ATB

BC

 

[dogman dan]

Wow that one could read two ways for sure. While I do hate how much my next software upgrade is gonna cost, I sure do appreciate Bill for taking fortran out of computers! Back in the seventies and 80's the computer system at NMSU really was scavenged out of the dumpster and housed in WWII army barracks. Back when my Wife would come home cussing, she was using dos commands on an IBM workstation with no mouse and a 10 inch screen.

 

[swbluto]

Ummmm... GUI and the Mouse were both developed at the Xerox Parc. You can thank Microsoft for making the combination of the two initially affordable(As opposed to the Mac, which was(and still is) god awfully expensive), and then eventually dominating, and then eventually jacking up the prices once its market-power was vast to become wealthy which we can see it has.

 

[dogman dan]

That is true, Bill didn't invent the mouse, or even the DOS my wife used to cuss at before it came with a mouse. No telling how long NMSU would have been stuck with crappy handmedown IBM mainframe stuff if PC's hadn't gotten cheap.