Replacing a DK96 wiith two 1j24b

[Mike Watterson]

No, they are like the EF80, barely manage more than 100MHz. Rated for up to 60MHz, which means just maybe a clever oscillator or mixer MIGHT work to 120MHz on a good day.
I'd be trying which ever PC86 or PC86 I have only at maybe 70 MHz or 90 MHz, not sure yet on the VHF test strategy. I'm baffled where it came from as all my TV stuff "vanished" between 1972 and 1976, I think. I found it last week among my meagre collection of Novals. I've so few valves that I have just a box for each base, apart from the battery B7G, Rods, Nuvistors and Sub-miniature which are in drawers.
Late 1920s B4 & B5 battery
B7 and related mains types
P8 (some battery)
Octal (some battery)
B7G AC heaters only
Rimlock (DK40 is with the B7G battery stock)
Noval

I've no Y8 battery types yet.

[Murphyv310]

The super regen I made with an 1j29b works up to at least 124mhz as I've actually heard Prestwick control tower on it.

[Mike Watterson]

The 1j29b has the most RF gain. A VHF superhet might need it rather than a 1j24b.
Did you try it on economy mode? Only one filament?

[Murphyv310]

Haven't tried it on half the filament.
I did try it with a 1J24B and reception was poor, it ran out of steam around 95mhz.

[Mike Watterson]

95 MHz isn't bad for a 60MHz valve.

[Murphyv310]

Totally agree.
The circuit for it was from a US magazine and for a single pentode acorn valve with 120v HT and a 6v battery for the heaters.
Mine uses 2 AA Ni-Mh and a little inverter that gives 100v, very neat indeed, I've the filaments in series via a limiter resistor. Like any super regen it's quite fussy so I put it in a tin box, it is actually quite loud and really sensitive.

[Mike Watterson]

Here is the first stage:

Make a screen, ideally with two compartments, a little shorter than straight part. Bend back the g3, shield and f- wires to solder to the shield. It will connect to the f- pin of the B7G socket.
   
   

Here is the plug made out of the bottom third of a pin protector and jumbo paper clips. The wire diameter is only slightly more.
   
I can't find the rest of my silvery ones. I don't know what the plating is, but it solders well. I bought two boxes in Eason's, Essential brand. They are allegedly brass plated and also solder fine.

[Mike Watterson]

Here is the top and an uncut one. Inferior solder. I've since bought 500g of decent 60/40 0.9mm Multicore/Henkel/Loctite solder from Farnell. This is maintenance / repairs. A NEW product isn't going to use home made B7G plugs!
   

The cut part is set on an uncut pin protector to get the pin length right and hold them. The jumbo paper clip is cut about 1.5mm above the surface and bent out. Then pressed with pliers. The hole is slightly larger on the cut top than the uncut surface as the protector holes taper to allow easy insertion (which is also why there are eight). Sometimes a tag or PCB socket for B7G has 8 pins, the 8th being for the centre screen. Often B9A has this also.

[Mike Watterson]

Here is a Pentode - Pentode in partial assembly. The green wire is the oscillator g2. I decided to remove that from pin 5 and connect g2 to anode to have triodised oscillator as all the mixer - oscillator circuits are for triode oscillators, The g1 and g2 of heptode and octode oscillators are really a triode; some even use two rods or plates as the g2 to feed electrons to the mixer stages.
   
I removed insulation from hook-up wire to colour code according to B7G pin. So yellow is the LO grid and blue is the RF grid. The two long wires are the B7G g2 (to oscillator anode) and final anode to IFT, the Pentode mixer anode. You actually have to ADD about 6pF to 8pF to the 1j24b pentode anode so you don't have to retune the IFT. It's a lower capacitance than most battery valves.

Here it is assembled, without a sleeve on the oscillator anode. You can see the g2 of the Pentode mixer (orange) and the g1 to RF grid (blue) and two 10 Ohm resistors to the violet filament + wires.
   

Anyway DON'T make a pentode - pentode. It's a triode - pentode that's wanted. So no green wire; that pin to HT has no connection, which is actually a second -f/screen grid to 0V pin on an 1R5/DK91.
Instead the oscillator g2 connects to the oscillator anode as well as the RF pentode mixer g2.

[Mike Watterson]

Here you can see that an outer sleeve of about the same diameter as a B7G valve is possible, but more height is needed.
   

Replacement of any series type, 25mA or 50 mA will need a parallel ballast resistor. Though the datasheet shows 13mA for the glow current (literal Russian for filament) it's closer to 11mA for a1.25V supply. The 25mA and 50mA is also a nominal value. I need to measure some DK91/1R5, 1L6, DK92 and DK96 at typical series voltages / currents to see what voltage to set on the triode-pentode, measure actual current at that voltage and then calculate the ballast resistors.


A variable capacitor had the two rotor pins soldered to the screen and stator to the exposed anode wire connection / join on the mixer pentode to allow adjustment of capacitance to peak the IFT tuned for a DK96.