Problem with Tuned Loop Aerial & Regeneration

[neon indicator]

I have been experimenting with standalone loop aerial regenerative radio as many people now are not in situation to add an external long wire aerial and an earth spike.

There are other advantages:
1) A 41cm diagonal/29cm square loop actually is more sensitive than a high quality 7" ferrite rod for MW & LW.
2) It's portable. A long wire isn't!
3) Large good quality Ferrite Rods are difficult & expensive to obtain.
4) LW is a manageable number of turns (< 60)
5) Looks vintage!

But I found that though the wound coil when tested was close to theory inductance, when connected to the radio the frequency was far too low. Either something dramatically adding inductance or capacitance.

Obviously the coil has some self capacitance. But I've not got a formula or "calculator" for that.

The culprit is the "tickler" coil. When it's in circuit the inductance is increased. I experimented both with capacitor and different circuits of potentiometer adjusted positive feedback for the "regeneration". A pot actually works best with "top" capacitively coupled to valve and "bottom" to Earth / 0V. The "tickler" or feedback coil is 5 turns. Adjusting the "feedback" definitely "pulls" the tuning off frequency.

Is the solution more separation between "tickler" coil and tuning / aerial coil?

Any articles anywhere on this?

Obviously if there is a diode detector added (there is none), then the feedback is not turned as high for "self mixing" or "homodyne" detection and the tuning would not be pulled. But then performance would likely be poor.

Any thoughts?


The schematic is earlier version using external aerial and ferrite rod. The frame-loop-aerial-coil version is similar, but no separate aerial and earth.

I've tried every imaginable wiring of pot / coil / valve. 10K to 200K pots.

Anode connection is MUCH poorer performance than screen-grid. Best "pot" value is 20K or 50K (22K or 47K).

[Herald1360]

I wonder if you would be better off with a fixed degree of feedback to the tickler via a capacitor from the anode with the other end of the tickler grounded, then use the pot to vary the screen grid voltage to control gain and hence the level of regeneration? It's a more conventional approach anyway.

Alternatively, perhaps try screening the tickler from the main coil by adding a shield- coax braid with a small gap to break the shorted turn around the tickler and connected to ground. This will add stray capacitance to both coils but this should be fixed so far as the aerial coil is concerned, so you should be able to compensate for it.

Neither of these suggestions is based on a thorough understanding of your particular circuit, but they are different from what you've already tried.....

[neon indicator]

The idea of fixed feedback at Screen or Anode and varying gain somewhere else is something to try.

Thanks.

My h'penny worth, is the reaction coil wound next to the earthed end of the aerial coil? And the 10pF trimmer in parallel with 100pF seems a bit small to cover a band, usually a varialble cap of some 300 to 500pF is used.

[neon indicator]

Good question. Not sure which end is beside reaction coil, I'll check.


The Schematic isn't quite reality. The variable cap is 12pF to 385pF. There is no trimmer. Nor is there a parallel capacitor.

[neon indicator]

The reaction coil (sod's law) was beside "live" end of the tuning coil. I've reversed connections on BOTH coils, making phase relationship the same, but reaction / feedback coil is beside "earthy" end of tuning coil.

The Reaction / Regen control is ever so much smoother and doesn't "pull" the tuning as much.

It doesn't work so well with feedback at the anode and varying screen grid (in fact hard to get it to oscillate!). These Russian valves are not traditional pentodes and I did notice in testing that varying Screen voltage by changing resistor affects lower frequencies more than high frequencies! I must take some photos of square wave drive to illustrate.


Why didn't I think of make the feedback coil be at "earthy end"?

Thanks merlinmaxwell

[neon indicator]

It's working very nice now with 3 x 1j18b as well as the 1j18b RF part.

It does work "faintly" direct with Maplin's "crystal earpiece" (which internally looks more like a ceramic "beeper" than 1960s type. But with the 1st of the 1j18b biased as "detector" it's more stable than "homodyne" point of regen/feedback and more sensitive than bias it as AF only and using 1N60 germanium diodes. So no diodes.

It's actually better with 4 x PP3 "on top" of the 3 x AA Alkaline than with 6 x PP3 Alkaline. 700uA vs 2mA

Yes, only 3 x AA Alkaline to run four filaments in series, 21.5mA LT 1.1V per filament. Fresh AA Alkaline measured 1.59V with 22.5mA load

I will tidy up schematic and publish final parts list and schematic.

LW & MW with one double wound coil switched parallel & series.

[Norm_N_Tam]

I look forward to seeing your final schematic! For now, here are a couple of Tuned Loop Aerial radios that use subminiature valves. I hope you will enjoy looking these over as well:

http://www.makearadio.com/tube/russian-loop-radio.php

http://www.makearadio.com/tube/loopradio.php

Norman
KA1GUK

[neon indicator]

Schematic attached. May be too shrunken

Full size here http://www.techtir.ie/construction/herge-wiring with write-up. (in progress).

Thanks Norm. Actually one of those is new to me.

While the feedback arrangement I have isn't common, it is on some old radios (or very similar).

If using it as a single valve radio the regeneration pot must be adjusted to "just oscillate" and it is very finicky to adjust and tune (homodyne mode). As a four valve radio, V2 is biased such that it will act as RF detector. In this case the "regenerative feedback" pot is a very smooth "volume control" below the point of oscillation.

If you squint with your ears on a good signal, it will drive a small loudspeaker (the 1j18b is more like a DF96 than DL96, a 1j29b would make a real driver like DL96).

I think the other less common feature is the way it does MW & LW. Though the Vidor "Lady Margaret" and similar superhets seem to use a dual winding, but not quite the same wiring.

R4LW in Limerick is "better" than a Sony ICF-2001D, as rotation of the radio to null local interference is sharper. So it should work in SE England for RTE1 LW and let you "null out" Algeria. R5Live "booms" in on MW in the evening as do many "continentals".

[neon indicator]

Well I'm happy with the radio now. I was spurred to do it by the Elektor kit https://www.vintage-radio.net/forum/...ad.php?t=70295
The Elektor "valve" radio uses two transistors, an IC and needs external aerial / earth. My design (which is like 1000 other 1920s to 1950s one valve regenerative radios) is hardly original, but does use VERY cheap common Russian pentode, less than £12 including postage gets TEN of them, and likely made between 1971 and 1991!

I was also spurred on by the BBC 1950s design. I originally made one like it testing plastic tube and ferrite coils, but the need for external aerial and poor LW performance annoyed me.

I realised that the large frame aerial would be higher Q without Litz wire and be sensitive enough that no separate aerial / earth would be needed. I'd seen some loop/frame aerials on the web, even one with 1j18b, but didn't realise till I compared with Ferrite Rod aerial just how well it works. We have no local MW in Limerick, and LW for RTE1 and good enough for day and night R4 LW was thus a goal.

The current iteration actually faintly gets German/French LW and English MW during day in Limerick, though not good enough to listen. Night time UK MW is fine.

Day and Night R4LW is fine. (If you are not sure, check where Limerick is!).

With the Maplin "crystal earpiece" ( €3) the RTE1 LW and R4 LW can be "turned up" too loud with the "regen" pot.

[neon indicator]

I put 10 Ohms in series of f- of the RF stage valve and changed to 4 x Alkaline instead of 3 as I think 4 x 1.2V nominal filaments on 3 x 1.5V nominal batteries would be too early end point. With 4 x AA that are reading 1.5V off load, the voltage per filament is about 1.35V, less than the recommended max of 1.4. The other 3 valves have 4.7 Ohms in series.

Batteries come in packs of four anyway.

I was asked elsewhere why I didn't "recommend" Rechargeable PP3s for the HT, well we had that discussion here http://vintage-radio.net/forum/showthread.php?t=72743 which has my take on it.

At 4 x PP3 and 700uA current and occasional use, plus the self discharge are Rechargeable batteries worth the hassle on a "retro" novelty valve Radio?

I'm finished tinkering with the design unless someone has a really good suggestion as it now receives UK MW during day quite well here in Limerick as well as RTE1 and R4 on LW 24hrs a day. The sharp null on loop aerial is good to ignore local interference.

My 1st valve project since my 6V6 home brew stereo 'gram amp in 1972, even my family are mildly interested! My son went and bought me some panelling board in B&Q as he thought it would be nice if I did a hexagonal version of the loop to replace the square version sometime. Mind you it was only €3.50, the 250g reel of 0.56mm enamel wire in Maplin is the most expensive single part!

[allan]

I'm wondering about the grid bias for these valves, not having checked their spec I'm a bit ignorant about their requirement. If you reversed the LT connection to the HT- supply what effect will this have on total HT drain for example?

I've been experimenting recently with an early radio, changing the grid bias supply and finding completely different overall sensitivity.

In your circuitry, by adding a ballast resistor in the filament supply you probably alter the bias slightly?
In one leg it would be more positive and in the other leg more negative.
For the regenerative detector, if you tried lifting the grid leak from ground, decoupling it to ground and connecting it to a negative supply could you, by experiment increase the gain? Depending on the characteristic curves maybe a small positive bias would help even? The same goes for the other stages.
Allan

[neon indicator]

Actually the f- and f+ direction is quite important.

Indeed since the "ballast" resistor is in f- rather than f+ and grid connected to "far end" of resistor is deliberate, it makes the grid more negative and improves bias point a little. Also a small amount of negative feedback. It's advantageous.

The "grid leak" resistor on RF isn't for bias! There is only 70mV difference between coil direct (no resistor & capacitor on grid) and a 4M7 Ohms! The grid current too tiny!

So why did I add series C and 1M Resistor? To reduce Hum pickup from chokes on Florescent lamps!

The ONLY ways to really change bias on the Russian "rod pentode" valves is:
1) Increase f- series resistor. Limit is realistically about 150mV to 250mV, but that's equivalent to infeasible 10M to 22M "grid leak" resistors!
2) Series filaments. Note that V3 grid is about -2.5V below its f- voltage! This allows larger anode resistor and more gain. V2 is deliberately at "wrong" operating point to work as AM detector on larger signals when regen is backed off. For very weak signals and presumably CW/SSB, you can very carefully adjust regen pot and tuning such that the RF stage is Mixer/Osc. V1 and V3 is only 56K load, so grid to near f- is about correct. We need V1 load lower for stability. Also there is no more negative supply. I have a 10 Ohm f- to ground ballast, at 25mA grid = 250mV more negative than f-

3) A grid bias supply. Since the grids take about 18 nano amps, this is likely to be long lived if a few CR2032 coin cells. This needed for 1j29b and 1p24b unless you running them "hot" Class A. It's especially needed for Class AB bias.

I think the LT supply is the correct way round to increase HT for V1, V2, V3 and V3 by 1.5V to 6V depending on position in series filament chain.

At 9V HT you can run with the grid to f+ (a big positive bias!) for more current. An unusual feature of this valve family is that you can legitimately have Vgk up to +1.5V on some with still under 1uA grid current and thus higher anode currents.

The 1p24b can be used pulsed at +5V grid 800W peak for 2uS!

See
http://vintage-radio.net/forum/showthread.php?t=69320
and
http://www.techtir.ie/blogs/watty/miniature_valves_1
and
http://www.techtir.ie/blogs/watty/miniature_valves_2

If you haven't already.

These family (1j18b, 1j18b, 1j24b, 1j29b, 1j30b, 1j37b, 1p22b and 1p24b ) are all "Pentodes" in that they will give a "Pentode" characteristic and have supposed control, screen and suppressor grids, anode and filamentary Cathode. They will also work like triodes and apart from 1j18b as Tetrodes. But they are not really Pentodes in conventional sense. All rod electrodes and in some senses more like original Gamatron. You can even use the 1j30b and 1j37b as "Gammatrons" as they have two identical control grids (not hexodes even though four grids). You can use one grid as "grid" and the other as "plate". Works at very low voltage.

On V1 a more positive bias reduces gain. There is no RF gain difference with 4M7 grid leak or no isolating capacitor and coil as bias path to 0V.

[allan]

OK om
It's my first meeting with these valves. Very interesting. Being made in Russia makes me wonder if they were introduced to combat EMP? We heard that the Russians were equipped to deal with this problem when I worked in Defence. I saw a few of their military radios when we re-equipped the Egyptian Air Defence comms system in the early 80s. We used nuclear hardened chips in those days but who knows if they would have worked when push came to shove?
I have played around with deaf aid valves, which from memory were DF70 series Something like that, and I seem to recall the odd UK military surplus sets used them, with the valves lined up on tagstrips with resistors etc.
I wonder if you could use reflex techniques in your experimental circuit? Another possibility is a super-regen or a homodyne receiver where you wouldn't need to bother with a reaction control. I'm sure there won't be many complaints nowadays of squealing, although I'm not sure where Irish broadcasting is heading? Maybe most people (sensibly) still rely on steam radio over there in limerick?
Allan

[neon indicator]

The valves as far as is known used purely in Military applications and some early satellites (Sputnik used them). Manpack radios, vehicles, Aircraft. Usually NiCd batteries for LT and often vibrator packs for HT. The issue was not so much about EMP but that the valves are very much superior to Germanium Transistors. Then having used them from late 1950s the inertia of Soviet Bureaucracy meant they kept making them and using them even when they did also produce Military radios with Silicon transistors.

Limerick and the South East have DAB (no where-else), but it only RTE and quality is dire. The same channels and poor quality is replicated on DTT. Plenty of good coverage on FM. AM basically only has RTE1 LW 252 from Trim Co. Meath. Athlone is long gone.

The Reflex techniques are poor and existed to save expensive valves. The Russian 1.2V Rod Pentodes work out about 55p or 65c each!