This is a topic that really does interest me. “Most” regen circuits do NOT use a grid cap as high as 250 pF. It’s more likely

to be 100 – 150 pF, pro forma; i don’t think anyone really thought about this; they just plug in the “standard values”.

The grid resistor, i have only in my years seen one with as low a grid resistor as 150k – 250k, and that was such a surprise

to me, i had to do a double take to see if i read it right. It was in one of those circuit collection paperbacks. I was going to

say a ------ ------- editored book, but i don’t know that for sure, so i won’t associate his name with it.

I did a simple back of envelope t=RC calculation for some standard regen values of 1M and 100 pF and it says, good for

audio up to about 15 kHz, so your 20 pF with that grid resistor is about right and reasonable. Why the grid cap should

have anything to do with selectivity puzzles me, tho. It’s easy to understand why the grid resistor, the diode load, affects

selectivity. The grid cap however unlike a lytic has no practical, in these time periods, self discharge; i can only see that

it might if too large, kill your higher audio, which might not be a bad thing, “depending where”.  This topic just reminded

me of ‘crystal radios’. There, the cap across the load has NO, zero effect on selectivity; only too large, it dampens the

higher freq audio. So what’s different about the grid – cathode diode detector in a regen receiver ?

This topic makes me think of the last generation of regens, the MFJ and Vectronics and TenTec type thing. These all

used tiny RF chokes as the L component in the LC. I could not see how such low Q items would be good; it seemed

to contravene anything we learned in the 1930s. I could not prove it mathematically, just picture what it would do as

regen ( pos feedback ) was increased. You have the needle-nose thing, with the center peak rising dramatically, but….

you also have some w-i-d-e  overall response and that wide bandspread does not rise as dramatically, as much as the

center freq. So you have a similar regen max selectivity – sensitivity but the overall freq response is not that selective

compared to a high-Q coil. That’s my theory anyway. I think this was proven out in the 1930s by experience.

-Hue Miller

From: kgordon2006

Sent: Friday, February 27, 2026 8:14 AM
To: Hubert Miller
Subject: Re: [ARC5] Way Off Topic

Yes. Many, if not most, regen circuits used a 250 pfd grid cap.

In many discussions with BoatAnchor Bob Keyes, we discovered that reducing the value of that cap very noticably increased selectivity.

Then, adjusting the grid leak resistor to a value much greater than the usual 100 or 150K also improved selectivity, at least on "weak" signals.

Of course, like almost anything else, it is possible to go too far, so experimentation is in order.