Battery Managment System

[iss407]

I am converting an electric car and one of the big concerns is the battery pack. My car will have 20 6v golf cart batteries in series but many combinations of 6v, 12v lead-acid batteries are possible, even 3.7v Lithium packs. No two batteries are exactly the same and some will charge faster or slower than average, and some will run out of juice before others and become over drained. A good BMS will regulate each battery's voltage individually.

A common simple BMS is just a shunt regulator that diverts some of the charging current through a resister when the battery reaches ideal voltage. This is quite a waste and doesn't help when the pack is discharging. Most chargers charge the whole pack at once. Common average continuous currents are 15-30 Amps charging and 100-200 Amps discharging.

My idea was to capture the energy shunted around the battery and push it back into the whole pack. It is a simple DC-DC Boost circuit but regulating the input voltage, not the output. The charger will regulate the overall pack voltage. When discharging the motor controller will use the unregulated pack voltage.

I've attached the basic design I've come up with. Each battery will get identical BMS modules. No external control is required. The BMS modules attach to the main battery leads, of course, as well as the main battery pack. A daisy-chain of precision resistors will provide each module with the average battery voltage. Two voltage dividers feed the PIC microcontroller with the divided voltages so it can determine how much above or below the average voltage the battery is. I use a simple pot in one divider circuit so I can fine-tune the dividers without needing precision resistors.

The microcontroller has built-in comparitors and A/D converters to use. It also has a PWM controller that I can use for the boost circuit. I am using a microcontroller both to save costs and to allow me to add additional functionallity later. Daisy chaining the PICs with opto-couplers will allow centralized control or data collection, for example.

An inductor and a couple of diodes provide a simple and effective way to shunt the current around the battery and send it back into the entire pack. By only shuting off one FET at a time it would even be possible to shunt the voltage up or down the pack instead of through the entire pack, but I don't see that being needed. Since the battery pack voltages will never be inverted from the battery itself the two diodes alone will isolate the BMS module from the pack.

While charging any batteries that charg faster than average will shunt the extra charging current to the rest of the pack, protecting them from over charging. While discharging any batteries that drain less fast than average will provide extra current to the load reducing the load on the overall pack, protecting weak batteries from over-draining.

The size of the FETs and the inductor will determing the max current that can be shunted. I was thinking 10A, but it would be easy to adjust that. Overall component costs would be relatively low, especially compared to other BMS systems: **broken link removed**. There is one system that uses a similar idea but only shunts the current to an adjacent battery: http://www.evamerica.us/BMS_Brochure1-19-07.pdf.

What do you think, is this feasible? What am I leaving out? High-voltage PIC microcontrollers are available that can run on up 16V+, so no on-board regulator would be needed.

Thanks for your input. This is my first post.

-- Paul

 

[Diver300]

It looks OK to me. It certainly looks like a more efficient way of keeping the batteries even.

I would question the idea of using a battery pack with significantly mismatched cells.

 

[iss407]

No, mismatched batteries would be bad. I will be using identical batteries from the same lot bought at the same time, but even minor difference in manufacturing between the batteries can cause problems when pulling over 100 Amp continuously.

 

[Willbe]

Cuidado, senor. . .these batteries can probably vaporize a screwdriver.

There might be quite a bit to designing, or even just specifying, a strategy.
I guess the worst event is running out of juice; you may want to use an extrapolator h/w or s/w device or program that anticipates this outcome before it happens and warns the operator.

And will you be fusing? It protects the operator and equipment but may leave someone stranded. So for emergencies you may want a fuse bypass.

This is pretty ambitious.

If you come up with something "novel" or something that "wouldn't be obvious to one of ordinary skill" in this field, for a few bucks you could get Statutory Invention Registration to protect your invention, so that the company who steals your idea can't sue you for patent infringement.
Yeah, it happens.

 

[iss407]

EVers know to be very careful around battery packs. Of course they are fused and normally have a circuit breaker at a lower value. I will use a Paktrakr system that measures each battery voltage and the pack current to track depth of discharge and usage data. All that is standard. Part of the reason I want to use a microcontroller instead of a pwm chip is so I can add that kind of functionallity.