Strange Reaction

[Variometer]

I recently decided to build a two valve TRF using a DAF96 as conventional grid leak detector, resistance capacity coupled to a DL96 outptut valve. Built it on small chassis about 3.7 inches by 1.9 inches. Coupled it up with 54V HT and although it worked, the reaction control (coil, capacitor and potentiometer) had no effect at all. Tried reversing reaction coil connections and no change!

What surprised me was the deafening volume it produced on headphones. After a lot of messing about, I removed reaction circuit altogether and it continued to work without any change. The HT came from my completely variable power unit (0 - 300 Volts DC). I turned the HT volts down and as I did so, it got louder, before finally starting to decrease. I turned it up again and when I got to about 60V, the whole thing burst into oscillation. The HT control appeared to be acting as reaction. I wondered if the feedback was coming from stray capacitances in the wiring. I then removed the screen dropper from the DAF96 and connected a 100K pot across the HT with the slider going to the screen. I then connected it up again with 54 Volts HT and the screen pot then gave me a perfectly normal reaction control, but without any extra components. The HT current varies from about 2.8mA to 3mA, depending on the setting of the pot.

It is all roughly built at the moment and I am wondering if this strange reaction will still be there when I rebuild it neatly.

Do you think I may be correct about the stray capacitances?

Bob

[Variometer]

This is a picture of it. Very rough at the moment, but works beautifully. I always do it this way, fumble about until I get it going and then re-build it properly on polished baseboard, and then try to figure out how it works! As you can see, it started off much bigger, but I have compressed it down to its final size by moving everything closer together.
Bob

[MichaelR]

Hi ,

Good way to learn

Might be easier if you could roughly draw out the circuit

regards
Mike

[Variometer]

It is nothing out of the ordinary, apart from the reaction control. It even works better without an earth!
Bob

[MichaelR]

Hi Bob,

First glance... do not like the headphones connected directly into the HT line...accident waiting to happen springs to mind.

Are the LT-/HT- floating or connected together

What is the value of the inductor/capacitor in the tuned circuit

Is the frequency of oscillation about 10 to 15khz?

Mike

[paulsherwin]

Originally Posted by MichaelR

First glance... do not like the headphones connected directly into the HT line...accident waiting to happen springs to mind.

While this is true, it was standard practice in the 20s and 30s.

Paul

[mickjjo]

It may be worth adding a bit of decoupling, some ideas attached.

Regards, Mick.

[Variometer]

Thanks for replies. Headphones in HT no particular problem because it is only intended for 54V HT. This is only the rough build anyway. The final build will have a matching transformer on the baseboard and low impednace headphones will plug into socket connected to secondry winding.
The two tuning coils are each 47uH chokes, The tuning capacitor is about 500pf.

Don't understand question about whether LT- and Ht- are connected together. Circuit shows there is a 560 ohm resistor between them. HT current has to flow through this to get through valve. Voltage drop across it is applied to grid via 3.9M resistor thus providing grid bias.

The circuit as shown works perfectly. The pot behaves as a normal reaction control would do. Voulme increases until it oscillates, then back off a bit.

If I put decoupling in as suggested, the pot may not provide reaction any more and I would have to replace conventional reaction circuit.

I have never come across an arrangement like this before and it was discovered by accident. I do like it though and if it works when I re-build it neatly, I may use it in future circuits!

It certainly makes a good source of discussion though.

Bob

[Variometer]

Ref headphones in HT circuit. The above circuit was drawn out in haste and is not exactly the way it is. Where I have shown the headphones, there is acrually a 1/4 inch jack socket on the prototype, see picture above. I made a small matching unit some time ago from an old 6 Volts tape recorder PSU. This is plugged into the 1/4 inch jack socket and my normal low impedance headhones plug into small 3.5mm socket in matcher. here is the picture of it. So no danger of blowing my head off even in the prototype.
Bob

[cmjones01]

A fair few years ago I built a TRF short-wave set with a reaction control which worked on the screen grid of the detector valve (an EF91). It did have reaction windings on the coils, though, so it was at least supposed to work that way! The reaction seemed to work just as well as the usual variable capacitor from the anode to reaction winding.

I still have the set in a box at the back of the workshop. I'll dare to switch it on again one day, though the extremely elderly TCC smoothing electrolytics might need a bit of TLC...

[MichaelR]

Hi Bob,

The reason I was asking the questions was because the grid leak detector and the circuit shown can also act as a grid stabilisation circuit needed for an oscillator to work.

This seems to be shown as you vary the HT voltage , suddenly there is a point whereby bias conditions are such that the first valve oscillates, with the values shown on the grid leak it would sustain about 1Mhz

If you carefully adjust the HT or the reaction control can you get squegging?

My question about the HT and LT in my previous post was realy to ascertain the LT and HT are floating from each other i.e LT- not tied to HT-

Mike

[Variometer]

Hi Mike,
It doesn't squegg! I feel pretty sure that the feedback is coming from the fact that the circuit is quite physically cramped under the chassis. Altering the screen voltage reduces the anode current of V1 to a level where it stops oscillating. The LT- and HT- are separated by the resistor soley to produce grid bias for V2.
Bob

[jimmc101]

I suspect the effect may be caused by RF current flowing in the DL96 and hence 560 ohm resistor.
Assuming that HT- is at RF earth, this will place RF on the filament of the DAF96.
Add in the capacity of the aerial to earth and the RF appears across the input tuned circuit giving a path for the feedback.
This feedback may not be very well controlled.

Decoupling caps on HT+, DAF96 screen and DL96 cathode and anode (RF only) all returned to HT- should remove the instability and allow more controlled reaction to be applied via the usual feedback winding.

Jim

[MichaelR]

Ok Bob,

Why not decouple the Ht +/- and LT +/- lines to earth and see how it operates. Maybe I am missing something here but i would have thought there is feedback possible through heaters as the circuit is shown.


Mike

[PJL]

You need to decouple the 560 ohms bias in the HT- feed.

[Oldtimer74]

I am puzzled about the use of 47 microhenry chokes as tuning coils. Surely these will be low Q? Are they the objects which appear to be in glass tubes parallel to each other? They would seem to be wound on bits of ferrite rod as they are so small. Can you get several stations on this set or only the local one?
Pat G3IKR

[murphyv310]

Hi.
You have forgotten an RF choke on the anode of the detector, it should be in between the anode & the 100k, audio should be tapped off at the junction of the choke and the 100k, you should decouple it to HT- with a 500pf-1000pf, your reaction can them be refitted and should work.
Cheers
Trevor

[Variometer]

Thanks for all the replies. I will try a bit of decoupling before I complete the final build.

Pat.
I don't particularly like winding coils for MW sets, so I use small RF chokes that cost about 30P each. I put them inside old cartridge fuses so I can change them easily. One-upon-a-time, I knew about things like Q, but over the years, I discarded a lot of accepted theory in exchange for things that worked, hence my current procedure of making things work and then figuring out why. The set gets lots of stations, the local ones being at speaker strength, but have just tried it a few mintes ago (0800) and got a few continental ones as well. It works better with no earth!

Trevor.
I didn't forget the choke, it was there in the original circuit, but subsequently discovered that it didn't make the slightest difference whether it was there or not!

Jim.
Initially there was a screen decoupler on the DAF96 and the set operated with a loud buzz, rather like a mains hum. I clipped it out whilst the set was operating, intending to hold a different value across it until I found a satisfactory one. The moment I clipped it out, the buzz disappeared, so it didn't go back!

Summary.
I really like to experiment and get more satisfaction from "messing about" than trying to understand conventional theory (that I never liked anyway).
I appreciate all your replies and will try a number of the suggestions. I can actually follow what you are all saying. I qualified as ship's radio officer in 1960 (the days of valves) and remained in that capacity until late 1992 when I quit at the age of 48 because the modern technology was wearing me down.
I am strictly a "dinasaur" these days as far as radio is concerned.
Bob

[MichaelR]

Hi Bob,

I totally agree with your approach to "seeing what works" and then figuring out why. Steve Webb who designed the cobwebb antenna is a very talented engineer and this has always been his approach in developing his understanding.

I am intrigued if you would not mind just decoupling the LT and HT connections as I am sure the feedback is through the heater filaments.

This may be why attaching an earth to the LT- is reducing your volume , it is reducing the sensitivity ( reaction through the filaments).

One step at a time works for me.

Mike

[Variometer]

Hi Mike,
In what way should I decouple them? not really sure what you mean. What value of capacitors and where? Do you mean a capacitor across the 560 ohm resistor? I can do this easily enough as the aluminium chassis wires are long enough for me to turn it over. Off shopping at the moment, but I can do it when I get back (about 2 hours).
Bob