Sziklai Pair NiMH battery conditioner design

I'm trying to use a Sziklai Pair of transistors to discharge a battery to a precise voltage. Regardless of the ongoing debate on the "memory effect" afflicting NiMH batteries, there are a other situations where it's desirable to perform a discharge cycle to 1.0 or 0.9v on a NiMH cell. In my case, my NiMH charger only charges in pairs but I don't always use my cells in pairs; in such a scenario, one cell will get undercharged and the other will get overcharged. By discharging both to exactly the same state, I can then charge them evenly and thus extend cell lifetime.

**broken link removed**

The idea is at any battery voltage above 1.0, R2 allows the NiMH battery to discharge at .5C (1A for AA, or 400mA for AAA). As soon as voltage falls to 1.0v, R1 will cause T1's base voltage to fall below its natural 0.7v cutoff which switches off T2 and the battery experiences no further drain (over-discharging NiMHs cells seriously damages them or else I could do this with only a resistor!).

I have not built this circuit yet but I'm trying to simulate it in software and I'm not getting the desired result. Could just be I don't know how to run the software right but either way I'd like some input on the above circuit because it seems the Sziklai Pair is shutting down the current (green) at about 1.2v (blue) no matter what value R1 is given? I guess what I'm looking for is a "knob" to adjust the voltage cutoff point for the Pair and R1 isn't doing it (at least, in simulation):

**broken link removed**

I am admittedly a semiconductor noob so I suspect this is just me being dumb with either my design or my simulator (or both)?

Input welcome. Thanks in advance!!

Works for me. Q1 will dissipate in excess of 1W, so it needs to be a TO-220 on a heatsink.

Ah, the simulated transistors had too high of a saturation voltage configured. If I set them to 0.3v the circuit works as expected:

**broken link removed**

Hopefully I can still find such a thing in real life!

MikeMl your current seems to gently taper off as voltage drops, rather than the hard cutoff my simulator gives. The hard cutoff is actually why I used a Pair and not a single transistor... I'm curious why the different behavior between the two simulations?

One more quick question... do I need a resistor to limit T2's base current so I don't fry it?

Thanks!

You DO NOT have a Sziklai pair!
You simply have two cascaded complementary transistors that are completely different but they have a lot of negative feedback.

The goal is for the current to stop when the Battery Voltage reaches a cutoff voltage, say 0.8V.

The slope of the battery current vs the battery voltage is primarily determined by the 1Ω resistor.

Here are four different cases:

Yours. Another with the resistor moved. A resistor in series with three Shottky diodes, and just a resistor by itself. Note the slope of all four currents as the battery voltage is swept.

Seems like it would be preferable to have a constant-current sink with a cutoff rather than a resistor in-series with a cutoff...

Yep I'm thinking a constant current 0.9v sink and precise voltage cutoff, sounds like a job for a dual opamp and a power transistor.

Use one opamp with a current sense resistor to keep the discharge current constant, then the second opamp turns the whole thing off if the cell voltage drops below the setpoint.

It's still getting pretty complex to have that on each cell in the battery...