Strange Reaction

[YT2095UK]

I had a similar thing happen a week or so again whilst experimenting with a pentode (I`ll find the post a bit later), these seem to be extremely good for regen radios if wired up a certain way, and don`t seem need any sort of conventional feedback tickler coil to work well.
a Very simple peice of kit that I find invaluable is an RF sniffer, made with a a simple Micro amp swing arm meter (I use an old signal meter out of a CB) with a 10k resistor and a germanium diode.
you connect one end of the 10k to one of the Diode, and then the remaining end go across the meter (observing polarity of course), the Junction og the resistor and diode is the sensing point, solder a bit of wire to that and probe your circuit with this single wire.
any RF present will move the needle, the amount depending on the strength, it`ll help you to See where RF is and how strong it is.
this may help you find the bit that`s creating the feedback

found it, Post #7 in here: https://www.vintage-radio.net/forum/...ad.php?t=45793

[MichaelR]

Hi Bob,

The idea of the capacitor is to bypass any RF down to Earth.I am thinking that RF is getting into the heater arrangement hence a 0.1 uf capacitor from each valve filament pin to earth .... then try also a 0.1uf capacitor across the 560ohm resistor.

I am assuming your earth connection was to the chassis and your LT- is connected to chassis.

If you stop the oscillation put your regen circuit back in again.

The sniffer probe by yt2095uk is a good idea also.

Mike

[kalee20]

With the circuit as is, there will be amplified RF on the DAF96 anode.

This then is applied to the DL96 grid, and the DL96 will amplify it again. So the slightest whiff of feedback from DL96 anode to the input such as from headphone leads to aerial,, will make it oscillate.

Bypassing the headphones with a capacitor will tame things a lot. But as there's no need for RF on the DAF96 anode, I suggest bypassing this with 100pF or so as well.

If you do add, at a later date, a conventional reaction winding, fed from the DAF96 anode, you'll have to re-enable the RF here. The usual trick is to use an RF choke, taking reaction current directly from the anode - and bypass the other end of the choke to chassis with a small capacitor. Again, this stops the RF from getting into the following stage and wreaking havoc!

[Variometer]

I tried a 0.1 uf between F+ and earth and it made no difference at all. Then I put a 0.1 uf across the 560 ohm resistor and ended up back where I started. i.e. The receiver still works, but selectivity now dreadful and it wouldn't oscillate at all, with original reaction circuit put back in. Not even when turning HT up to 90 Volts! Reversing connections to reaction coil made no difference either. Will have to do some more "fiddling," but if it fails, will remove capacitor from across 560 R and just leave it with its peculiar reaction control of screen voltage pot.
Bob

[MichaelR]

Hi Bob,

I am assuming you mean that you have stopped the oscillation. Surely this is what we were trying to do. You were getting instability, unintended oscillation.

Now we have to get the reaction control to work. What component values have you got in the reaction circuit and how is it wired. Incidentally when trying out the circuit did you adjust the screen grid variable resistor.

Mike

[PJL]

How's the reaction wired?

[Variometer]

Mike,
No, I never wanted to stop the oscillation, the capacitor across the 560 Ohm brought me back to square one and I have now removed it and it is working perfectly again, but with an unusual reaction control, i.e. no choke or capacitor, but simply a pot on the screen volts of V1!

To recap. When I completed it, the volume was very load, but it wasn't particularly selective. There was no oscillation at all, and I couldn't get the reaction to work at all (I wanted the controlled oscillation that is reaction). I found that increasing the HT voltage, I could obtain a reaction effect without the normal reaction components. Varying the HT is normally not a practical proposition, so I varied the screen voltage and that worked fine. Excellent volume, lots of stations and good selectivity and controlled reaction.

PJL
The original reaction circuit was one I had used hundreds of times without problems. The reaction coil was a 100'h choke mounted next to the lower half of the aerial coil. The ends of a 47K pot was connected across it and the slider to earth. One end of the pot went to the anode of V1 via a small capacitor. Initially, I had a choke between the anode and the bottom end of the 100K anode resistor, but it made no difference in or out. I tried various values of reaction capacitor!
The fully functioning circuit as it is now is shown in #4 above.

It is working perfectly, but is rather unconventional as far as the reaction is concerned.

Normally, reaction circuits are no mystery to me and the fact that this one is makes it more interesting.
Bob

[MichaelR]

Hi Bob,

I think the screen control is sdjusting the gain of the pentode and causing your circuit to oscillate.

What stations do you receive ?

Mike

[YT2095UK]

just out of curiosity, is your pot a Wirewound sort?

[Variometer]

Mike,
Yes, it is. That is what I want it to do. It is an unconventional reaction method, but it works! It receives all sorts of stations from local to continental ones in the evening.
The pot is a miniature carbon track.
Bob

[MichaelR]

Bob,

Can you set the pot so that it is not oscillating.

When you vary the pot does the volume of the audio vary.

If you remove the 2nd valve does the oscillation stop.

What frequency range on received signals have you got.

Mike

[kalee20]

For reaction to work, there has to be (positive) feedback. The circuit diagram as shown has no feedback mechanism. So what we're about, is trying to understand where this fedback is arising.

It's satisfying to make something that works, but immensely mr so to understand it! Otherwise there's no confidence that making a second model will work the same.

The fact that bypassing the 560Ω resistor kills the reaction, to me, says that the HT battery is 'hot' at RF. Is it possible that capacitive coupling from this to the aerial is taking place? I'd predict, that if you put a 0.1µF from HT+ to chassis, you would also find that reaction disappears.

[jim_jobe]

What wavelengths is it operating on Bob?
I built a one valver last week using a 3D6. This has the phones in series with HT but all I got was horrible noisy crackling when I operated the tuning capacitor (a 50pF air spaced). I scratched my head, as one does, and it began working after a fashion bit only when I touched the top of my balding pate! I later discovered that the tuning capacitor wasn't earthed. I had assumed that it was earthed through the spindle but it wasn't and needed an actual soldered connection. I still don't know why it worked when I scratched my head though and I don't intend repeating the experiment!
Jim

[Variometer]

Mike,
Yes, as I have said before, the pot is acting as a very good reaction control. When it is at its lowest, the volume is weak and the selectivity poor. As I increase it, the volume increase and the selectivity improves. When I increase it to a certain level, it bursts into oscillation, I then back off a bit and the set is functioning at maximum efficiency!
If I removed V2, I would not know whether it was oscillating or not unless I moved the headphones to V1.
I don't know what frequeny range it is apart from the fact that it is generally MW.

My initial post was simply wondering how it is working with such a strange reaction method and whether anyone else had come across this configuartion before.. I am now going to re-build it properly (polished wood base, brass terminals etc etc) and see if it still works. It is even an advantage to me for it to be "out of the ordinary," as I will be writing an article about it.

I don't actually want to have to put a conventional reaction circuit in it.

Bob

[Variometer]

Jim,
Reminds me of the first (non-kit) radio I ever made in 1954 when I was 10. It would only work when I had my left forefinger on valve top cap and right forefinger on tuning capacitor frame. It was built on an upturned cardboard shoe box.
Bob

[jim_jobe]

I forgot to mention Bob that it only worked with one finger on head and other hand on tuning capacitor spindle. I wasn't earthed in any way so return must have been capacitive through headphones and power supply!
Jim

[Variometer]

Kalee
My own idea is that because it is on such a small chassis, there was initially a certain amount of feedback from stray capacitances that made it work at loud volume, but not enough to make it oscillate. The introduction of the screen pot allowed me to bring it up to oscillation as in a normal reaction circuit. It is not on my work table at the moment as I am doing something else not connected with radio. I will try the capacitor later though. However, I don't want to stop the reaction working. It is working great as it is and if it still works when it is re-built, I will be very pleased.
It doesn't have an HT battery, by the way, it is a mains variable power supply (homemade), so I can have any HT voltage I want between 0 and about 300V. It is working well on 54V at the moment.
Bob

[MichaelR]

Hi Bob,

Fully understand now what you are trying to do.

My questions are really there to find out how it is oscillating because as Kalee quite rightly points out it may not be reproducable by other people.

For instance it might be a number of differing things that are causing the oscillation, wiring layout even the power supply and the particular individual valves that you are using or the combination .

Nobody is disputing that it is not oscillating.

The point is the pot is altering the gain of the pentode for sure which will obviously affect the oscillation (as you would expect it to) and you have some degree of control.The positive feedback method needed for oscillation is unknown.

Mike

[PJL]

The positive feedback is due to the 560 bias resistor in the HT- rail. This puts an RF signal relative to the LT (cathode) on the HT line. The feedback path is via the 100K anode resistor. I am then a little unclear. Varying the screen voltage will change the output impedance of V1 which would effect the feedback but I am surprised it works so well as the resonant circuit does not appear to be in the feedback path.

[Variometer]

Thanks for further replies. In the last few minutes, I dispensed with the two 47uH chokes in fuseholders and soldered two chokes (47uH each) under the chassis. This was not to improve performance, but simply to make room for me to include the matching transformer on the baseboard next to the tuning capacitor when I rebuild it.
It is still working fine. I also tried using three small 9V batteries in series for the HT and it still works perfectly with good reaction control.
When I rebuild it, I will use a completely different set of components and valves. I am still happy with my idea that stray capacitances are causing a certain amount of feedback and that the control of the screen voltage allows me to bring it to oscillation before backing off again.

Bob