Searching an answer, found this:

"
Ah, the sadly neglected shunt regulator - haven't had to make one of those for quite a while. The TL431 keeps the voltage across it constant. So it pulls the base voltage to that voltage with reference to the bottom rail. Now, if doing that makes the base sufficiently positive of the emitter, the transistor starts to pass current from emitter to collector. Doing so passes extra current through the supply series resistor - so the voltage across that rises. Which decreases the emitter base voltage. A stable position is arrived at, with a particular current through the transistor. Now place a load on the output - the voltage across the series resistor tries to increase, but, as it does, the collector current will fall. Result, constant voltage output - with current being swapped from transistor to load, as the demand for it by the load grows - and vice versa.

Putting a resistor from base to collector would just current share with the TL431 until there was no current in the TL431 - at which point it could no longer hold a constant voltage."

I'm thinking my first diagram with the resistor in power lead was correct. It was copied from data sheet. Not specific for BMS, just a generic shunt regulator.

 

Continued test with 22 ohms in base circuit.
Down to 3.99v and still shunting. Six previous tests only shunted down to 4.00v.
Removed resistor. Base current is <50ma again.
Now I'll see if switch point is back to 4.00v.

 

First time I have heard of them. Interesting!
Here is a basic circuit base on a TL431. I think you would need the low voltage one (1.25 volts) and adjust the resistor divider for your batteries.

Where did you find this?
I'm having an issue with base current.

 

Where did you find this?
I'm having an issue with base current.

How are you testing it? Don't forget in actual operation the difference between the batteries will only be a few tenths of a volt and the total voltage applied will only be 4.2 volts times the number of batteries.

 

View attachment 110159
"As testing" diagram.
22ohm in base limits current
Omitted 100R from top rail to TL431L
Added feedback R8.

I have 2 positions of 12 populated on board for testing.
I'm hoping to adjust all to <.05 volts difference.

Now set to 2.40 volts.
12S LTO battery pack. Each cell of LTO 1.5 to 2.75 volts range.

 

Ps.

Thanks for all the help.
I'm hoping someone will "spice" it up to save my trial and error.

Good turn around on circuit boards, eh?

To bad I have to add a few parts. Still better than breadboard!

ps. ps.

If someone wants a board, let me know. I had to get 10....................