Home Made TRF

[Neil Purling]

If you look at the attached images you will see the circuit diagrams I have married to produce my home made TRF.
The detector diagram was contributed by heatercathodeshort. Thank you very much sir. If I can make it work, so can you!!!
I have tried it with a EF91, but here the valve with metalised coating is a CV1053 (EF39). The black metal valves are a 6J5 Triode and the 6V6 Tetrode, from fleabay and Colomor Electronics respectively.
The whole shebang is powered by a transformer courtesy of settsappeal and the vintage RS output transformer is from the same forum member too.
I was going to use a EZ90 rectifier but m1ecy had some 6X5s.
The appearance is of a parts bin mongrel, which is true.

So to the £1M question: How does it work? Well, with the triode boosting the signal the output starts to clip after about 30% volume with a strong local signal. It is flipping loud!

Many moons ago there was another thread about a home made TRF started by Hunts Smoothing Bomb. That inspired me to have a try. I have picked HSB's brain and that of heatercathodeshort. Their advice has been invaluable and I apologise if my PMs were asking so many questions.
The audio stage was culled from my 1952 edition of Radio by Bradley. It is the AF stage of the AC Mains superhet. I amended the design to use the 6J5. Mr Bradley, wherever you are, thanks to you too!

Now all I need is to encase it. I was at woodwork at school.

[quantum]

How's the selectivity? I remember when I used to make TRFs with a single tuned circuit, the BBC still used the medium wave frequencies for Radios 1, 3 and 4 and a single tuned circuit had difficulty in getting good separation between Radios 1 and 4. I found that I had to used a tuned RF stage as well, and an infinite impedence detector to ensure that at least one of the tuned circuits tuned sharply.

If you've got plenty of gain to spare, you could experiment with negative feedback to improve audio quality at the expense of gain. The simplest method would be to connect a resistor between the anodes of your 6J5 and 6V6, anything between 1 meg and 2.2 meg start with, or a feedback circuit from the secondary of the output tranny and the cathode of the 6J5 stage.

[Neil Purling]

When I added a longer length of wire to my TRF radio to boost the signal strength I noted that Magic 1161 and Talk Sport 1089 Khz were the two shortest wavelength stations. I appreciate that I gave the stations frequency but their Web sites don't give wavelength.
If I remove all aerial wires I get Radio Humberside (1485 Khz/202 metres).
If I connect as little as 2ft of wire magic 1161 appears in place of the afore mentioned local radio with no change in setting of tuning capacitor or reaction pot.
At present there is no coupling capacitor in the feed from the aerial to the coil.
This would normally be about 100pf and I am looking to source a 100pf variable capacitor.
Please tell me why this is happening.

[mickjjo]

Hi Neil, The effects you describe are quite normal when there is only one tuned circuit. The aerial wire acts as one plate of a capacitor and the ground as the other, so varying the length of the aerial will affect the tuning. The effect can be reduced with a small capacitor or trimmer in series with the aerial, or by using a smaller coupling coil next to the main tuned winding for the aerial connection. Hope this helps a bit .

Regards, Mick.

[Hunts smoothing bomb]

Hi Neil,

I have read in a book or two that changing the legth of the aerial wire on these primitave recievers has exactly the same effect as tuning the variable condenser.
I think what happens is the aerial will be more sensitive to certain frequencies depending on it's length, e.g. If the aerial is 2 metres long then it will be more sensitive to wavelengths (in metres) divisable by 2.
It will also affect the tuned frequency because it is like changing the inductance....... Ithink?
I do not understand why it does not affect the tuning of a domestic comercially built set though, perhaps because there are many tuned circuits?

I'm not totally sure that what I have said above is correct but, I think it is along the right lines. I'm sure a knowledgable sole out there will put me right

Regards
Lee

[Neil Purling]

Is the presence (or lack of) a capacitor between the aerial and its winding on the tuning coil a factor?

[Station X]

The length of aerial makes a difference to the tuning because it possesses capacitance to earth and inductance. This is particularly noticeable with some crystal sets which rely on the capacitance of the aerial to tune the set and will not work without a good earth connection. Some comms receivers have an aerial tuning control (antenna trim) to adjust the tuning to suit a particular aerial and frequency. There is a noticeable peak as this is adjusted. The more wire you put up the more signal you will receive.

Addressing Lee's point A resonant aerial of say a quarter wave length has an impedance of about 37 ohms which is not too bad a match for a receiver with a 50 ohm input. However a quarter wave resonant aerial for use on MW would need to be 50 metres long to resonate at 1.5 MHz (200 metres) and 150 metres long to resonate at 500 KHz (600 metres).

A quarter wave aerial 2 metres long is resonant at 37.5 MHz or 8 metres.

You can use an Antenna Matching Unit (AMU) sometimes miscalled an Antenna Tuning Unit (ATU) to fool the receiver into thinking that a short aerial looks like 50 ohms, But this is hardly worthwhile with a TRF receiver and contrary to popular belief it won't improve the aerial.

[newlite4]

I have been looking for a TRF circuit like this for some time. A year ago I bought a cabinet for a 1933 Halston (USA) radio at the car boot for 50p and ever since have been looking to build a replacement chassis for it. I did download a chassis layout and schematic for the Halston from the Nostalgia Air site but felt that the cost of 'tubes' could mount up. This could be the answer.
Neil

[Neil Purling]

I think that with the 100 turns aerial winding there was too much inductance.
In the attached image of my amendment I wound 2 turns of insulated wire around the aerial coil, connecting one end to the aerial and the other to chassis.
Without the aerial connected I tuned what I thought to be my local BBC station, (hard to tell at nine pm). However it was in its usual place, just after the vanes of the tuning capacitor started to mesh. Connecting the aerial produced a large jump in signal strength of course and NO frequency shift.

[Neil Purling]

After trying the set on Sunday morning I got the local gardening programme.
However there was not much else in evidence. I need more inductive coupling to get the other medium wave stations during the day.
I think that if I wound a aerial input coil of say fifty turns on a bog roll tube, then the usual 65 turns,tapped at 5 for the detector circuit on another tube that is slightly larger. I can thus slide one inside the other to get the ideal balance and wedge the thing with slivers of cardboard to hold the coils in position.
What better excuse do I need to have another play?
What do you think?

[Heatercathodeshort]

Hello Neil and well done with your project. Its fantastic to get a signal from something you have built and by the sound of it you have got it about right. Thanks for your comments by the way and feel free to pm me if you desire.
Regards. John.

[Neil Purling]

I now have proper reception. The stations are in their appropriate place and my local at 1485 Khz is present as is Magic 1161 Talk Sport and I think what is Virgin radio just below Magic 1161. The two turns have been disconnected from the chassis and thus are a loose coupler.
It has been a educational experience.

Some masochistic impulse tempts me to repeat the experience and build another to give short wave......
Anybody have any ideas about modification of the coil?

[Ed_Dinning]

Hi Neil, for short waves the coil should be self supporting and made of thick wire 16/18SWG. Reduce L & C pro rata, you may need to reduce the coupling windings even further. You seem to be getting thehang of experimenting, have fun.

Ed

[Neil Purling]

I thoroughly enjoyed myself with this little TRF. When I got one of those kit radios it got bunged in the loft.
As the Easter weather has been so vile I dug it out again.
It has been through various incarnations. I wound a new coil on a paxolin former. The aerial coupling winding was on the outside wound around a bog roll core. Best results came after I reduced the winding down from 100 to 50 turns.

I did have a hum problem for which I seperated the mains transformer from the small signal stages & added a choke and additional smoothing. Not much difference.
It got shoved in the loft when I acquired my 3 valve TRF kit radio.

As the Easter 2008 weather was so vile I attacked it with pliers & soldering iron, removing the 6J5 stage entirely and driving the 6V6 directly from the detector.
With the wire I have got in the loft a signal input I suppose is about 20mV. It is hard to say exactly if you are looking at the raw carrier of a off-air AM signal.

The only test I could do was to inject a signal from my signal generator into the aerial socket and monitor the signal on the grid of the 6V6. The injected signal was set to 20Mv p-p after I let the generator warm up.
The detector circuit is with post #1.
Some of my valves were old stock, others of uncertain condition.
The attached oscilloscope trace photographs show the signals obtained with the various valves I had.
It is strange that a fresh out the cardboard wrapper metal 6K7 give joint strongest signal with a Hytron VR53. I also have a un-screened 6K7G in it's original box, but it must have been used or the internal construction differs.
The screen grid controls reaction via a voltage fed from a potentiometer.
This unscreened valve needed the voltage very low to keep it out of oscillation.
I looked at a Rogers VR56. It looked to be NOS. However I only got a 600mV p-p signal. Maybe the operating conditions do not suit this type?
The VR53 and CV1053 is a EF39? The only known new example was a CV1053, giving 1.2V p-p, half the o/p of the Hytron equivalent.