brewdude wrote: ↑Wed May 27, 2020 1:36 am
Will you please explain the recovery circuit?
What are the spec's of the reverb pan?
Sure. So, the recovery circuit is a hybrid cascode. It leverages the high transconductance of a semiconductor (a JFET in this case) to drive the current through the top triode to achieve very high levels of gain. In this way, it’s similar to a traditional all tube cascode. Even though there are some operational differences, Merlin has a article which gives a good overview of designing a traditional all tube cascode here:
http://www.valvewizard.co.uk/cascode.html
I recommend giving that a read to start.
I opted for a "hybrid" approach because the project presented a couple of challenges;
First, in order to keep the amp architecture and function true to the original design, I only had one tube position to work with. The 6U8A pentode section of the tube was ideal for the driver, but a recovery stage requires very high gain and low noise. Unfortunately, the triode section of the tube is a low/medium mu–high gm type, and is not well suited for this type of task.
It was so frustrating, because all I really needed was just
one more friggin' triode stage to give me a few different options to get the level of gain I needed. But, it wasn't an option without it causing wholesale changes to the rest of the amp. When faced with the prospect of drill out the chassis and installing another tube, or squeezing every last drop out of that single triode and settling for a voltage gain of
22, I said fuck it, I'm cheating.
I started thinking about different kinds of hybrid circuits I'd seen used in other applications. I remembered seeing various hybrid cascode combinations popping up from time to time, but I really didn't have any experience with them previously. I started searching out examples, and discovered that using a jfet as the input stage, and transconductance amplifier in the lower section of a cascode has a lot of advantages. Using it in the recovery amplifier of a reverb circuit, it's almost ideal. The LSK170 jfet in my example is legendary for its low noise characteristics and is uniquely well suited in this application.
Single stage. Very high voltage gain. Low noise. That with the fact that I don't even need to provide a separate low voltage supply for the fet makes it like the perfect cheat code here.
Does that answer your question? I'm not sure how in depth you wanted me to get as far as choosing the right jfet, the saturation region, how to draw up the load lines, etc