The Amīcaflex 4 transistor reflex radio

[Amie]

.... If a floating detector is used, and it only has its own leakage as a load then the actual signal output voltage would be as  high as it could possibly be. It would need an amplifier of very high input impedance to make use of it though, valve or fet for instance. Remove either of the detector diodes from my radio and I will guarantee it won't work. 

It was found that the Grundig 80U that I worked on had components missing, that meant the am detector was floating, (Lawrence I think noted), it shouldn't have worked, but it did. When I checked it with a modulated RF signal the waveforms were very distorted though.

Detector diodes in low level circuits aren't necessarily rectifying the incoming signal, they are acting as non linear mixers, so anyone claiming that the floating detector doesn't work isn't taking that into account, if you want predictable results though, it has as Peter says got to be done properly.

Thanks.
Amie.

[Geordie McBoyne]

From the January 1961 Radio Constructor magazine the design and construction of a super-regenerative transistor receiver.  This receiver will offer much better performance than the very simple receivers I've posted up in this topic.
The detector operates in the squegging principle and is explained in the text.

The white spot transistor is equivalent to the OC44. Red spot OC71.

Geordie McBoyne.

[Murphyv310]

Hi.
Well after 56 years I cobbled up the EE8 reflexed radio using one AF114 instead of the 116, it had tin whiskers but I blasted them with a charged cap at 150v. I matched all the values as per the old circuit with the exception of the HF choke which is a modern 2mh item. As I thought the resultant AF waweform is horrible but in fact that little one transistor circuit drove a pair of high impedance phones quite easily and surprisingly sounded quite good. Again I tried different combinations of diodes and got an improvement in output by fitting another diode with its cathode connected to the anode of the existing diode and the new diodes anode grounded. There was a noticeable improvement in audio.
Another improvement which didn't reduce top end audio noticeably was to increase the 47nf on the cathode of the detector Diode RF filter, that also cleaned up some RF leakage but that could be due to rubbish modem chokes with high capacitance. Getting good RFCs these days is like looking for a needle in a haystack.

[Amie]

From the January 1961 Radio Constructor magazine the design and construction of a super-regenerative transistor receiver.  This receiver will offer much better performance....
Geordie McBoyne.

Thanks for the interesting circuits.
I would be interested to know if you've ever built this circuit? For it to be efficient, the quench could only be a very low frequency on the MW band. For audio, some recent experiments I was doing with help from Kalee 20, on band 2 VHF with SR detection, a quench frequency of about 30KHz was possible using a triode valve as a detector.

Thanks
Amie.

[Kalee20]

.... If a floating detector is used, and it only has its own leakage as a load then the actual signal output voltage would be as  high as it could possibly be.

That's (debatably) true, though there are complications!

If the leakage resistance is super-high, then the detector output load capacitor will really charge right up to the signal peak, with virtually no load. And as the AF signal increases to the modulation peak, the detector output will follow the modulation.

But as the AF signal cycle passes its peak and heads downwards, with the AM signal heading towards the modulation trough, the detector load capacitance will stay charged because the leakage resistance is so high. If the AF is at a low frequency, it might discharge so that the detector output can follow the modulation envelope. But if the AF is at a high frequency, it very likely won't, so not only will the AF output drop, it'll also be distorted as the rising portion of the AF cycle will be correct but the falling bit will just be a segment of an exponential discharge curve. Which could well be the distortion you mentioned.

So under these conditions, reducing the load will actually give an INCREASE in detector AF signal output voltage.

[Amie]

Hi.
Well after 56 years I cobbled up the EE8 reflexed radio using one AF114 instead of the 116, it had tin whiskers but I blasted them with a charged cap at 150v. I matched all the values as per the old circuit with the exception of the HF choke which is a modern 2mh item. As I thought the resultant AF waweform is horrible but in fact that little one transistor circuit drove a pair of high impedance phones quite easily and surprisingly sounded quite good. Again I tried different combinations of diodes and got an improvement in output by fitting another diode with its cathode connected to the anode of the existing diode and the new diodes anode grounded. There was a noticeable improvement in audio.
Another improvement which didn't reduce top end audio noticeably was to increase the 47nf on the cathode of the detector Diode RF filter, that also cleaned up some RF leakage but that could be due to rubbish modem chokes with high capacitance. Getting good RFCs these days is like looking for a needle in a haystack.

By adding the additional diode you've made the circuit as per mine basically. The transistor must be biased better, and the AGC action will be feature now too.

+ you shaved another AF 1xx transistor

The old fashioned multi section pi wound chokes are the best ive found, bit like unicorn poo if you're looking for one though.

Amie.

[Amie]

.... If a floating detector is used, and it only has its own leakage as a load then the actual signal output voltage would be as  high as it could possibly be.

That's (debatably) true, though there are complications!

So under these conditions, reducing the load will actually give an INCREASE in detector AF signal output voltage.

That's quite so! Its as in the world of crystal radios, efficiency is in the matching of the load...


Amie.

[Kalee20]

Yep! Of course, the heavier loading, while improving the high AF situation, will give a reduction in the low AF performance, till the two become equal... And selectivity will be worse.

Loving the new avatar, BTW!

[Murphyv310]

Ah crystal sets.
Have you ever tried the ALD chips?
There's a few circuits online......... but!! There is a trick you can do with them to really boost the output on strong signals. Basically you use the resultant DC voltage from detection to power the chip. This can double the output. In fact with just 0.5v (fed externally) they can power a speaker, audible but not loud with a short aerial. My frame aerial and the ALD chip is impressive.

[Murphyv310]

Hi.
The method of obtaining bias in some older circuits with germanium transistors as in the Philips kits was quite common. We are aware of the shortfalls of not using a potential divider but in fact this simple method has two advantages.
Number 1 it introduces a little NFB which can be useful in fact the negative feedback leads to number 2 which gives a small degree of safety from thermal runaway, if this occurs the collector volts reduce which is reflected to the base thus reducing its voltage and then the transistors collector current. As can be seen on the output transistor it's only 27k, and in this case the AC126 ran slightly warm or quite hot if you forgot the heat sink, an emitter resistor though and a potential divider bias arrangement would have been the best route of course giving a better overall result. No doubt the main reason would be to make the kits as simple as possible.