Grid leak VS Diode detection in regen set, advantages disadvantages

[neazoi]

Hi, I consider building a small tube regenerative set and try some things out for experimentation.

I have found two detection techniques.
The first uses the "grid-leak" and this is common in most regen receivers I have seen, like this http://www.makearadio.com/tube/12at7-6ak6.php
The second uses the internal diode in some tubes and they say that this has a less distorted sound
**broken link removed**
Also I have found this circuit, a diode-tube combination that he praises for its sensitivity and low distortion http://www.valveradio.net/radio/high-sensitivity-am-detector.html

Have you got any experience, comments or ideas on the above to suggest?

 

[betwixt]

I doubt you would notice if there was any difference in distortion. Grid-leak works perfectly well, the grid/cathode form a parallel diode anyway.

It doesn't state what the B+ voltage is but I would be cautious about the rectifier diode types, personally I would go for ones with higher PIV rating, maybe 1N4006 or 1N4007.

The anode supply to the first stage is unconventional, normally the load resistor goes at the top end of the choke rather than the anode itself. If you move it you might get a marginal increase in gain. Also, I can't see any advantage in the first stage cathode using that 10uF capacitor. The purpose is to allow the choke to block RF and let the resistor do the grid biasing but it would be simpler to just place the choke in series with the resistor and it would avoid a potential resonance peak as well.

Note that in the second link, the 'power amplifier' only consumes 1.5mW (excluding heater) so don't expect to drive a loudspeaker!

Brian.

 

[neazoi]

I doubt you would notice if there was any difference in distortion. Grid-leak works perfectly well, the grid/cathode form a parallel diode anyway.

It doesn't state what the B+ voltage is but I would be cautious about the rectifier diode types, personally I would go for ones with higher PIV rating, maybe 1N4006 or 1N4007.

The anode supply to the first stage is unconventional, normally the load resistor goes at the top end of the choke rather than the anode itself. If you move it you might get a marginal increase in gain. Also, I can't see any advantage in the first stage cathode using that 10uF capacitor. The purpose is to allow the choke to block RF and let the resistor do the grid biasing but it would be simpler to just place the choke in series with the resistor and it would avoid a potential resonance peak as well.

Note that in the second link, the 'power amplifier' only consumes 1.5mW (excluding heater) so don't expect to drive a loudspeaker!

Brian.

How about the third link I posted? He seems to know what he is doing. I consider the circuits in figure 6 and 7

 

[betwixt]

Figures 6 & 7 show regenerative rceivers, the first link is to a reflex receiver which is slightly different in operation. Regardless, the control grid to cathode diode in any amplifying tube will be just as effective, if not slightly better than a crystal diode. The control grid reduces anode current when it is negative with respect to the cathode an it works like the anode of a diode when it is positive, conducting current to the cathode and also not restricting anode current. Whether the diode is external, internal or just the grid to cathode make very little difference to the performance.

Brian.

 

[neazoi]

Figures 6 & 7 show regenerative rceivers, the first link is to a reflex receiver which is slightly different in operation. Regardless, the control grid to cathode diode in any amplifying tube will be just as effective, if not slightly better than a crystal diode. The control grid reduces anode current when it is negative with respect to the cathode an it works like the anode of a diode when it is positive, conducting current to the cathode and also not restricting anode current. Whether the diode is external, internal or just the grid to cathode make very little difference to the performance.

Brian.

Here are some changes I did on the third circuit to hopefully operate on a low voltage.

1. Instead of using a variometer for L2/L3, use static coils and connect a shunt variable capacitor in order to control the feedback. Kitchin in his high performance direct conversion receivers article, states the capacitor method is better than the rheostat.

2. put a 1K:8R transformer and take the audio from the 8R for further AF amplification.

3. Use a VCC of 12v for the filaments and the anode of the 12DW8 (max anode voltage of the tube is specified to 15v). I will probably need to replace this 680k resistor in the bias with something smaller (what value?)

4. Optionally, remove L1 and feed the RF input directly to the top of L3 using a small value capacitor, just to make things simpler (not implemented yet).

5. Optionally, use the other triode inside the 12DW8 as an RF preamplifier and AF preamplifier, as it is done in this regenerative receiver http://www.makearadio.com/tube/12at7-6ak6.php I do not want to talk about this in this topic, it is a future idea I have. However I mention this here because I want to ask, which of the two dissimilar triodes of the 12DW8 should I use for the RF/AF preamplifier and which one for the regeneration stage?

Your comments and suggestions to the above points will be valuable.

 

[betwixt]

You need to experiment with this, there are no hard rules and designs of this nature are very critical of construction and component values.

My observations for what they are worth:

1 & 2. The output transformer will not work reliably like that. You are effectively adding the load of the transformer across the feedback capacitor which makes it's value subject to changes in the load. It also risks resonances in the feedback path which could make it unstable. The feedback capacitor is in series by the way, not in parallel. I would agree with capacitive control being far better than variable inductive control.

3. The 680K has nothing to do with the 12DW8, all it does is provide a small forward voltage to the 1N4148 to ease it into conduction. Choose a value that gives say 0.5V highest voltage. You could experiment with using the 12DW8 internal diode instead of a 1N4148, it might work better.

4. Won't work, the phase is wrong. The feedback coil winding decides whether you get positive or negative feedback, you want positive feedback but a direct connection will give negative feedback.

5. My preference would be to use the half with lowest capacitance as the RF stage and use the other half as an audio amplifier. Replace the audio transformer primary with two resistors (or a choke in the lower position) of about 10K each and take the audio from the tap between them. That isolates the output from the feedback path without losing too much audio level.

Brian.