Author Topic: Cheapest/smallest way to make an integrated 60s BMS front-end  (Read 1204 times)

0 Members and 1 Guest are viewing this topic.

Offline uer166Topic starter

  • Super Contributor
  • ***
  • Posts: 1026
  • Country: us
Working on 60s1p HV pack out of 18650s with an integrated inverter for fun (camping/burning man power pack). The BMS cost seems to approach ~$1.5-2 per cell when using a stack of integrated daisy-chained 14S Maxim BMS ICs, seems to be not economical for a pack of this type. Any existing ways to make a discrete or otherwise cheaper BMS? I've seen and am using some chinese BMS PCBs with a whole lotta passives and ICs, do they have individual references per cell and comparators etc?

Balancing is not a strict requirement, although I suspect I will come to it eventually considering how long the strip is. Really all I need is safe under/over-voltage cutoffs per cell that will disable the load/charging.

I am aware of a 48V pack 12S alternative with a different inverter, but the main goal of this design is size/efficiency, so if I have to use the Maxim ICs at that price, I will, but curious if there's any alternatives.
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4705
  • Country: au
  • Question Everything... Except This Statement
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #1 on: May 05, 2020, 10:43:45 am »
low power dual comparator per cell, and 2 shared status lines between all cells back to the controller, 1 for over, 1 for under?

probably something like a P-fet to the top of the cell with a resistance to the status lines, and have the power switch with 2 big low side mosfets to break on that condition,

Can also have that low side switch control some LED's to tell you why it cut.

An older external laptop battery of mine used a similar approach, where a second P-fet off the same status signal switched in a very weak balance resistance, to allow the battery to resume charging once a cell gets below the over condition
« Last Edit: May 05, 2020, 10:48:14 am by Rerouter »
 

Offline digsys

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #2 on: May 05, 2020, 11:59:43 am »
Quote from: uer166
Working on 60s1p HV pack out of 18650s ..... I am aware of a 48V pack 12S alternative with a different inverter, but the main goal of this design is size/efficiency ...
I would definitely NOT go with a single string at 220-240VDC, and go with the 48VDC idea - for a few reasons -
48VDC is a "safe" voltage, with little chance of electrocution
With 4-5 cells in parallel,  you have a heck of a lot better redundancy / life cycle expectancy. Single cell strings of that number can require a lot more conditioning.
With 48VDC, the higher current = lower ESR you can achieve a much better efficiency anyway
48VDc is becoming a standard everywhere - vehicles, PoE, security / safety systems. Nearly everything I work with is / has moved to it, so parts are becoming very common.
I work with BOTH HVDC, in electric vehicles, and currently developing several 48VDC systems, so very familiar with both .... so unless you're building an EV, I suggest go with 48VDC ... it also simplifies your BMS as a bonus ... oh and you can plug any PoE equipment (cameras etc) straight in.
 
Hello <tap> <tap> .. is this thing on?
 
The following users thanked this post: uer166

Offline Marco

  • Super Contributor
  • ***
  • Posts: 7044
  • Country: nl
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #3 on: May 05, 2020, 02:13:04 pm »
Working on 60s1p HV pack out of 18650s with an integrated inverter for fun (camping/burning man power pack).
Bit of a dangerous pack voltage to lug around ... that's how you get burning men.
Quote
Really all I need is safe under/over-voltage cutoffs per cell that will disable the load/charging.
Two TL431s and two transistors, the transistors pulling up a signal OR wise.
Quote
I am aware of a 48V pack 12S alternative with a different inverter, but the main goal of this design is size/efficiency
I don't see how the tiny bit more copper needed for 12S5P makes much impact. The greater level of insulation needed probably wipes out the advantage for 60S.
 

Offline uer166Topic starter

  • Super Contributor
  • ***
  • Posts: 1026
  • Country: us
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #4 on: May 05, 2020, 10:44:01 pm »
I am fully aware of the danger, this is not an average homemade pack and not my first time building HV packs (the previous ones were for Formula Hybrid and then Formula Electric cars, with all the requisite safety thingys).

Anyway, the pack will be fully isolated from the enclosed metal chassis, and the output of the inverter will be floating relative to output GND (which is bonded to chassis). It will have an embedded isolation monitor to shut off in case of an isolation failure. If someone grabs a live output while touching the GND, nothing happens, since, again, the phases are floating. The IMD will then shut the pack off due to isolation fault within a few seconds, to handle latent fault scenarios (not strictly necessary to handle only single faults). Obviously no GFCI is needed here either, there's nowhere for the current to leak.

Think of it less as a UPS and more like a very small Tesla/EV pack with an embedded non-isolated inverter with all the same safety functions. Anyway, the BMS cost will be way out of whack in this case no matter how it's done, but it's worth it for me because of the high-power (600W) embedded inverter, and having 98% efficiency (single-stage non-isolated) vs ~90% (48V-> boost to 200V -> invert to 120V AC) is a huge difference in the size and weight of the cooling system.

I keep coming back to the lingering feeling of needing balancing, at which point might as well have a full BMS setup with a stack of 14S monitors.
 

Online jbb

  • Super Contributor
  • ***
  • Posts: 1265
  • Country: nz
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #5 on: May 06, 2020, 02:04:57 am »
If you want to have a crack at this then go for it.

ABLIC make little 1 cell chips with OV / UV / balance comparator. They have daisy-chaining pins to stack up.

However, you will need to add your own thermal and current protection, and won’t get any pet-cell info.

Note: 48V to 170V converters could likely be made to 95% efficiency...
 

Offline uer166Topic starter

  • Super Contributor
  • ***
  • Posts: 1026
  • Country: us
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #6 on: May 06, 2020, 03:44:35 am »
Oh wow those ABLIC S-8209A ICs is exactly what I'm looking for, knew they existed but couldn't find them easily. A bit pricey still, but even with 60 of them it's cheaper per cell than most/all stackable BMS front-ends I was looking at. Does balancing and over/undervoltage detection, which is just right for my application.

It doesn't give any diagnostics unfortunately, would be cool if someone made the same ICs but instead of on/off outputs it's a minimal embedded ADC with a daisy chain interface.

This is only the under/over voltage part of the whole system, my temperature monitoring target is at least 20% of the cells (probably 10 or so thermistors in strategic locations). A fuse for short circuit, an overcharge FET protection (but no over-discharge FET protection, that can be done in the inverter itself).
 

Offline mark03

  • Frequent Contributor
  • **
  • Posts: 749
  • Country: us
Re: Cheapest/smallest way to make an integrated 60s BMS front-end
« Reply #7 on: May 06, 2020, 04:36:45 am »
All I can think is, they had BMSes in the 60s??
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf