Question about glass encased germanium transistors

[Liquitone]

I'm working on a fuzz face at the moment and measuring germanium PNP's to use.
Now I have these mystery germaniums that are clear glass encased and noticed the leakage and gain changes depending on the light conditions. Is this normal?, and is this why for instance these black OC44/45's have a black sleeve around the glass?
One transistor says OCP70, which I assume is actually OC70? I can't find anything on OCP. The other has no markings at all, but I figured out the collector, base and emitter.



In this photo I'm using R'G's gain/leakage circuit and you seen how the numbers change. top row is with the connection to base open, bottom row closed.



I'm really bad at math, but I'm thinking under dark condition this one measures around 200uA leakage and a true gain around 34hfe? Or am I getting the calculations totally wrong? (the OC45 The webshop said had 70-100 gain measured 20uA leakage/74 gain, so it sounds about right) It did sound good in the Q1 position, but the OC45 with a gain of 74 sounded better.

[bluebunny]

OCP70 was a Mullard photo-transistor, which explains the reaction to light.  Click here and scroll down a bit.

[tonyharker]

Yes these are germanium phototransistors - OCP70s. Basically an OC70 or OC71 with no black paint.  
I can remember scraping off the black paint of OC71s to make them.
In general all semiconductors are light sensitive but most are packaged in a lightproof casing to prevent them being affected accidentally.
These would normally be used with the base open circuit.  Light would increase the Base/Collector leakage current thus turning the transistor on.

Tony.

[Liquitone]

Thanks guys! that clears  things up a bit.(no pun intended)

[amptramp]

There are people who use LED's as clipping diodes in fuzz circuits, which is OK, but in some cases, they mount them on the front of the box so you can see when they start clipping.  The problem is, they also respond to ambient light, inserting 60Hz or 120 Hz hum in response to electric lighting and reduced clipping voltages in sunlight.

I used to work on glass cockpit displays for aviation and in some cases we needed a touch screen.  The overlay type caused too much contrast loss, the capacitive type would fail radiated emissions, the sonar type could not accommodate condensation, so we were stuck with optical scanning touch panels.  We used infrared LED / photosensor pairs which worked as long as you had flight gloves on - but IR goes right through a finger, so it didn't work for most of us.  We always had Wayne, our only black engineer in the group, demonstrate it because it only worked for him.  The big problem is to get the screen to operate in bright sunlight and the infrared content is always higher than the visible - even on a hazy day, a car still heats up inside because of it.  We tried using green LED's as both emitter and photodiode in one test and it worked.  Sort of.  The leakage current in the LED's used as photodiodes was substantial and went up rapidly with temperature, so they were not that useful but the photoelectric effect should be accounted for in the design if you use them as circuit elements.  In other words, think inside the box.    

[R.G.]

One sneaky, dirty trick that is possible is to use an LED as BOTH an indicator and a communications pickup. Things done very quickly, say at faster than 120 times per second, tend to be missed by the human eye. So it's possible to use an LED as a mostly indicator, partly photo pickup. You can connect it directly to a pin of a uC where it's driven by the pin as an output indicator, but for a few hundred microseconds you can flip the pin to being an input and read samples of the light impinging on the LED lens. There's some tricky design to be done to make this a reliable com link, but, hey, if designs were easy, we'd all have our George Jetson flying packs now.