automated battery charge/discharge cycler

[ice595]

Hi, im planning to make a automated battery charge/discharge cycling tester for AA ni-mh cells.
I wanted to test a whole bunch of ni-mh cells and see which one will hold more of their original capacity after a certain amount of charge cycles. (e.g 1000 like it says on the packaging)

I've whipped up a simple circuit to do constant current charging and constant resistance discharge. (see attached schematic)
The only thing that bothers me is the PNP transistor that's configured as a constant current source.
Is there any easy way to use a NPN transistor as a high side constant current source?

[David Hess]

I do not know of a simple way but why does it matter?  Is using a PNP transistor a problem?

If it is an issue, a low side current sink with an NPN could be used with the battery connected to the positive supply.  The discharge transistor could be NPN or PNP.

[macboy]

You could use an NPN as a switch on the high side, and use the series resistor to set the current. If Vcc is 5 V, the current should be "constant enough" as cell voltage goes from approx 0.9 V discharged to approx 1.5 V charged. I like your PNP circuit better.

[Audioguru]

Why measure the battery cells from different manufacturers? Some will last for many cycles and others will fail soon but have a warranty so they can be replaced.
The Ni-MH cells in my Chinese solar garden lights last for about 2 years (365 x 2= 730 cycles). Their old Ni-Cads failed in a week or sooner.

Energizer guarantees their Ni-MH cells for 5 years and the number of cycles can be millions and they will still replace them if they fail within 5 years.

My worn out Ni-MH cells developed a lower than normal fully charged voltage and a higher series resistance that reduces the maximum current. Their capacity is also reduced.

[ice595]

I do not know of a simple way but why does it matter?  Is using a PNP transistor a problem?

If it is an issue, a low side current sink with an NPN could be used with the battery connected to the positive supply.  The discharge transistor could be NPN or PNP.

Its not really a problem if I visit the local component shop, as I don't have any PNP trannies in my parts collection other than the dinky 2n3906
I've also noticed that PNP transistors aren't as common as NPN in most circuit designs, why is that?

[David Hess]

Its not really a problem if I visit the local component shop, as I don't have any PNP trannies in my parts collection other than the dinky 2n3906
I've also noticed that PNP transistors aren't as common as NPN in most circuit designs, why is that?

In general NPN bipolar transistors and N-channel FETs are higher performance than PNP bipolar transistors (unless fabricated with germanium instead of silicon) and P-channel FETs but this was more of a problem in the past.  Today NPN and PNP bipolar transistors built on silicon are comparable although P-channel FETs are not comparable to N-channel FETs.

There is a reason that old bipolar integrated circuit processes were NPN only or NPN with very low performance PNP transistors and logic families like TTL and ECL were designed with this in mind.  The same applies somewhat to CMOS as well.