Pygmy Club four band

[Mike Watterson]

The plot thickens!
I decided the 33K feed to the LO primary coils doesn't need a decoupler because it's not used to also feed DF96 g2, that connects to HT.
I disconnected all the Silver Mica capacitors to give more visibility in tracing the wiring and also to test them. All but one (196 pf, labelled 200 pF) are 20% or more out. I had then the idea of connecting the inductance meter between DK96 g2 and the HT feed point. The LW value was LESS than the MW. I was suspicious of the 33K as the leads were full length and it was all bunched up. The next tag across was more nearly level with the HT end of the 33K. I moved the meter lead there and got sensible readings on inductance per band. Naturally the OC1 and OC2 are the same, as OC2 (lower frequency) adds extra capacitance.
The four coil is a single 18uH coil switched across the RF tuning capacitor to 0V.

The rebuilt IFT was compared to the first stage on the Meguro Geniscope. The trick is to feed with a 10:1 probe from sweep out and use 1:1 probe on the secondary to the scope. The Genescope has a low impedance out and an input designed to go to the detector, but it derives the sweep from the mains, so 5mV scale on scope with trigger set to ~ gives a nice RF response. Despite using single core rather than the Litz (which is no more than four enamel wires) the rebuilt IF looks much the same.

Still some tracing of the wiring to do for posterity, then re-assembly. I have confidence it will work. All the switch contacts now seem fine.

I wonder could LW have ever worked?

Still figuring out how the aerial wire for the whip gets to the RF coil and tuning cap to see why that should kill the OC1 / OC2 LO drive.

No grid leak or AGC on the RF grid of the DK96. It's simply connected to 0V via the 18uH coil on OC1 and OC2 or either coil on the ferrite rod.

[Mike Watterson]

I replaced one Silver Mica (label is 106 pF or something). It measured 85 approximately. After padding with a ceramic disc I had less! Took disc off and capacitance had dropped to 64 pF!

One cap is 196pf and labelled 200PF, so it went back in. The other three had ceramic disc caps put across the backs.
All of coil / switch pack reconnected.
First IFT remounted and wired in except for AGC resistor.
Second IFT (the one with new winding on detector) is mounted, but not connected.

I cut grass today, so might go at it fresh tomorrow. I'm optimistic. The RF tuner seems OK and the LO tuner now only shorting at one spot. It's 500 pF + 250 pF approximately, all the similar size ones I have are 500 pF + 500 pF, and I'm not going to rip half the vanes out! I'll try some ultra fine sandpaper on card to correct width.

It's been interesting reverse engineering it and figuring out why they did some things. I'm delighted it was relatively easy to wind the IFT. I do have a project in mind to make 1.7 MHz IFTs. The current version of the ZN414 / MK484 from China works to 2 MHz. So half way between 160m Amateur band and end of European MW (US maybe extends to 1.7 MHz, but I'm not planning on visiting).

[Mike Watterson]

I couldn't resist it.

Assembled and a dead short on the HT to OV. It was the untouched primary wire of the IFT I'd worked at. Corroded and a whisker touching can. It snapped there so retinned and resoldered to cap track, put thick PVC tape. Replaced an eyelet style ribbon on can to lug. Remounted. BBC R4 LW loud. RTE R1 LW weaker, so alignment.
The detector / secondary of IFT doesn't quite peak, so I'll take that out and check it more exactly, take some turns off.

Audio quality isn't great, may be the replacement speaker, which was chosen purely on size. I'll try another. Touching SW aerial wire STILL kills LO. That's a mystery. I don't think it's even connected on LW & MW.

Really I was ready to just have this for "display" and rob the valves and audio transformer. I'm surprised I've got life now on LW (the earlier MW was very faint at full volume on really strong stations). Next time I'll see about SW and MW as well as adjusting the IFT a bit.

The high resistors where even higher out of circuit, so most replaced. It's certainly not the quality of a Pye or Philips.

[Mike Watterson]

Works on all bands (though higher SW bands needed a DK96 swap) on the "new" tag by inspection for the 33K HT feed to the LO primary coils. RTE1 at times quite loud, there is an intermittent connection somewhere.
The aerial wire to whip, rather than being only on SW bands, is connected also in some fashion on MW & LW. Add too much wire and the rod aerial reception is swamped by local noise on LW. Though right now at 10 am there are only stations on LW. There might be something on the 40m to 20m SW band, but likely needs a better SW set. It does pick up the signal generator even at 20MHz, just pick up via the aerial wire, no coil or connection, just the unconnected test lead on generator radiating.

The detector IF coil, secondary of IFT2, is just a bit lower in Q, as you'd expect from solid wire rather than the approximately 4 to 6 strand enamel Litz. AFAIK the only reason for fabric covered wire is to reduce the self capacitance on large LW & MW windings by increasing turn/layer spacing. It's also the reason for wave winding. Lots of "modern" ferrite rod coils are actually single core enamel wire with a fabric covering.

The primary is about 5 Ω and my new secondary is slightly thinner wire than can be used, my next size up or my fabric Litz would only fit about 100 to 150 turns rather than 240 approximately. L is proportional to N squared, due to mutual inductance, which is why two identical inductors give 1/2 in parallel and x2 in series, if not coupling. If they are unshielded and very close, then series will give near zero or x4 depending on polarity and parallel doesn't affect L, but doubles Q due to halving the resistance. RF flows only on the "skin", thinner at higher frequencies, hence hollow pipes or steel core works at VHF and higher. So Litz increases the Q for the same amount of copper due to skin effect. It's multiple coils in parallel.

There is something intermittent on gain. Perhaps a valve socket, bad joint or intermittent switch contact. I had it working on batteries for the first time. A B101 using four stacks of CR2032 paralleled via 1N4148s and wire fuses and 2 x Alkaline D cells in the LT holders. Up till now I've been using 63V from 2x bench PSUs in series with current limit set (both have reverse diodes on outputs and other protection) and using the LT on a 90V/1.5 B126 sized battery eliminator, which gives about 1.42 quickly dropping to about 1.32V on no load as the pair of series shunt rectifiers across the o/p warm up.

I always use that sort of LT or D cells as that clamps any accidents with HT.

Also I disconnect HT when changing parts on battery sets as many only switch HT- and either LT+ or LT-. It's best anyway on anything to disconnect supply and not rely on on/off switches.

[Mike Watterson]

How did I know how far wrong the Silver Mica caps were?
Previously I only remember replacing SM caps when they had loose rivets or other obvious damage.

I have a precision LC meter based on a PIC and kit, that has a relay and uses a 5000pF capacitor as a reference. I bought it maybe 15 years ago, only buying the PCB and pre-programmed PIC. A year or two later I bought a better, one the AADE model in work. I don't have it. My friend near Limerick with loads of gear had an expensive thing with either 0.1% or 0.05% accuracy, I forget. He also had a box of 350V epoxy encapsulated 5000 pF 0.5% caps. Maybe SM. So one of them went in and he gave me a small handful of others that tested close to 5000 pF, though not as close as the one selected for the reference. Mostly as good as the 0.5% on the label.

So I rechecked it, my DMM which reads about 1nF with nothing connected, a bare Chinese universal component tester, a dedicated Chinese DM6293L (which does maybe 200H max), and I have an old Bridge, like the Hunts style of thing. It's neon is hidden inside, instead the Y63 magic eye used to null bridge is the leakage display. The precision kit was bought as it sensibly measures VHF coils and capacitors maybe below 1uH and 1pF at the terminals. You can press Calibrate on C (open) or L (short) to zero the display for probe lead capacitance or inductance.

Measuring the 5000pF picked at random, in stock maybe 15 years:
Precision kit: 5060 pF (I should have removed leads and used terminals)
Chinese Uni: 5044 pF
DMM (ΔRel): 4980 pf (Using regular leads, and hitting ΔRel button before connecting DUT.
DM 6243 L : 4960 pF (On terminal socket rather than leads)
Old Bridge: 4600 pF approximately. It's a pot with a big perspex pointer on the 1 to 10 scale.

[Mike Watterson]

I can have BBC R4 LW really loud, OR RTE LW really loud. Works with both loud if the R4 is more like 1600 than 1500. So pointer maybe at wrong place, except the spring is in the way to move it along more. I'll hear what MW is like after dusk and then tomorrow consider where the pointer should be with cap fully closed. I'll examine the scale for alignment marks tomorrow.

My memory was in error on the B101 pack. It was made up years ago and is used on a few sets. It's got 12 columns of CR2032 in a 4 x 3 layout, with two columns in series for each total stack of six. The recommended maximum CR2023 current being about 1.5 mA, so 9 mA without exceeding discharge recommendation. Most 67V sets are a bit less than that.

Still nothing heard other than sig gen on either SW band, but at least even with about 2m of wire the bands all still "run" even if you touch the uninsulated joint. My ancient B101 pack is down to 50V, but any 67V radio should work down to about 45 or 46V on the HT pack. The end point is roughly (pack voltage label) / 1.5, though the maximum volume will be low. In contrast the radio will stop quite suddenly, depending on condition of the 1R5 / DK91 / 1L6 / DK40 / DK92 / DK96, just over 1V for a new one and about 1.1V for a used one. Really used ones don't go on SW except with fresh cells, or at all.

[Murphyv310]

You've done well with the set. It was a basket case in fairness.
I've a mains battery Rhapsody here that was always deaf and a bit hummy on mains. Left it for years so perhaps will give it a go once I get the other things off the bench.
Great write up Michael.

[Mike Watterson]

It's very challenging when you've no schematic or alignment information and it's more than one wave band. I always find multiband (more than MW & LW) coil / switch units hard to reverse engineer.

Now that's it's dusk the MW is getting lively and some stations on OC2 (49m to 20) quite loud just with a temporary "whip". The RF SW coil is nearly fully out to peak, so I'll check SW LO setting on OC1 as OC2 just uses the same coils. One station at the 40m end of OC1 and a loud Arabic one at 3/4 way to the 20m upper limit of OC1, the higher SW band.

I was very suspicious of the 33K HT feed resistor to LO coil/switch pack as two SM caps had obviously been re-soldered and two SM caps had snapped leads. Always hard to see where a stray MW coil wire and stray LW coil wire came off when they are Litz. I got confidence when there was SOME life on MW, but it couldn't work properly on the other bands with the resistor going to the MW coil tag, hence willingness to take out the IFTs and all the SM caps and trace it somewhat.

Also early on it was evident something was wrong on the secondary of the 2nd IF. No doubt the selectivity would be slightly better if Litz had been used, but I counted the wire I unwound.

So in addition to all the original listed issues:
One IFT coil needed rewound.
All the wires had to be redone on the B7G plug of the power connector, and the socket holes attacked with the finest Dremmel cone in/out by hand. Works better than spinning it with the tool.
The HT feed 33K to LO coil/switch pack was on the wrong tag. It had full length leads and was folded up. The correct tag is close to the point on the DF96 were it's soldered to.
Two intermittent connections
Resistors and caps

I've never replaced so many resistors or Silver mica caps, ever.
150, 330, 1M, 2 x 2M2, a 47k gone to nearly 80K, 3M3, 33K. I think only the 10M DAF96 grid leak and the DK96 osc g1 27K to f+ are original.
One SM was OK.
Three had small (10 to 33 pF) ceramic caps soldered across the back.
One Silver Mica replaced because every time I soldered it the value reduced. It was labelled 106PF or 105PF, started at 86 and went down to 62! Replaced with 100 + 10, which was 1pF higher only than label. I think 177 pF might really be a design of 180pF and 106 pF a 100pF? My theory is that the caps were marked with actual test value to nearest 1 pF before coating in wax. The white paint is on the main paxolin substrate. The other outer layer is very thin.

There are 3 spots near the closed end of the LO tuning cap that short. It's OK fully closed. So fine dry "wet & dry" paper and gently sand before final alignment. I must polish and varnish the actual brass pointer so it's visible against the black hard board.

I've seen the PSU circuit for the older 5 band Club that uses the DK92. It's on DocTSF . An autotransformer for HT of about 82V and LT of 8V. The 8V dropped to 6.2V. Then that's folded on 2 x 1.5 D cells for battery, with the 8V to 6.2V dropper resistor across the battery!

The model I have ominously has the isolating capacitor on the Terre (Earth) socket, but a weird glass thing that looked home made from a glass fuse (it seems to be a capacitor) had its covered wire enter the Earth socket and was soldered to the Ant socket. Surely that wasn't original?
I presume the Ant socket connected to the same place as the telescopic whip, which doesn't seem isolated! So perhaps that mysterious capacitor fed the whip too. The whip does seem to connect on LW & MW, via switch contacts, as well as the OC1 and OC2.

I've no idea how representative of French battery portable valve radios this is. I had French 1962 approx AM/FM mains set and the only odd aspect was a tall rotating card cylinder with LW & MW loop coils at right angles and at right angles to that a loose helical wire. Also the VHF tuner head used an ECF82 which I'd not seen before. Otherwise it was quite like a German set.

[Mike Watterson]

I've a mains battery Rhapsody here that was always deaf and a bit hummy on mains. Left it for years so perhaps will give it a go once I get the other things off the bench.

At least those have approximate circuits. Mike/Crackle will tell us how many variations. I've cream one that had a faulty mains/battery switch which is like relay contacts. It's working well when I finished.
I've a second that Mike & I think is a home made case (RGD badge). It's quite a different version of the family. I've just lifted it off the top of the bookcase.
Had it years now and not restored it. Two pin (flat) cable socket missing. It looks like it takes different batteries and maybe has an aerial socket. The battery compartment had an unopened pack of tissues styled like $100 bills.
115, 200, 220 and 240V mains selector.
Insanely complex mains/battery switch so perhaps it switches from 7.5V (probably 6.4V to 7V really) series to 1.5V parallel.
Selenium rectifier with only one wire?

[Mike Watterson]

Rebuilt the tuning capacitor.
Important to draw the cord layout and know which of the three wires going to coil/switch pack are RF, LO and 0V.

1) Remove it, noting screw locations and a safe container to collect them. Remove drum first noting position fully meshed. This also has 2 x alloy spacers.
2) Remove the trimmer caps. Use a marker to indicate RF mica in case a different thickness. There is usually a metal washer, an insulating washer, sprung connection to stator and then the mica.
3) Unmesh fully and remove the rear lock and grub screw. There may be a ball bearing, this one did. Catch it!
4) Depress rotor spring and lift out rotor from rear catching the ball bearings at the front in a tub. Move the spring out of the way.
5) Clean the parts. Put on WD40, but it MUST be washed off later with alcohol.
6) Cut fine emery paper with a half circle to match rotor, for the stator. Use folded over, if need be add paper /card. On this one a folded emery/wet&dry was perfect.
7) Gently rub along rotation arc.
8) Use straight bit for the rotor
9) Clean everything with alcohol and blow it off.
10) Stick the front balls to shaft in the groove with pure BP petroleum jelly, which is cheaper than Vaseline which has additives. Remember one is for the rear. Put a dab in the rear depression of the shaft.
11)Drop in fully open, then fully mesh, gently, using pliers on the front shaft.
12) Depress and put slotted or holed spring behind rear shaft. You may need to jiggle rotor from front to get it straight.
13) Drop in the rear ball and add the grub screw. Centre the rotor plates in the stator plates.
14) Check using current limited PSU, beeper or DMM/Ohmmeter that there are no shorts on full travel on RF or LO sections. There weren't! But if there are, note position and redo from step (3).
15) Add lock nut and ensure grub screw doesn't move.
16) Clean the mica, insulated washers etc of trimmers. I replaced the steel washers as they were very rusty. Reassemble trimers.
17) Check again for shorts.
18) Remount and reconnect.
19) Re-string the pointer. You DID do a nice diagram? Ensure free running and not catching. Replace cord if in doubt of condition. The entire cord, wheels, drive shaft can be waxed, oiled or Vaseline coated. Replace spring, if needed, with a ballpoint pen spring if you have nothing better. I've used steel guitar string, the first that's not overwound, crab line, monofilament fishing line, beading thread and linen button thread on various sets.

It was perfect. No cracks/pops across any band.

Now, any experts? With a OC1 20m to 40m scale roughly marked with dashes, which of the two tuning points on SW is correct? They are of course 2 x IF apart. Which is why dual conversion was invented and why some military sets use 1.6 MHz or higher single conversion.

Next: Any better ideas than toothpaste to remove abrasion on some places on the scale?