EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: Rob190 on October 15, 2021, 05:28:35 pm

Title: Help needed understanding an old PSU circuit
Post by: Rob190 on October 15, 2021, 05:28:35 pm
The attached diagram is the PSU of a compet-32 calculator, dating from the late 1960s. The PSU does actually work but I'd like to understand the function of the two NPN transistors in the yellow box. They are both Hitachi C641 NPN but connected in a configuration that really does not make any sense to me. I've included a photo of the relevant area of the PCB illuminated from the underside. Of course it's possible I have made a mistake with tracing out the circuit but it's fairly easy to see how they are connected.

Interestingly, the PCB is marked with a diode symbol which indicates the transistors might be acting as a zener diode.

The calculator uses a number of 'imaginative' circuit techniques in various places e.g. BJTs used bi-directionally, so anything is possible.

Any ideas?

Rob.
Title: Re: Help needed understanding an old PSU circuit
Post by: TimFox on October 15, 2021, 05:38:28 pm
The dual transistor is connected as a temperature-compensated Zener voltage reference:  a reverse-biased diode breaking down (Zener) in series with a forward-biased diode.  (Normal TC Zeners, e.g. 1N821, use the same circuit.  Other TC Zeners use higher-voltage Zeners with more forward-biased compensation diodes.  BJTs have relatively low breakdown voltage for the reversed B-E diode.)  The temperature co-efficients of the two diodes are opposite and cancel (at a given current).  Where does that line to nowhere connect?
Title: Re: Help needed understanding an old PSU circuit
Post by: tooki on October 15, 2021, 05:39:42 pm
I think it might be a transistor zener clamp, similar to this:

https://youtu.be/BGcKjy_UNQ4 (https://youtu.be/BGcKjy_UNQ4)

See also: https://www.eevblog.com/forum/beginners/multimeter-input-protection-what-are-these-bjts-doing/ (https://www.eevblog.com/forum/beginners/multimeter-input-protection-what-are-these-bjts-doing/)
Title: Re: Help needed understanding an old PSU circuit
Post by: MRIKen on October 15, 2021, 05:42:28 pm
Another reason for using transistors instead of diodes, marketing.  This way they can claim x transistors to make the product sound better, even though they are not used as transistors.  I have heard of this marketing ploy back in the day
Title: Re: Help needed understanding an old PSU circuit
Post by: Rob190 on October 15, 2021, 07:05:00 pm
Quote from: TimFox on October 15, 2021, 05:38:28 pm
The dual transistor is connected as a temperature-compensated Zener voltage reference:  a reverse-biased diode breaking down (Zener) in series with a forward-biased diode.  (Normal TC Zeners, e.g. 1N821, use the same circuit.  Other TC Zeners use higher-voltage Zeners with more forward-biased compensation diodes.  BJTs have relatively low breakdown voltage for the reversed B-E diode.)  The temperature co-efficients of the two diodes are opposite and cancel (at a given current). 

Ok thanks. So how would you know what the zener voltage is? It is really consistent enough or would they have tested different transistors until they found one with the right breakdown voltage?

Quote
Where does that line to nowhere connect?

Which one? I don't believe I've missed anything happy to check again.
Title: Re: Help needed understanding an old PSU circuit
Post by: TimFox on October 15, 2021, 07:21:35 pm
The line connecting the two collectors that crosses (without a dot) over the +12 V output.  If it were dotted, it would short out the C-E of the transistor.
Typically, NPN bases will break down around 6 V.  The 1N821 reference zener is rated for 6.8 V, of which 0.7 V is the forward-biased zener.
Title: Re: Help needed understanding an old PSU circuit
Post by: Rob190 on October 15, 2021, 08:04:33 pm
Quote from: TimFox on October 15, 2021, 07:21:35 pm
The line connecting the two collectors that crosses (without a dot) over the +12 V output.  If it were dotted, it would short out the C-E of the transistor.

Assuming you mean the 2SD175 collectors, that's just the unregulated output from the rectifier.

I didn't label everything. The 'o's represent the outputs from the board. From the top:

Quote
Typically, NPN bases will break down around 6 V.  The 1N821 reference zener is rated for 6.8 V, of which 0.7 V is the forward-biased zener.

I didn't realize the Vbeo of BJTs would be so consistent. Not something I've ever thought to measure!
Title: Re: Help needed understanding an old PSU circuit
Post by: Rob190 on October 15, 2021, 08:14:08 pm
Quote from: tooki on October 15, 2021, 05:39:42 pm
I think it might be a transistor zener clamp, similar to this:

https://youtu.be/BGcKjy_UNQ4 (https://youtu.be/BGcKjy_UNQ4)

See also: https://www.eevblog.com/forum/beginners/multimeter-input-protection-what-are-these-bjts-doing/ (https://www.eevblog.com/forum/beginners/multimeter-input-protection-what-are-these-bjts-doing/)

Interesting. Thanks!

It actually answers another question I was going to post here: Is it still possible to get hold of bidirectional BJTs? The core memory driver circuit uses 2SC17s in a bidirectional configuration and I wouldn't be surprised if some of them have blown. 2SC17s are hard to find and the price is quite high. The video lists a few options and searching for 'Reverse HFE' found a few more.