DIY Electric Car Forums banner

Estimate BMS cost per lithium battery?

12K views 24 replies 12 participants last post by  Sunking 
#1 ·
I think the title sums it up.

lets say you wana use 3.2v 40ah batteries with a total pack of 20kwh.

thats 156 batteries.

All yo experienced converts please stand up

P.S it has to be a realistic figure.
 
#2 ·
#4 · (Edited)
[Let's] say you [want to] use 3.2v 40ah [cells] with a total pack of 20 [kWh]. [That's] 156 [cells]. All [you] experienced converts please stand up.
I assume that all the cells are in series.

I looked it up in the BMS selector:

  • Cheapest analog BMS, centralized or distributed: CleanPower MiniBMS: $ 1926*
  • Cheapest digital BMS, distributed: Elithion Lithiumate Lite: $ 1763
  • Cheapest digital BMS, centralized: Ewert Orion BMS: $ 1490*
*) Wiring harnesses not included (for 156 cells, for the Orion BMS, about $ 154)



In order of ease of installation and configuration:

  1. MiniBMS, distributed
  2. MiniBMS, centralized
  3. Lithiumate Lite, distributed
  4. Orion BMS, centralized
 
#5 ·
wow, over $1000?

Ok lets be a little more specific. 3.2v 40ah batteries. say 125v battery pack in series of 4. 39 batteries per series.

you want the BMS to do balance charging, cut off power if voltage is to high when charging and cut off power if voltages gets too low. and whatever extra the system may come with.

Would this still be over $1000? does anyone have experience installing a BMS system?
 
#7 ·
why not simplify number of connections with bigger ah prismatics? and.... a BMS is *not* an absolute requirement; there are a number of us that subscribe to the top-balance and 'go commando' let the charger do its thing at end of charge without any BMS or cell level monitoring. If you do a good initial parallel balance, then good final tweak in series before buttoning up, and have no parasitic loads on partial pack to unbalance, Li cells seem to stay remarkably balanced.... so far, in my experience.
 
#6 · (Edited)
You can be as specific as you want.... Go to the links above.... you have everything you need to make all sorts of different configurations. Play around with it.

And you should parallel the batteries FIRST, then series.... If you have 4 series packs of 39 cells and 4 of those in parallel, you still need 156 BMS channels because each cell could have a different voltage. If you parallel first, you get all 4 cells on the same channel at the same voltage, and would only need 39 channels of BMS.

Since you want Balancing, choose "YES" under "SOC Balancing"

I have experience with both Elithion and Orion BMS systems and own an Elithion myself.... but all of the things I've learned are on the web (www.elithion.com and www.orionbms.com).
 
#8 ·
I have been thinking about designing a BMS which would consist of one PCB per cell. It would incorporate a shunt MOSFET which would offload charge current above a specified limit, such as 4.2 volts, and send a signal which could be used to reduce or stop the charging at that point. It would also send a signal if the cell voltage dropped below a minimum level, such as 2.5 volts. I think such a device can be made for about $5 each, or less in quantity. So a 150V (50 battery) pack BMS would cost $250, and a 600V BMS would be $1000. I'm assuming a 10 amp maximum charge current so that the offloading device would be limited to about 40 watts of dissipation. But if the BMS can control the charger it could allow much higher initial rates and then start bypassing cell charging current toward the end of charge when the first cell reaches its peak level. :)

It's possible that a device such as this could be made for much less, or perhaps designed for groups of 4 (12V), 8 (24V), or 12 (36V). My idea is to have this device always connected to the battery pack and constantly sampling the SOC, while drawing no more than about 100 uA average, which would take 50,000 hours to take a 10 Ah cell to 50%. :cool:

The same basic design could be used to perform a full load capacity test by discharging the cell through a load at a specified rate and tracking the voltage over time. The information could e stored in EEPROM or dumped continuously to a PC over a USB port. There are USB PICs that sell for less than $1.30 each in 100 pc quantity. :)

If you think this is a project worth pursuing, please give me your "wish list" of specifications and actual charging and discharging parameters so I can see what it would take. ;)
 
#9 ·
I have been thinking about designing a BMS which would consist of one PCB per cell. It would incorporate a shunt MOSFET which would offload charge current above a specified limit, such as 4.2 volts,
you obviously don't understand *much* and hide it behind technobabble. waiting till 4.2v would certainly be too late for any cell ... the rest of your proposal thus loses all credibility.
 
#14 ·
I am ordering some of these:
http://www.mouser.com/Search/ProductDetail.aspx?R=FDD6530Avirtualkey51210000virtualkey512-FDD6530A

They are 20V 21A with 47mOhm RdsOn and Vg of 2.5V. They are only $0.79 each and $0.53/100. They are rated for 33W power which would be pretty close to the 3.6V or 4.2V cutoff at 10 amps. But actually I would like to use it in PWM switch mode and dissipate the power in an external power resistor which would be about 0.2 ohms. But that is wasteful. I may just make a boost converter with an inductor that will take the 10A charge current and then boost it to a useful level of 12-15 volts which could be used to charge a battery. Or even just transfer the charge to another cell. As a far-out idea I could make a microinverter and dump the energy into the grid. But as I see it, the charge bypass will usually just start with a few amps until the other cells come up and they will all balance quickly.

The idea is that each cell will have its own processor which can run on 1.8 to 5.5 volts, well beyond the range of any individual cell. The PIC can activate an optoisolator to provide a wire-or indicator to the charger that one or more cells have reached full charge. It's also possible to use resistors on each opto output so that the total resistance would drop as each cell is charged and thus determine how many have reached full charge. It could be done in banks of 10 to 20 cells and then a PIC with 10 ADC channels could monitor a total of 100-200 cells and control the charger.

I just ordered 4 of the LiPo 18650 cells for $14.16 including shipping:
http://www.ebay.com/itm/4-Pcs-18650-3600mAh-3-7V-Rechargeable-battery-silver-Tab-W-Tab-/400371084253

That's about $0.26/Wh, and if nothing else I'll have a 14.4V battery supply to run automotive accessories. But I want to test them thoroughly and see if they are any good. The same company also has LiFePO4 as well as NiCd and NiMH. In fact, I also bought these:
http://www.ebay.com/itm/150974934354

That's 8 AA NiMH rated at 3000 mAh for $5.39 + $2 shipping, so that's also about $0.26/Wh, if they are as stated (and I intend to find out). They also have packs of 50 for $32.91 + $2 shipping or $0.19/Wh.
http://www.ebay.com/itm/50-pcs-AA-2A-3000mAh-Ni-MH-Rechargeable-Battery-Yellow-/400381655519

They claim 1000 charge cycles. At least NiMH is not quite as dangerous as LiPo and charging is not as critical. I could make a 360 VDC pack from 300 of these for $210 for about 1.08 kWh and that would be good enough for my tractor project. They are about 20 grams each so 300 pieces is 6 kG which is sure a lot less than the equivalent in lead. It may not be totally far-fetched to use 20 such packs for your 20kWh EV pack, for a cost of $4200 and weight of 120 kG.
 
#18 · (Edited)
I think the point dtbaker was trying to make is to keep it simple for other people to understand, otherwise whats the point in putting up information if it cannot be used by other people?

Thankfully I understand it, and that Mosfet is interesting. Ill compare it with others on digikey.

Ok so the reason Im asking BMS cost is because I have this design on paper which is about $100 - $150. and it does balance charging, sends signal to charger when one cell is maxed out, heat sensor, does all the minimal requirements nessesary to prolong battery life.

I have tested it on a bunch of large capacitors with different values and it works extremly well! im so excited! (still have to test it on actual batteries)

Now I have to compare this with others on the market to see if its better just to buy one or make the one i have designed.

Am I nutz making one? should i just buy a prebuild and proven BMS? Thoughts peoples!:D
 
#19 ·
I'd like to see your design. No sense reinventing the wheel or duplicationg efforts to come up with an inexpensive and reliable DIY design. Of course there is a huge liability issue that might be assumed and expected for someone who knows what they are doing, but inappropriate and dangerous for a beginner. For my own purposes, and small batteries, I'm willing to take that risk, and I would offer the design as open source for anyone to copy, use, and improve upon.

If your design is $100-$150 for a full pack of 150 to 300 volts (50-100 cells), then your cost per cell is actually less than what I stated for mine. Of course the cost would be less for a single board with multiple inputs. I will start a new thread for my proposed design, so it can be discussed (or not) as people choose, and I won't intrude on someone else's thread where it may not be appropriate.
 
#21 · (Edited)
I designed an analog BMS a while back that was a series of 4 cell modules (so number of cells in the pack are variable) and ended up costing about $2.50 per cell in materials. It was for my ebike, but could easily scale up to a larger vehicle. I think the $5/cell estimate is very reasonable for a larger version, especially if you're building it yourself. The cost difference between an analog and a digital one is more the time cost than the material one, analog being the clear winner there.

The thread from my design is here on Endless Sphere http://endless-sphere.com/forums/viewtopic.php?t=48961

[EDIT] I just looked up the BOM from my boards, and at the time it was $7.65 per 4 cells, and another $2-3 per 4 cell PCB. I did it on the cheap and made my own SMD stencil, but for a large pack like the OP is asking about a kapton stencil would do fine and is not expensive.
 
#23 · (Edited)
It's basically the same as the high current shunt regulator (figure 21) from TI's TL431 datasheet with a power resistor between the transistor's collector and ground. http://www.ti.com/lit/ds/symlink/tl431a.pdf

I also added an opto-isolator that is triggered when the TL431 starts shunting to act as a HVC signal. LVC is similarly handled via NCP301 monitor IC.

I've been meaning to put together a page with schematic, BOM, etc but just haven't had time to lately. I will let everyone here know when I do though.
 
#25 ·
I am not going to read all the replies but rather answer on what I learned. It cost $11 per cell based on 100 AH cells. Having said that it might not fit the description of BMS as they are only passive Cell Balance Boards with no communication or monitoring. They simply Bleed the Cells for a slow charge to EQ all cells.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top