LiFePo4 active Balancing/ bidirectional Flyback BMS

[PeterG.]

Hello Community..

I am trying to build a bidirectional Flyback Balancer BMS.
At the beginning, i would like to balance 4 Cells… (at the end over 100 Cells, but not directly in a automobile)
So there will be 4 individual Flyback converters…. one for each cell.
I would like to operate the BMS in a Automotive, with LiFePo4 Batteries.

The Batteries are directly charged by the Alternator.
The ECU, Engine etc. is directly connected to the LiFePo4 Batteries (like normal with Lead Acid but there is no LeadAcid Batterie any more).
The 4 flyback primary sides are connected directly (0.5m cable) to each Cell.
The flyback secondary sides are connected directly to the Batterie pack 0V and 14.4V (Bottom Cell1 / Top Cell4)

while discharging: the primary side will switch (4x)10Amps @ 60-100Khz (2.5 - 3.6V (4.5V))
while charging: the secondary side will switch (4x)5Amps@ 60 - 100Khz (10 - 14.4V (18V))

The maximum allowed conducted Emission (in a european automobile) @ 100Khz is around 70dBµA that means
1uA*10^(70db/20db) = 3.16mA.

http://www.fordemc.com/docs/download/FMC1278.pdf (Page 25)

it is just a little Problem to design a Input filter for each Flyback converter, but it is a huge Problem to design a output filter… because it is just toooooo big, and i would need 4 Input Filters and 4 output Filters…

My Question is relatively simple, but i cannot find the answer, and People i have asked… are not willing or not able to answer that question.

My question is.

If i connect the BMS and the flyback converters DIRECTLY (0.5m cable) to the Cells, and the LifePo4 has a typical internal resistant of 1mOhm… (in my case 0.06Ohm?)  is it then really necessary to design a filter for 70dBµA?

If it is not necessary because the cell resistant is that small, how much current ripple would be okay?

There are 6Cables

0V
Cell1 +
Cell2 +
Cell3 +
Cell4 +
and the Flyback secondary side is separately connected to Cell 4+

for better understanding attached a simple (experimental) Spice model...

Thanks for any help,

Peter

[Jay_Diddy_B]

Peter,
Welcome to the forum !!

Linear Technology has an integrated circuit designed to do the balancing. Have a look at the LTC3300

Link: http://www.linear.com/product/LTC3300-1


This is normally paired with the LTC6804 BMS chip.

Active balancing, generally only makes sense if the pack is large typically 100Ahr. If the pack is smaller, it is cheaper to add more cells than do the do active balancing.

The LTC6804 supports passive balancing, these are just discharge resistors that can be turned on across each cell.

For active bi-directional balancing you need a transformer, 2 MOSFET, part of a chip, some passive parts. You should expect to pay $3-5 per cell.

I think the EMC specs are for normal vehicles, for the 12 and 24V sub-systems. I am sure that the main motor drive in an EV or HEV will have much higher conducted noise.


Good luck with your project !!!

Regards,

Jay_Diddy_B

[PeterG.]

Hey Jay...

Thanks for your answer...
Yes I will use the LTC 3300 but i will not pair the LTC3300 with LTC6804.
on this board i will use BQ76PL536PAPR.. the LTC6804 need minimum 11V (to close, if batteries are empty).. 
But i will use the LTC6804 for the bigger project.

for the 4S system i use 24 CALB CAM 72 Cells...
4S6P... so

14.4V@440Ah.

The system should operate in a normal fuel powered car ... it is not a Electro Vehicle.
Thats why i have to take care about the "normal" Automotive EMC requirements.

[Jay_Diddy_B]

Peter,

With a 440A/hr cell the active balancing makes sense.

Regards,

Jay_Diddy_B

[David Hess]

EMI can be reduced by limiting the loop area of the wire pairs between the balancer outputs and the battery cells.

Most of the designs I have seen mounted the balancer circuits directly adjacent to the batteries.

[PeterG.]

EMI can be reduced by limiting the loop area of the wire pairs between the balancer outputs and the battery cells.

Most of the designs I have seen mounted the balancer circuits directly adjacent to the batteries.

Mounting a passive Balancers on the Top of each Cell is easy!
Mounting a active an bidirectional balanceres on the top of each cell, is not easy.. and the Problems would be still the same.. because the flyback secondary side ist still connected to the top of the stack and the bottom of the stack.
Means big loop area of the secondary side.

since i use Input filters, the HF loop area of the primary side is only on the PCB.

the problems with cell mounted balancers would be exactly the same = big unfiltered  loop area of the secondary side.

What i try to find out is, how much current ripple on the Secondary side would be ok.
The 70dBµA would only make sence to me, if the wires are not connected directly, to such a low impedanz batterie.

The batterie will swallow every ripple.
I can not see any Problem with conducted Emmision at all, an i believe that the allowed current ripple is much higher than 3mA.
But i neither have the experience nor the knowledge.

[PeterG.]

No ideas?
I thought this question would be easier to answer... :-)

Regards
Peter

[Jay_Diddy_B]

Peter and the group,

I think the solutions are in the details. The Ford EMC specification talks about measuring the conducted ripple currents with an CISPR 25, ISO 7637-2 LISN in the circuit that is not terminated.

This puts a 5uH inductor in parallel with 1k Ohm in series with the battery input.

I would look at the details of the measurement.

Regards,

Jay_Diddy_B

[legerme1955]

Hi, i rod your post and i just notice you whe are working out , in Spain an active balancing BMU with LINEAR TECHNOLOGIES, LTC 3300 and companions , very few people try to dare with this tecnologie may be the most performant , actually ltc 3300 is a mosfet driver lt 6804 cell monitor and lt 6820 like spi comunicaction and an mcu to control overall .Whe are working in order to obtain more than 10 amps transfer energy balancing. If you are interested whe can share info about. thanks best regards. G.Seguy, Barcelona Spain