Golborne Vintage Radio

The Rod Tubes are lower voltage than the regular 1.4V valves. Most will work well to below 0.9V.

For RF where you want to feed the "cathode" the well known approach with direct filaments is a 1:1 transformer wired to f- and f+. The f- and f+ are shorted for the signal via a capacitor and either pin is the input, or two caps and feed the junction. The transformer windings are in series to the filament supply such that the DC currents cancel, but it is a choke to the common mode RF (or indeed 38 kHz or audio). The input side of the transformer, a common mode choke, has both connections to RF ground. The f- and f+ point may be tuned to ground if desired. The g3 must also be at RF ground, usually zero volts. The supply is then 0V and -LT or whatever to the filament, nominally 1.2V.
You can only use regulated DC due to the filament design. A DEAC approach on trickle charge has close to 1.5V using NiMH, it's a bit lower on NiCd (the original DEAC). There is no need with modern regulators. Though shorting the Adj pin to 0V of almost all adjustable regulators is about 1.25V, slightly high, but FAR better than a trickle charged NiMH.

But the whole point of these valves is the low battery current on LT, apart from the 1p24b, which are about 220mA each. Obviously you can design mains powered gear, but if using mains it's far better to use EF80s and ECH81s. Common and cheap. You can make a decent push pull audio amp with EF80 or ECH81. You could even make an entire AM/FM radio with 2 x EF80 for VHF preamp and Mixer / Osc. One can be the AM LO. The 1st IF FM amp can be a Pentode Mixer using LO to g2 on AM. A pair of EF80 as push-pull audio can use a small 115-0-115 mains transformer without saturation. Copy the Vidor battery AM/FM set that uses all DF97s. Use an AGC amp and high level inverted AGC on the g2 pins.

You can also make an entire AM/FM radio using ECH81. Note ECH83 is the same valve. Apart from audio out you can make an entire radio using EF89, ECH81/ECH83, ECC82 etc all with 12V heater and 12V HT.

The clamp diodes need things like 1N5401 to manage the current if filaments are disconnected and are more suitable as last line of defence safety devices. Note the 1N4001 or 1N5401 are OK for 1.2V, but for the D series valves you want the 1N4007 or 1N5408 as they have a slightly higher voltage at the same forward current. Some parallel 125mA D series UK sets did use a selenium rectifier as a shunt regulator. Using either the 1N series or selenium, the regulation is poor and voltage rises alarmingly if the o/p valve fails or is removed. The series resistor is fairly critical to the usual filament current.

Hi
Thanks Tracy and Michael.
I've scrapped the idea of using the 1Ж42А for a mixer in the meantime. So today I used one as an RF amp from a ferrite rod antenna and tuning cap with a little positive feedback from the G2 and HT of two turns around the ferrite rod. The anode was loaded with a 2mh choke. The stage worked well with only RF on the anode, using a set of high impedance phones, coupling cap and a germanium diode I was able to receive half a dozen stations with no external aerial. Three were well loud enough for comfortable listening, performance being far more than I expected. Later I added another 1Ж42А wired as a leaky grid detector, the result was disappointing. I then decided to use the germanium diode as the detector and the second stage as an audio amp. The results are truly brilliant, I'm sitting in the living room listening to Spirit Radio from Ireland at good volume. I'll put up the circuit tomorrow.

The Rod Tubes are too linear to use as a leaky detector. You can use one as a diode! Connect all pins except f- & f+ as the "anode".
Or ground all pins except f+, f- and use g1a and g1b as anode.
Or just g1a as an anode and experiment with g1b with -ve to 0V as volume control (gamatron mode)

Note the Ge Diode was invented long before Rod Tubes, in 1930s. A modern 1N60 is still the best diode detector for AM. Came out in 1946.

Note unlike ALL regular valve contemporary to Rod Tubes, there is NO internal screen. Absolutely EVERY radio used a shield to clip the Rod Tube to the chassis, alumina Thick film circuit or PCB.
Coffee tin or other tinned steel is best as you need magnetic screen as well as electrical. Compare bringing a strong magnet near a Rod Tube and a DF91 / DF96.

The 1j42a and 1j37a are fantastic mixers, but only using a separate oscillator and f1 as common mode to g1a and g1b and f2 as differential on g1a and g1b, otherwise poor.

Without the transformer with RF and LO on the centre tap it's best to tie g1a and g1b and use it as a mixer the same way as a 1j24b, that is a separate triodised 1j24b (or 1j42a) as an oscillator and tie the oscillator anode direct to the mixer valve's g2 as the only g2 bias. RF then to mixer g1 and IF from mixer Anode. I've tested this plugged direct into a DK96 socket, with NO additional parts, using the 1j24b, other than a small resistor in each of the f- leads to drop voltage by 0.25 volts.

Because the Rod designs are very linear, the best mixing action without the g1a and g1b special drive is signal on g1 and LO on g2. That's how mixers were done with regular valves before Hexodes or Pentagrids were invented.

A single valve as mixer oscillator will work, but rubbish compared to two valves unless it's a Heptode, Octode or Nonode.

Thanks Michael.
Later last evening I loosely coupled the signal from my modified Wellbrook loop and received at excellent volume from some Spanish stations then lo and behold Caroline on 648 came through, the tuning is pin sharp with no splatter from adjacent stations. I tried to do a video using the phone but the mic is too insensitive for decent audio off the headphones.
I'm thinking of using an Epad chip for detection as it's very high impedance so should not load the RF amp as much as a diode, what's your thoughts on that idea?

PS. Do you have any practical circuits for mixers
for the 1j42a?

A diode doesn't load. It's the resistor after the diode that loads.
We did extensive tests in a lab. Best RF detector is a 1N60. Better than a valve detector in a DAF96 or EABC80 etc. Note the valve sets often used about 500 K load, the 1M pot in parallel with AGC filter / line. Certainly you can use 100K.

I think Joe put up some circuits on the radio museum for the 1j37b from Russian schematics. They work for a 1j42s. But it's really simple!

Add a second winding of about 100 turns (you can try more later) using two wires at the same time to the ferrite rod tuned aerial on a separate piece of card. Put it at the middle if it's got a MW coil one end and LW at the other. Two colours of enamel is handy, that's bifilar winding. Wire start of one wire to the end of the other: That is the centre tap. The free ends go to g1a and g1b.
Put a 1 Megaohm "grid leak" from the centre tap to either f- or f+ depending on your supply volts.
Connect 100pF from centre tap to the Local oscillator output (just copy the part on g1 & g2 on any battery valve heptode/octode oscillator using a second 1j42a or a 1j24b if you have 45V Ht). The LO is triodised, but g3 to 0V / f-. If using a 2nd 1j42a as an oscillator short g1a and g1b.
Then on the mixer the g2 is HT (if only 9V) or a potential divider, g3 is f- and the anode drives an IFT. You can use any double tuned valve IF or two single tuned transistor IFTs with the secondarys back to back.

For SW, add 5t to 50t on the bifilar coil (really an accurate centre tapped coil), aerial, earth and tune that coil instead of MW or LW. Leave unused band coils open on a ferrite rod and separated as much as possible.

Verify that on MW, the LO leakage is less than a transistor set.

Always put earthed (f-) cut up coffee tin (or similar) sleeves on every Rod Tube, RF, IF or AF.

If you are going for small, also use 2 x transistor IFTs back-to-back on the IF amp anode to drive the 1N60. Try a 100 K to 250 K ohm pot to drive audio amp. Use it +ve like a transistor set not valve. Then if you use a DM70 or the Russian version as a Magic eye, the junction of anode and HT resistor can drive the g2 pins of the mixer and the IF amp for AGC, max volts = max gain, min volts = min gain on g2. Regular negative AGC on the g1 as per normal valves hardly works, because the rod tubes are linear sharp cut off. The variable mu / remote cut off valves use a progressively wound grid, i.e. the spacing varies. Impossible to do on a Rod tube design. But the 1j24b with 85V HT certainly do decent gain variation varying g2 from 20V (min, strong signal) to 55V (max gain weak signal). The 1j42 will probably do something similar, but at a lower voltage.

If anyone wants some bifilar enamel wire I have a good sized spool of it. The wires are twisted together. Not sure about the best way to send small quantities. Free except for postage and you might like to make a small donation to the BECG: www.becg.tv

I'd be interested Jeffrey.
A donation will be sent for sure.

This doesn't actually need twisted together wire, just winding the two coil in parallel (at the same time) is good enough, though obviously actual twisted bifilar is better balance (less LO on the ferrite rod or radiated on the aerial. The actual twisted pair enamel is easier to use on cores or binocular cores, but using regular wire is easy on a former on a rod.

A wideband RF transformer with a centre tapped primary, secondary or both, on a suitable binocular core of course needs the actual twisted bifilar wire to do a centre tap as otherwise the balance is poorer and it's hard to wind. A typical example that might use a small core or toriod is the pair of centre tapped transformers to feed a ring diode mixer. The RF in on one transformer. The ring of diodes is between the transformers and the Local oscillator feeds the centre taps. One centre tap can be antiphase LO or 0V. The second transformer has the mixer products.
So a 1j42a or 1j37b simply replaces the two transformers and four matched diodes (in a ring, not bridge) with one centre tapped transformer on the input and any sort of suitable transformer (no centre tap needed) on the anode. A ring diode mixer is very good but needs well matched diodes, two accurately made bifilar transformers for centre taps and quite high level Local Oscillator. The unique split grid Rod Pentodes with a g1a and g1b are doing the same thing.

The Heptode/Octode and transistor combined Mixer/Oscillators are particularly dreadful for radiating LO on the RF in. They also suffer more from intermodulation due to a strong signal causing pulling (FM) of the local oscillator so that the strong unwanted signal has FM on the IF, which then will convert to AM via the IF filter slope.
The Hexode and Pendode mixers take the loc oscillator on a later grid than RF, so less leakage. Obviously a Heptode or Octode with RF on g1 and the LO from a separate oscillator fed to a later grid leak less LO.