Home Made TRF

[Skywave]

Originally Posted by Neil Purling

I changed the circuit of the detector by utilising a new Jackson Type 0 365pf tuning capacitor, rather than the previous 500pf one.
Reaction is applied to the second grid of the EF39.
Could the change to the tuned circuit mentioned above change the way the circuit behaves in respect of the amount of reaction required?

Obviously, you have reduced the tuning range by reducing the total amount of variable C in the signal-selection part of the cct. However, as regards any impact on the reaction control setting and its behaviour, I would expect only a very small noticable difference, if any.

[Neil Purling]

Max capacitance equals the longest wave-lengths, and there isn't anything locally at the top of the band worth bothering with.
Anyone got any informed technical opinion on the effect of the change I have made on reaction (if any)?

[Neil Purling]

I did not intend to hurt the feelings of G8DLH by my comment above.
I did not mean to imply anything other than a desire to understand the effects of making changes to the basic circuit I used for the detector.

[Neil Purling]

I have had a distinct lack of volume since I rebuilt the radio and I am wondering what it is wrong.
I have looked at the anode voltage on V1 (EF39) and it is now 270V. It was only 78V before. I assume that therefore the valve isn't drawing any current.
I don't know a great deal apart from that.
What should I look for?

[Biggles]

I think you're probably right in assuming that the valve isn't drawing current. If it has a cathode resistor check for any voltage developed across it. Also check the screen grid (g2) voltage is present. If this is missing then the valve will only draw a small amount of anode current. First port of call would be to check all connections particularly cathode connection to chassis or via a resistor. The valve socket may have bad internal connections. If all that is ok, then you may have a faulty valve. (Check also that the heater is lit!)

Biggles.

[Neil Purling]

When I rebuilt the thing I shifted the position of the tuning coil. i also changed the tuning capacitor from 500pf in the diagram to 365pf.
surely this wouldnt make any difference beyond changing the tuning range and perhaps the selectivity.
Reaction is applied by varying the screen grid voltage. The thing seems to oscillate much easier than before. The reaction pot is the same as before.
I wish I had taken more voltage measurements before, particularly the voltage swing on the screen grid. the audio o/p seems more muffled in its present state.
When it is working it is a very good radio, considering I built it.

[Skywave]

Hi Neil.

Re: your cct. diag. A few observations / comments that may help.

(a) I am puzzled by the anode load of the EF39: 220k plus 330R. Should there be a decoupling cap. here at the junction? Or?
(b) I think that you will get better quality audio if you can arrange for a RF filter between the anode and the top of the AF gain control.
(c) Presumably, the mains transformer sec'y. has a centre tap to gnd. ?
(d) If you are still troubled by hum, replace the 1k5 resistor in the PSU filter by something like a 5H choke - or more.
Finally:
(e) What is the value of the cap. across the pri. of the O/P transformer? (I'd guess at something like 0.005 uF)

BTW I'm sure that (a) and (c) are draughting errors . . .

Hope some of that helps.

Al.

[Neil Purling]

The circuit digram ought to have a 250-0-250 HT winding.
The detector is exactly as in this sketch:
https://www.vintage-radio.net/forum/...8&d=1141415399
It was contributed by member heatercathodeshort in a thread on a home built TRF made by member Hunts Smoothing Bomb.
This thing worked fine before. What sound does come through is very low and somewhat muffled. It is not surprising because it isn't presently drawing much current.

[Skywave]

Originally Posted by Neil Purling

The circuit digram ought to have a 250-0-250 HT winding.
The detector is exactly as in this sketch:
https://www.vintage-radio.net/forum/...8&d=1141415399

. . . . which, if I may respectfully point out, is not quite the same as the schematic you submitted earlier, re: the 330R resistor in series with the 220K resistor in the EF39 anode feed . . .

As for this valve (possibly) not drawing any current, let us know what the voltages are at the anode, screen grid and cathode, please. This will help us to start the diagnosis.

Al.

[Neil Purling]

Point taken about the 330R restsitor, it doesnt make much, if any difference.
On the EF39:
The anode voltage varies from 260V-100V as the reaction control is advanced.
The screen grid voltage swings from 0-40V as reaction is advanced.
Since the cathode is effectively connected to chassis (through the coil) it is at 0V.

Voltages measured under no signal conditions with AVO 8 Mk2.

[Neil Breward]

Originally Posted by Neil Purling

Point taken about the 330R restsitor, it doesnt make much, if any difference.
On the EF39:
The anode voltage varies from 260V-100V as the reaction control is advanced.
The screen grid voltage swings from 0-40V as reaction is advanced.
Since the cathode is effectively connected to chassis (through the coil) it is at 0V.

Voltages measured under no signal conditions with AVO 8 Mk2.

Neil - A couple of suggestions to improve the stability of your circuit:

1. The anode of the 6J5 Af amplifier needs more decoupling to remove any hum and possible RF feedback. Connect a 22k resistor between HT and the 220k load resistor and decouple the junction of these two resistors to chassis with a capacitor - a ten microfarad 350 v electrolytic would be OK.

2. Lose your 330 ohm resistor in the EF39 HT circuit. Replace it with a 2.2k item and decouple the junction of these two to chassis with a 0.1 mfd cap.

3. Connect a ceramic or mica capacitor of about 100 pF between the EF39 anode and chassis. This will act both as an RF filter and make reaction more reliable and increase the gain as it will reduce the Miller feedback at RF through the valve's internal capacitance. Not huge in a pentode, but can still cause problems.

Good luck and keep the soldering iron warm!

Cheers,
Neil

[Neil Purling]

If anyone knows about the working of the electron-coupled detector, as used here please enlighten me.
Are they especially dependant on the coil being precisely arranged. I am thinking of the bottom of the coil, the first five turns. The bottom of the coil goes to chassis of course. The five turn tap is for the cathode. There isn,t much tension on those five turns and I wondered if they get out of place if this is going to totally muck up the operation of the receiver. The only things disturbed were the coil and tuning capacitor.
As member heaterchathodeshort presented this circuit to us, perhaps he (or someone else) can tell me what is important to have right for proper operation.

I am thinking of re-winding the coil. I have already changed the valve base and tried substituting another EF39. I even subbed the 6J5 in case it was giving trouble. Of course I now know from my probing around the EF39 that the problem seems to be around there.

[Skywave]

Just a few thoughts, Neil.

Since this problem has arisen since you changed the tuning cap., could it be faulty? Rare, I know, but if you could temporarily put back the original 500 pF variable, there might be a surprising result.

Secondly, the connections to variable capacitors are sometimes a little confusing. With all due respect to you, I just wonder if you have accidentally wired the 365 pF cap. into the cct. incorrectly - I know, I have done sillies like this myself!

Thirdly, if you move the tap on the coil further up the coil from ground, the valve will require less reaction to make it oscillate. Doing this could throw some light on the problem.

Sometimes, with faults like this, it helps to think laterally.

Just a few ideas.

Al.

[Neil Purling]

Al, why is the EF39 also not drawing any current?
If the bottom of the coil incl tap no longer lie parallel to each other this will mess up reaction, yes?
The new variable capacitor was wired with the moving vanes to chassis & the fixed to the top of the coil, just like the last one.

[Skywave]

Hi.

Based on your reported measurements, it would seem that the valve is drawing current. As the screen grid volts are varied, the voltage at the anode changes. This change can only take place if there is a changing voltage across the 220K anode load. And this can only take place if there is a changing current through it - coming from the valve anode.

I think that the problem you have is that the valve is not getting to the self-oscillation point . . . which leads to the next para. . . . .

I don't quite understand your comment about the coil being parallel. The whole coil, including the tapped portion, needs to be co-axial in order to produce the mutual inductance between the section of coil above the tap and the section below the tap. That way the required feedback is produced by this mutual coupling. If there is insufficient feedback, you will not be able to get the set to oscillate - or more precisely, to the point just before oscillation begins - when the sensitivity of the detector is at its greatest. The threshold point at where the oscillation begins is determined by the voltage on the screen grid in this type of design of TRF.

BTW Some designs include a variable capacitor in series with a seperate feedback coil - this capacitor controlling the amount of feedback - instead of the method in your cct. I'm not aware of any significant advantages of each method over the other - but there may be. Perhaps others could develop this in a seperate Thread.

Al.

[Neil Purling]

I think heatercathodeshort presented this design to the forum in post 23 of this thread:
https://www.vintage-radio.net/forum/...home+built+TRF

It appealed to me because it was something that could be created eaily out of junk, did not use a commercial coil. Most importantly it would work, even if you were the sort of person who doesn't have much luck with building stuff.

[Neil Purling]

The amount of inductance (L)in the tuned circuit never changed, but capacitance (C) did.
The change of variable capacitor plus I moved the coil so I could shorten the wires from the coil to the rest of the circuit.
I stripped out the de-coupling I have already fitted, so it is like the original diagrams.
I have exhanged PM's with heatercathodeshort, thanks mate.
I experimented with the original design of coil where a 100 turn aerial coupling winding fed the signal to the detector. I wound it on a bog roll tube so it would slip over the detector coil. It produced some loud signals, but loaded the detector so I couldn't tune the bottom of medium wave.
My exchange of PM's produced the advice to reduce the number of turns of the aerial coil to fifty because it was placed directly over the detector coil.
For your interest:A aerial coupling coil of 100T is more appropriate when the coil is wound next to the detector winding.
So, it is running properly. I get my BBC local on 1485 Khz, Virgin 1215,Magic 1161, Talk Sport 1089 and, finally BBC Radio Five Live on 909Khz. I think that slow motion (36:1) dial definitely helps.

[Neil Purling]

I had trouble with the valve base. I changed the base.
I experimented by eliminating the 6J5 stage entirely. Obviously the output volume was greatly reduced.
I was thinking about re-making the receiver with all the suggested modifications of this thread.

I have several EF36's. Maybe I could use two. One as detector & the other as AF pre-amp. It's just a idea.
I want to improve stability of the whole set and add a 'Gram' input.
Should I therefore have multiple switch wafers: One for the AF, one to switch the detector's heater off and maybe a third to cut the HT to the detector.

[Neil Purling]

Post #31 contains some useful advice I intend to follow. However the only capacitor I have of around that value is 120, not 100pf as reccomended.
The capacitor was to act as a RF filter, stabalise reaction and something was said about Miller feedback I don't entirely understand. Some designs have a choke on the anode of the detector to block RF from the AF stages.
As I have not proceeded beyond the PSU so far I would like members input on what sort of value of choke one should have, if one is desirable. From looking at some other circuits it seems the choke should be placed between the anode resistor and the valve base.
The anode pin of the valve the choke and the coupling capacitor to the AF pre-amp.
It was suggested I change the value from 0.01uf to something else to change the frequency response of the radio, but I can't remember what value was suggested.

[quantum]

A 120pf cap will do just as well as a 100pf cap - the value isn't that critical. Miller feedback refers to possible internal feedback within the valve itself which can cause instability, some older designs do use a choke in the anode load of the detector stage to prevent RF from getting into the HT but that wouldn't prevent the Miller effect. Personally, I've found that if a detector is properly decoupled a AF choke shouldn't be necessary.

You could probably increase the AF coupling cap up to 0.05uf without problems, any larger value might cause 'grid blocking' unless the value of the grid leak of the following AF amplifier was 470k or less. The larger values increase bass response, but I'd be surprised if you'd notice any real difference above 0.02uf.