Rod Pentode TRF

[Murphyv310]

Hi.
Another interesting thing. Listening to two "Amateurs"? On 1934khz LSB right now, been discussing football but not given any callsigns, Yorkshire accents. Good signal on the TRF and very clear, the reaction control gives a bit of fine tune. On the XHDATA D808 receivable but only just on the telescopic aerial.

[Kalee20]

Your delayed amplified AGC sounds about right Kalee.

Here it is, as at present.

I have a main IF amplifier (2-stage - the advantage of two stages is that each needs to have a gain variation of 10:1 to give an overall 100:1 range). The main IF amplifier feeds the audio demodulator in the normal way.

Prior to the audio detector, I pull off some IF voltage and feed to a side chain, as shown. First, there is a low-gain additional IF amplifier which serves as a buffer. This feeds a leaky-grid detector, the output of which is heavily filtered to remove audio. Because the leaky-grid detector is reasonably linear, the filtered output has a DC level which depends only on carrier strength and not on programme content.

The output of the leaky-grid detector goes more positive with signal strength, not what we want, so a further valve inverts it. A negative bias on the DC amplifier's grid keeps this cut off (so the anode voltage high) until the signal strength is enough for the leaky-grid output to be high enough to overcome the delay bias. Once this happens, the DC amplifier output starts to fall.

The outbid the DC amplifier is scaled and level-shifted and applied to the main IF amplifier screen-grids and suppressor grids.

The circuit is not optimised as yet - the various potential dividers could be combined, for instance - but it works!

Readers may raise their eyebrows at three power rails, however the 1.2V and 120V are standard LT and HT rails; the negative bias line is needed elsewhere (bias for 1p24b output valves) and also for a heavily-morphed infinite impedance detector, so it is already present.

[BusyBee]

Your delayed amplified AGC sounds about right Kalee.

The outbid the DC amplifier is scaled and level-shifted and applied to the main IF amplifier screen-grids and suppressor grids.

The circuit is not optimised as yet - the various potential dividers could be combined, for instance - but it works!

Readers may raise their eyebrows at three power rails, however the 1.2V and 120V are standard LT and HT rails; the negative bias line is needed elsewhere (bias for 1p24b output valves) and also for a heavily-morphed infinite impedance detector, so it is already present.

It's an interesting circuit / approach which will give me food for thought. Thank you for sharing. Your negative supply line is maybe unusual but does not quite come out of the blue to me. My earlier mention/ idea of using an op amp in the AGC line would require similar to handle negative going output (for reverse biasing grids in the normal way). Normally in simpler sets this is derived from signal so it is not obvious. In many ways it may be another way of approaching what you have done but by taking reference from the main detector or signal relative point and having more involved processing. Although it does seem a bit involved at first look your three valves do make for a more traditional look. It will be interesting to see the full finished set.

Tracy

[Amie]

Hi.
Another interesting thing. Listening to two "Amateurs"? On 1934khz LSB right now, been discussing football but not given any callsigns, Yorkshire accents. Good signal on the TRF and very clear, the reaction control gives a bit of fine tune. On the XHDATA D808 receivable but only just on the telescopic aerial.

Have you much aerial on the regen Trevor for receiving top band, or are you just using the Ferrite for that too?
I find that too much aerial / too heavily coupled aerial has adverse effects... Like overloading and damping of the detection

I was very excited once to hear a conversation on top band between what I thought was radio hams, one was talking about being in Paris! then I realised that they were transmitting AM, Naughty... Dawned on me that it was someone local with an old AM house phone. 

Peter...
Well done! Lots of thought gone into that! So no sand or soot anywhere to be seen in the signal path: design driver met

[Murphyv310]

Hi.
No external aerial just the ferrite rod. Already heard two German stations late in the evening on Top Band at the weekend. The coverage is around 2mhz to 480khz. Quite wide indeed.

Definitely LSB and managed to record it onto my phone directly from the headphones.

[Murphyv310]

As an aside.
The low voltage 1J42A is greatly effected by external magnetism the others to a lesser extent.
Put a coil close to the valve with an AF signal and low and behold you get audio out, a magnet when it gets close results in lower gain, we of course know the reason, perhaps a crude AGC with a coil strategically placed .......... OK where's my coat

[Amie]

As an aside.
The low voltage 1J42A is greatly effected by external magnetism the others to a lesser extent.
Put a coil close to the valve with an AF signal and low and behold you get audio out, a magnet when it gets close results in lower gain, we of course know the reason, perhaps a crude AGC with a coil strategically placed .......... OK where's my coat 

Oh I'm trying that one! AGC deflection coil!

[Kalee20]

Look forward to the results Amie, wacky ideas which work , result in huge amounts of lab cred

Tracy, I probably should've started a new thread for AVC rather than hijack Trevor's thread, but were already talking about it before my post. Your suggestion of an op amp is well worth looking at, why don't  you get a traditional superhet and put an op-amp in that to drive the AVC line and see!

My circuit has the high-gain DC amplifier. At the moment the gain is limited by the 33 megohm feedback resistor (I could've looked at system stability to avoid AVC motorboating but chickened out and limited the gain!) but that's just the first cut. In principle I'm aiming to take that out and let the gain in there go skyhigh, as high as the 1j24b can provide, which will approximate your op-amp anyway. That should keep the output level absolutely constant until the AVC system runs out of range.

[BusyBee]

As an aside.
The low voltage 1J42A is greatly effected by external magnetism the others to a lesser extent.
Put a coil close to the valve with an AF signal and low and behold you get audio out, a magnet when it gets close results in lower gain, we of course know the reason, perhaps a crude AGC with a coil strategically placed .......... OK where's my coat

I have tinkered with this before, I forget for what reason but it was not sucessful. The Russians always seemed to put screens on these valves. I generally follow the same practice, but individually make them from tinfoil as the only Russian examples I have seen are in old equipment which tends to be expensive.

Look forward to the results Amie, wacky ideas which work , result in huge amounts of lab cred

Tracy, I probably should've started a new thread for AVC rather than hijack Trevor's thread, but were already talking about it before my post. Your suggestion of an op amp is well worth looking at, why don't  you get a traditional superhet and put an op-amp in that to drive the AVC line and see!

My circuit has the high-gain DC amplifier. At the moment the gain is limited by the 33 megohm feedback resistor (I could've looked at system stability to avoid AVC motorboating but chickened out and limited the gain!) but that's just the first cut. In principle I'm aiming to take that out and let the gain in there go skyhigh, as high as the 1j24b can provide, which will approximate your op-amp anyway. That should keep the output level absolutely constant until the AVC system runs out of range.

Yes I have fully realised that I am tramping a bit on Trevors thread with some of the things I am putting too although it's me being typically impulsive plus sidetracking, not intentionally and I apologise to Trevor for it. I have an idea which I may go to, still in the embryonic stage.

I've not seriously got to the op amp stage yet. It's just an option I thought of. I need to consider the detector and derivation of a basic AGC signal first. I only have two simple superhets and don't really want to risk messing them up as one is a bit cramped and the other works well and is a nice radio at present which I use.

[Murphyv310]

I love the input everyone, really interesting diversions.
Keep them coming.