A Germanium Regen

May 20, 2013.

Updated June 2, 2017.

I came across this schematic from an old GE publication. It looked pretty simple and I was looking for something to do with the batch of 2N107's I came across while looking for transistors for a friend's Fuzz Face build. The circuit looks odd, with no visible bias scheme and no bypassing. Sort of like they were trying to do the same thing as I did with the ultra-simple regens on this site.



 I found a loopstick antenna from a defunct Pioneer stereo receiver, and some variable caps from a similar source. I used hot-melt glue to hold all the parts down onto a plastic cutting board from Ikea. See, it has a handle! It's portable!

The thing sorta worked the first time it powered up. There is an antenna shown in the schematic, but I just let the ferrite bar pick up the RF. Using the coupling cap value shown in the schematic there was squegging at a low audio frequency, so I changed the value to 500pF, and the radio became more usable. The large value of feedback capacitor (365pF) was too much, so I used one of the original FM tuning sections instead. That was better, but the gizmo is barely controllable because the transistor characteristics drift so much with temperature. If I put my finger on the transistor for a few seconds it heats up and the radio oscillates no matter what you do with the regeneration control.

When it is oscillating, the radio responds to lots of carriers on the broadcast band, but it can only demodulate a few of the strong locals. The yellow wire in the picture leads to another link winding on the loopstick. Attaching a short antenna to it brings in a lot more stations, but it's not easy to tune. The feedback variable has a huge effect on the tuning frequency, so you really have to use both hands to tune a station in.



I might mess with this a little bit more. A pot in the battery line seemed to tame the regeneration, but reduced the audio level.

June 2, 2017.

Well, I did mess with it a little more. The "implied" biasing scheme really bothered me, so I stuck a resistor from collector to base implementing the cheesiest bias method possible. Now it works great. That change seemed to take care of all the problems the radio had. Regeneration is now smooth, and you can tune in a lot more stations because it's just more stable. I reinforced the hot melt glue joints for added reliability and generally cleaned up the project. Nothing like a well-engineered piece of gear. I have since moved to Santa Barbara, where the local NPR station broadcasts on 1340kc. That station comes in really well, so I listen to this thing a lot using a pair of WE 509W headphones. Kinda hurts your head after a while.

Back to Regens