Automatic Regeneration Control

[pwdrive]

Is automatic regeneration control possible ie: maintaining the regeneration on the verge of oscillation when tuning across the band?
I have seen articles where mechanical capacitive ganging as been used (mainly on old domestic TRF's) ie: regen feedback under capacitive control ganged with the RF tank circuit.
Also seen an article about a homebrew where the AF was rectified and via a valve shunt was used to control the screen grid voltage on the regen detector.
Is there an RF method?

Lawrence.

[pwdrive]

Done a bit of searching, treat the regen det as a stand alone oscillator then design for constant threshold of oscillation across a given tuning range, a modified Vackar seems to come very close.
The link below takes you to some work done by someone on this subject using a valve regen, the article is an interesting one, down towards the end of the article are two utube links worth watching about this subject, they are quite well done with no annoying music and no jerky camera.
Anyone who is familiar with using or building regens will note and hear how well this detector works with AM with the regen control set to a fixed position, I think it would be as constant for SSB/CW once just over the point of oscillation, it covers an almost 3 to 1 tuning spread which I think is quite impressive, the one shown tunes from approx 3.3 Mhz to just over 9 Mhz (I think) It looks like the Vackar is the way to go....?
Are there any examples of this approach in domestic vintage radios of the TRF/Regen period?

http://www.kearman.com/vladn/hybrid_feedback.pdf

Lawrence.

[Radio Fixer]

Seems to me you are describing super regeneration invented by Howard Armstrong. http://en.wikipedia.org/wiki/Regenerativ...e_receiver

I restored a US AM / FM set that got the FM with the use of just one extra valve. Not super quality but it worked. It used a circuit called a Fremodyne which was just the super regen principle. Here's an extract from an article I wrote:

"The Fremodyne FM Converter

The single, double triode, valve converts the FM band of 88 to 108 Mc/s to give audio output. When switched to FM this output is amplified by the AM pre-amp and the 35L6 output valve.

Wireless World, of March 1948, which included a short article on the circuit, put the functions like this:

One section of the valve works as the local oscillator necessary for superheterodyne conversion, while the other section functions as:

1) A superhetrodyne convertor to an IF of about 22 Mc/s
2) A super-regenerative I.F. amplifier of high gain.
3) An F.M. to A.M. converter and
4) A detector giving an AF output.

W.W. also said the circuit has been criticised for radiation but the Hazeleltine engineers said that it was 30-40 dB better than an ordinary super-regenerator. This would be one without superhetrodyne conversion and a separate local oscillator stage. Here, this runs at around 22 Mc/s higher than the tuned frequency and is not near the resonant frequency of the input tuned circuit. Another problem was said to be that stations could be tuned in two places, one on each side of the I.F. transformer response with both being correct.

Now from the schematic of Figure 1, how does this clever piece of circuitry actually work?

Below is an explanation ‘worked up’ by me. I don’t know whether it is totally correct?


  Fremodyne.pdf (Size: 42.02 KB / Downloads: 91)

From looking at Refs. 3 and 4, I consider the second triode (12AT7 b) to be a VHF Colpitts oscillator. The anode is at AC earth, due to capacitor C27, and so is connected to one end of the oscillator coil, L62, the other end of which is connected to the grid. Now the grid to cathode capacitance and the stray capacity from cathode to earth forms the tapped capacitors, across the coil, associated with this type of circuit. The RF choke, L67, allows DC to flow in the valve but has a high reactance so as not to effect oscillation. Ideally, the resistor R55 should have a value so as not to load unduly the ‘tank’, formed by L62, A8 and the tuning capacitor. No doubt by suitable choice of R55 and the capacitor C29 it is possible to control oscillator amplitude, in ‘grid leak’ fashion.

The first stage (12AT7 a) has to be basically an oscillator that is continually quenched (stopped) at supersonic frequency. Again it looks like a Colpitts with C31 and C32 forming the feedback capacitors.
They are also the IF transformer tuning capacitors, returned via C36. Choke L65b provides a high impeadance DC path.

Mixing takes place, between the input signal and the injected oscillator waveform, at the non-linear cathode to grid of the stage. R54 and L66 provide a high impeadance DC return path for the grid.
Capacitor C33 and resistors R48 and R49 are the quench components. This capacitor, charging via the grid cathode diode, leads to the valve to becoming biased off. If an oscilloscope is connected to C9 negative (or positive), then a sawtooth waveform is seen, of about 25 Kc/s, with audio superimposed upon it. The value of C9 is not critical, it being just an audio coupler with IF shunted away by C34.
As the frequency modulation sweeps across the tuned circuit skirt, the IF current through the valve changes and this is converted to an audio voltage change across R49. R63 and C25 provide de-emphasis.

So what of performance? AM performance is as good as any small radio of the time and is quite respectable. This can’t be said for the FM band. It can only be described as poor and the quality is worse than the cheapest solid state radio. It is possible that I am still not getting the best from the circuit; I was reluctant to start disconnecting and measuring the ceramic capacitors’ birdnested’ around this stage. However, these did look as good as new with solid encapsulation. Also, I have ‘talked’, via the Internet, with US collectors and old radio dealers. They said this performance was about all it was capable of, and some gave a lot of other more disparaging remarks as well!

It needs strong signals and careful tuning to minimise distortion and low level background whistles, to which it is prone. It does have reasonable rejection of AM noise, presumably the high gain of the super-regenerator, means it acts as a limiter. I have a fluorescent bench light that generates stacks of interference (I must get around to adding some line filters); it is so bad that I cannot have an AM radio in use in the same room. On the FM band of the Gilfillan, although it can be heard when switched on, it is at a very low level and not objectionable. So in 1948, buyers would have had a radio with which they could listen to the new FM stations, with a fair degree of AM noise rejection, for just $34. It was one of the lowest priced FM receivers. "

Gary

[pwdrive]

Hi Gary, not super regen I am thinking of as there is a problem with the quench frequency once the band being tuned falls below a certain frequency as I understand it.
The guy in that link has done a lot of work on the ideas he has come up with including all the maths, I guess the proof of the pudding is in the eating, if his utube videos are genuine (they appear to be) then he has done a good job in almost eliminating the age old problem of having to manually adjust the amount of regeneration required while tuning a typical ratio of 3 to 1 up to at least 9 Mhz, I reckon you could get it spot on in a tunable IF regen detector as the tuning ratio required is much lower.
I don't think the concept is new, there are the odd references to it somewhere from the old days but it all seems to have been forgotten or not built upon? maybe because of the extra component count/complexity etc?

Thanks for the link and explanation thereof.

Lawrence.

[Radio Fixer]

OK! Sorry about that.

Now you have given me something else to do (the list gets longer everyday!). I'll have to read and try to understand the article now.

thanks Gary

[pwdrive]

OK! Sorry about that.

Now you have given me something else to do (the list gets longer everyday!). I'll have to read and try to understand the article now.

thanks Gary

No worries Gary, I'm not very good at writing posts.

I think that guys (He's called Vlad) approach was to combine the properties of two oscillators ie: the Vackar and the Armstrong, we already know that one of the most common types of regen detector is the Armstrong type, ie: the amount of regen required for oscillation increases as the frequency being tuned decreases, I presume that the Vackar helps to compensate for this, the Vackar is relatively stable in amplitude and frequency across a wide frequency range ie: 3 to 1.
He refers to "tilt" this as far as I can understand is his word to describe the amount of feedback required in plus or minus terms around a nominal value as the frequency is increased or decreased
That's how I have read it so far, I might be wrong so any input welcome.

Lawrence.

[pwdrive]

Having just read through the paper he wrote it has become clearer now, so far as I can see his approach and methodology seems to confirm his practical results, all very interesting.

Lawrence.

[Radio Fixer]

Well, I read some of it. Skipped the math and didn't try to understand the pretty curves. Um! how do I make one for the MW band?

Think you are going to be on your own with this one Lawrence!

Gary

[pwdrive]

Well, I read some of it. Skipped the math and didn't try to understand the pretty curves. Um! how do I make one for the MW band?

Think you are going to be on your own with this one Lawrence!

Gary

Hi Gary, I vaguely got the curves ( I think) just a series of resonance plots in the usual amplitude/time domain with series, parallel and combined losses modeled.

The basic circuit seems to suggest Armstrongs feedback arrangement (tickler and tank) but the tank is a modified Clapp arrangement ie: Vackar, thus its a hybrid with what appears to be equal but opposite properties so far as loss versus frequency is concerned.
In it's basic form it shouldn't take to much to build one or modify an existing medium wave regen as the basis for the coil is already there.
If I ever get back to some experimenting I would like to give it a go.

Lawrence.

[pwdrive]

I think this pre dates Vackars paper (1949) There is a mention in this article of using a hybrid circuit in oscillators using a tickler/collpitts arrangement mentioning it's near constant amplitude when tuning. If you come across an unusual regen circuit in a receiver or Lo in a superhet then maybe it's an attempt to achieve constant amplitude....

http://www.vacuumtubeera.net/Superhetero...lators.pdf

Lawrence.