[bruceme]

In researching "old-school" balancers used in the early days of Lead-Acid, I return to the zener shunt.

Here is my take on a LiFePO4 shunt:


The zener will have to be tightly matched (tested and returned). It would be much easier to match if I could use pairs, but I can't find power zeners at 1.8v. 3.3 is the smallest in this series.

I am building a cost-scaling system for very high voltage packs. The system uses fixed balancers and Lee Hart's "Batt Bridge".

Watch this good YouTube on batt bridge

Any thoughts or critique?

[Qer]

Originally Posted by bruceme

Here is my take on a LiFePO4 shunt:

Well, after the transistor blow up (the diode between base and emitter will short the zener) it will probably work somewhat at least. However, you'll have a pack that will always discharge slowly due to the nature of zeners (the knee is kinda soft so they will pretty much always draw a bit of a current), but if you live in a cold climate maybe that's not only bad since it will keep the pack warm in the winters...

[bruceme]

Originally Posted by Qer

Well, after the transistor blow up (the diode between base and emitter will short the zener) it will probably work somewhat at least.

What's the right circuit?

Originally Posted by Qer

However, you'll have a pack that will always discharge slowly due to the nature of zeners (the knee is kinda soft so they will pretty much always draw a bit of a current), but if you live in a cold climate maybe that's not only bad since it will keep the pack warm in the winters...

The specs for this closed zener is 500ohms;

I=V/R=3.4V/500R=0.0068A

40Ah cells,

40/0.0068 = 5882 hours or 214 days,

Even I can remember to charge in that time frame

As for heat... I^2*R * 80 (cells) = 1.8watts, hardly a space heater. But good things to consider.

Thanks,

-Bruce

[bruceme]

Originally Posted by bruceme

What's the right circuit?



The specs for this closed zener is 500ohms;

I=V/R=3.4V/500R=0.0068A

40Ah cells,

40/0.0068 = 5882 hours or 214 days,

Even I can remember to charge in that time frame

As for heat... I^2*R * 80 (cells) = 1.8watts, hardly a space heater. But good things to consider.

Ok, re-reading your post and viewing the graph I saw what you where trying to say. It's not "binary", the further you are from the knee the lower the current, and it's not by a little, it's by a lot. So yeah, I could calculate this, but it's more than the spec numbers.

[Qer]

The right schematics.... Well. You probably have to do it a lot more complicated with some kind of comparator and a voltage reference etc. That's when things start to get so complicated I kinda prefer to put it all in a micro controller instead. I'm the software nut, remember?

[EVfun]

A very simple idea for a shunt regulator if minimal shunt current is required (between 100ma and 140ma.) Based on my experience with and without regulators I suspect that a very tiny nudge each cycle would be more than enough provided you initially top balance the cells.

View in fixed width font:

          D2 // 
+ ---------->|--
    |     |    |
    |     >    >
 R2 >  R3 > R4 >
    >     >    >
    >     |    |
    |     |-----
    |    --     
    |----/\     
    >   D1|     
 R1 >     |     
    >     |     
    |     |     
- ---------

R1, 2700
R2, 1100
R3, 33
R4, 120
D1, LM431B
D2, red LED

[bruceme]

Originally Posted by EVfun

A very simple idea for a shunt regulator if minimal shunt current is required (between 100ma and 140ma.) Based on my experience with and without regulators I suspect that a very tiny nudge each cycle would be more than enough provided you initially top balance the cells.

View in fixed width font:

          D2 // 
+ ---------->|--
    |     |    |
    |     >    >
 R2 >  R3 > R4 >
    >     >    >
    >     |    |
    |     |-----
    |    --     
    |----/\     
    >   D1|     
 R1 >     |     
    >     |     
    |     |     
- ---------

R1, 2700
R2, 1100
R3, 33
R4, 120
D1, LM431B
D2, red LED

Ok, I get it, voltage splitter driven transistor to a shunt. makes sense and it's simple enough.

[bruceme]

Originally Posted by EVfun

A very simple idea for a shunt regulator if minimal shunt current is required (between 100ma and 140ma.)

Two issues:

- The circuit described by EVfun is likely for 12v...[Edit: I was wrong! I swapped R1/R2 by accident] if you're reading this, you need to adjust the resistors for your battery cut offs. I will post an update later, but My R1 is 1.2k, R2 is 2.5k, producing the required 2.5v for reference trigger at 3.65 cell volts. If you have other chemistry, you will need to adjust.

I used this site to calculate the voltage splitter that drives the trigger.

- 140ma is a very small amount of shunting. LiFePO4's don't need much, but this is tiny. Many other examples online, use the LM431 (aka TL431) to drive a mosfet/resistor and drive much more current. Not saying either is right, just more research is needed. I'm always in favor of simple!

Here is a working simulation of the circuit described above for LiFePO4.





Thanks,

-Bruce

[Qer]

That is a better circuit since you use a voltage reference rather than resistor dividers that has a percentage error and a transistor that will have a Vbc between, ehm, some and some more depending on temperature, current, the value of the Wall-Mart stocks and some crazy chaos butterfly in China...

Biggest problem as I see it is that you have no overheat protection. Shunting means wasting voltage to heat and, well, things can get pretty hot if something goes wrong. It seems like a simple problem, shunting some current when the cells start to get fully charged, but it isn't. It's a can of worm...

[EVfun]

Originally Posted by bruceme

Two issues:

- The circuit described by EVfun is likely for 12v... if you're reading this, you need to adjust the resistors for your battery cut offs. I will post an update later, but My R1 is 1.2k, R2 is 2.5k, producing the required 2.5v for reference trigger at 3.65 cell volts. If you have other chemistry, you will need to adjust.

I used this site to calculate the voltage splitter that drives the trigger.

- 140ma is a very small amount of shunting. LiFePO4's don't need much, but this is tiny. Many other examples online, use the LM431 (aka TL431) to drive a mosfet/resistor and drive much more current. Not saying either is right, just more research is needed. I'm always in favor of simple!

Here is a working simulation of the circuit described above for LiFePO4.

My circuit was drawn for 3.5 volt shunting. Notice that R1 is under R2 in the schematic. Specifically, if 1% resistors are used and an LM431C voltage reference is used then the range should be about 3.48 volts to 3.55 volts turn on. Only a few LM431 types are rated for 150 milliamps, most only for 100ma. In that case the R3 should be 56 ohms and R4 150 ohms to make the device survive up to 4.5 volts (after that you have ruined the cell so who cares about the reg.)

          D2 // 
+ ---------->|--
    |     |    |
    |     >    >
 R2 >  R3 > R4 >
    >     >    >
    >     |    |
    |     |-----
    |    --     
    |----/\     
    >   D1|     
 R1 >     |     
    >     |     
    |     |     
- ---------

Of course the gate of a 2m3907 transistor could be connected between R3 and D1 with about a 240 ohm gate resistor. A 10 ohm load resistor and the LED with R4 could be connected between transistor and ground. Then, for hysteresis, a high value resistor could be connected, one end between transistor and the load with the other end connected to the reference pin of D1. I had a test rig similar to that working with a 2n3906 because that is what I had (very similar behavior, much lower load rating.) This resulted in a reg that sharply turned on, and then off a couple hundredths of a volt lower (helps keep the heat in the load resistor and not the reference or transistor.)

+ -------------------  
    |     |         |  
    >     >         |  
 R1 >     > R4      |  
    >     > R5      |  
    |     |--^^^--|/ T1
    | R3  |       |\   
    |     |         |  
    |-^^^-)------------
    |     |     |     |
    |    --     >R6   |
    |----/\     >     |
    >   D1|     >     >
 R2 >     |     |  R7 >
    >     | D2 \/     >
    |     |    --     |
    |     |     |     |
- ---------------------

R1 1500 ohms
R2 3600 ohms
R3 ? kohms
R4 240 ohms
R5 240 ohms
R6 120 ohms
R7 10 ohms 5 watt
T1 2N2907 PNP transistor
D1 LM431C Voltage reference
D2 red LED

For now I haven't been running with any regulators and I have only had to do a little rebalancing one time. I kinda put that work on the back burner but may revisit it.