Need help finding a bidirectional switching device

[rwolfe]

I'm making an active battery balancer for a battery system that needs a more aggressive balancer than just using resistors.

One of the candidate layouts is a capacitor-based system (see below). For this system, a pair of contacts/relays (shown as red and green switches) for a high cell will close to charge up a capacitor network. Then that pair will close. The capacitor network will reconfigure to put the capacitors in series (and thus boost voltage). Then a different pair of contacts/relays will close to absorb that energy.

I'd like to avoid using electromechanical contacts/relays as the cycle count will be high. To use a semiconductor device, it needs to be able to withstand polarity flips as the circuit operates. This is something I'm having trouble finding.

Can someone advise me how to find such a component?

For reference, average current would be at least 2 A, preferably in the 5-10A range. I could adjust stack size to accommodate voltage tolerances, but, the diagram shown would have the switches seeing roughly +/- 20V.

 

[KlausST]

Hi,

have you done some simulation or calculation ... on efficiency, peak currents, switching frequency, .. and so on?

Please do so...

For loss calculation you have to use RMS current and not average current.

Klaus

 

[rwolfe]

Hi,

have you done some simulation or calculation ... on efficiency, peak currents, switching frequency, .. and so on?

Please do so...

For loss calculation you have to use RMS current and not average current.

Klaus

For this iteration, I'm not needing a really efficient energy transfer. Passive balancers only pull down high cells. I need to also pull up low cells. These come in large groups, so, pulling several dozen down to within a maximum deltaV of the lowest cell is very inefficient. If it works at all, it will likely be more efficient than what's in place now. Conserving energy is not really an issue for this anyway, as energy input is available.

The only problem I might have with energy inefficiency is heat generation. I can add heat sinks, blowers and peltier cells if needed. I can't calculate heat losses until I know what device I'm using.

Switching frequency is undefined for now, but I can set it to whatever value is appropriate using the timers on an MCU. It'll end up being a little bit more than 5 time constants, which can't be defined until I have system impedance.

Having a component to build around will allow me to start working on transients. I can't figure peak current until I have candidate components, as they will add voltage drop/impedance to the circuit. That said, I will add some passive components to dampen transients if impedance is low enough to need them.

Since you asked about calculations I'll share with you this rough approximation:

V.delivered.noload=2V.cell.donor-6V.drop.cell.contact-3V.drop.capacitor.contact
This assumes the voltage drops across the contacts are low enough relative to donor cell voltage to allow charge to transfer every time it is supposed to.

That should be the maximum theoretical cellular output voltage after the system has had time to reach equilibrium at each step, assuming no other losses. The voltage drop across the various contacts is a problem and might make this not work. Using an electromechanical contact would get around the voltage drop issue, but I don't want to introduce so many possible points of failure at the cell contacts.

Right now my immediate obstacle is finding candidate bidirectional solid state contacts. From there, I can start defining the rest of the circuit. I'm looking for help on where/how to find what I'm looking for.

 

[D.A.(Tony)Stewart]

The energy 0.5CV^2 stored in each cap. for both polarities * frequency, E=fCV^2, must be the average power lost (in the switches used + cap+cell (total ESR in the loop)

This is not efficient. But if you could insert a DCDC regulator, it might. I've seen a Patent on this maybe 20y ago.

Lithium Ion Battery Management and Protection Module (BMS ) Teardown - Schematics, Parts List and Working

In this article we will be learning about the features and working of a 4s 40A Battery Management System (BMS), we will look at all the components and the circuitry of the module. I have done complete reverse engineering of this module to find out how it works so that I can show how the BMS works.

circuitdigest.com

one choice (not avail)

• Inductive, Switching control Scheme
• Up to 92% Charger transfer efficiency
• Accurate Balanced voltages down to 30mV
• Auto detect unbalance and auto balance
• Low sleeping supply current, 2uA
• Programmable balancing current up to 2A
• Precondition balancing current
• Status Indications
• Battery Over voltage protection
• Support small size inductor
• ETA3000 is available in one type of package, DFN2x2-8L
• 
  
• http://www.etasolution.com/upload/201905/09/201905091810050738.pdf hmm deadline
• https://forum.allaboutcircuits.com/threads/ballnces-based-on-the-eta3000-doesnt-work.193049/ < works with ceramic caps ( Film caps are better for Irms) but this $0.50 chip only does 2A and may not be cascadable.
  
• 
  Better choice
• Inductive Active Cell Balancer IC - Google Search
• https://www.ti.com/lit/an/slua524b/slua524b.pdf?ts=1729204328906

and finally for 10Ah cells

BQ78PL116 PowerLAN Master Gateway Controller With PowerPump Cell Balancing Technology_WOSENDONGLI代理TI 德州仪器

WOSENDONGLI 代理TI 德州仪器 BQ78PL116, BQ78PL116RGZR, BQ78PL116RGZRG4, BQ78PL116RGZT, BQ78PL116RGZTG4

wosendongli.com

--- Updated 58 minutes ago ---

alternative LT8584