Pilot BM109 Poppet

[Mike Watterson]

Warning, Danger, Will Robinson!

The chassis is potentially live with no interlock. Not a big issue on a mains only radio, but this also takes batteries.

NINE useless resistors packaged as paper dielectric capacitors, AKA Waxies. But not all paper dielectric capacitors are in waxed card tubes. Some are in metal cans with rubber bungs. Some use a plastic that goes brittle with age (Hunts Modseal). Some military ones and some used in EMI sets are well enough sealed that they might be OK. The so called Paper In Oil, may just the same thing with a better sealant. The paper is impregnated with a kind of paraffin, but paraffin does eventually allow water to dissolve in it. The really good PIO use hermetic seals and a more expensive oil more resistant to water.

This set had nine. One across the mains. Two on screen grids, one on HT, mysteriously a 100nF (0.1uF) in series to switch to switch in the LW Local Oscillator (LO) capacitor. I suspect a piece of wire works there.
1S5/DAF91 anode to 3V4/DL92 grid, 1S5/DAF91 grid, AGC time constant, 'tone correction' (snubber) on T4 primary output transformer.

The 1R5/DK91 filament was open circuit. I first checked LT from bench PSU at 7V. About 6.5V to 7V is right for mains or external PSU. The 7.5V is a nominal FRESH battery, and the set should work somewhat at about 5.5V or a little less. The LO stops first. I used an external 90V HT PSU also.

I could see leakage of LO on the scope on volume control, so it was tuning and also waveband changing.

Two turns of wire on ferrite rod and 470K, close tuner and select MW. The four IFT coils spot on.

About 600 KHz and 1.5MHz also worked on signal generator.

Tried LW with 250 KHz. Where are the pesky trimmers?

The MW RF & LO for 1500 kHz peak are on the main capacitor, but under the chassis, so you kill yourself if it's not an isolated supply. Really awkward. No LO trimmer for LW (not extraordinary, hence suggestion to adjust 500m/600kHz LO coil before LW. LW RF trimmer is just under the back of the wave change. Tweaking it (it was fully tightened!) brought up the signal massively. So much so that RTE1 LW was breaking through. Disconnect SG and peaked up RTE. Only a faint fistle at R4LW.

So fitted it all in the cabinet and put a 5 amp three pin plug on. My boxed shaver adaptors have a hole drilled for the 3rd earth pin. I considered putting a polarised flex, but left the anoymous two core on it. I'll add a warning label on chassis, cable and plug.

Next time I'll align MW properly now it's in the cabinet and I can see the scale. Then redo LW.

Finally I'll test Franken Mixer/Oscillator with a ballast added so it's 50 mA series. The 1R5/DK91 has no screen grid supply like the DK92, 1L6 or DK96, but my Triode-Pentode oscillator/mixer doesn't use it. So it should "go" in the DK91/1R5 socket.

Note that the Trader Sheet 1331 14th December 1957 and the R&TVS 1958-59 volume page 212 use different component numbering.

IF amp is a 1T4/DF91.

I should check what voltage & type the 1nF to the aerial socket is, as the other end of it goes to the chassis via a tap on the MW coil. It's not shown on the Trader circuit.

The Selenium rectifier appeared to be open circuit, so I soldered a 1N4007 direct from mains in wire to the 32uF smoothing cap (C26 Trader, C23 R&TVS). I've 230V here (because that's the Irish harmonised 220V) so I'll check HT caps with 30V PSU first and then try on the 250V tap. It's possible no extra series resistance is needed.

There is what might be an asbestos disc on the end of the dropper. Such things are best left alone or soaked in repair cement or similar to stabilise fibres. The DAC90 (not a DAC90a) has rather more asbestos. It will get the cement treatment.

The later Vidor mains/battery sets with a potentially live chassis (some had isolated PSUs) had a cunning arrangement of an ordinary in line cable two pole plug and socket that unplugged when the back was opened.

[Mike Watterson]

I couldn't resist looking at the PSU.
At the 230 to 250V setting the DK91/1R5 (the last valve in the chain) had just about a volt. So no reception.
At the next setting it had 1.26 but dropping. The HT is a healthy and steady 87V at the 210 to 225 setting. Should be about 84V on the 250V tap with 240V AC. The intermediate DC taps are actually also AC intermediate voltages.
The HT to LT dropper is supposed to give about 6.5 to 7.0V. It was 5.6V. The 1800 Ohms measured 1970 Ohms. I replaced that with 1k in series with 750 Ohms. The LT was briefly 1.37V but dropped to 1.28.
The 750 R is a vintage resistor, it may be going high.
I also suspect the 100uF capacitor that's decoupling the last 3 of the five filaments. It's one of those 2 part plastic cased electrolytics that are often faulty. The main HT caps seem OK as the HT is stable 87V.

Some reception at the HF end of MW, still no R5 and I think a short on one of the tuning gangs just after the half way closed point. Arrgh.

Definitely no more tonight. Reading time. Patricia Wentworth, "The Silent Pool" tonight. Miss Maud Silver would give Miss Marple or Jessica Fletcher a run!

[ppppenguin]

Note that the Trader Sheet 1331 14th December 1957 and the R&TVS 1958-59 volume page 212 use different component numbering.

Looks like you're doing a grand job on this set.

Trader sheets usually have different component numbering to original maker's data. Not sure why, though they always redrew the circuits in their own house style. I don't know if R&TVS also differ. So as a general point, in any thread, when referring to C99 or R34 it's worth saying which source of data you're using.

[Mike Watterson]

Here is a scan from my own R&TVS.
   

So I snipped out C25, 100uF. This set uses too few or badly placed earth tags, so the wires snake around each other making unsoldering even more difficult than usual.

It started at 1mA, on a 6V supply, and rose to over 2.5mA before I lost interest. My electrolytic caps are now all old so I have the replacement 100uF 16V on 10V for a while. Current quickly dropped to under 0.1mA. I'll check does it actually have capacitance. So the original would have been over-running the 3V4 output valve filaments as well as under-running the other three.

The mains supply and power switching is a no-win on this other than using a shaver transformer (2kV or 4kV isolation and 20W power limit, most have a special core design and/or a thermistor on the primary side. ). If the neutral is to chassis and live to the rectifier, then EVERYTHING including the chassis can be live when the power is off, as it's the mains connection to chassis that's switched.

[Mike Watterson]

I calculate that the heater ballast should be closer to 1500 Ohms than the specified 1800 Ohms. So certainly 1970 Ohms is too high.
I'll need to find something or put ordinary resistors in series and parallel. Needs about 4W rating, probably 5W. About 80V and approximately 50mA, perhaps a little less of both.

[Amie]

That c5A is a definite afterthought, I wonder why....

[Mike Watterson]

Maybe they had excess 0.1uF 350V caps to use up? It doesn't even have DC on it. Because it's in series with a 390 pF even if there was AC (there is RF), any RF or other AC is going to be all across the 390 pF. So it does nothing.
Maybe the 390 pF was originally to go there and the two connected ways of SW1b for LW were supposed to go to 0V. That's how I'd wire it.
It's generally a terrible layout of chassis, despite the space. The Pygmy and Murphy I did recently are both better layouts, though the Pygmy is a bit mad at the volume control and the output valve wiring. Hence the tendency it has to have feedback whine unless the wire from anode to transformer is perfectly dressed.

I remembered I had a box of Marconi/HMV (EMI) battery valve PSU parts that came with some old IFTs I bought. It had four power resistors. Two were obvious scrap. One looked OK but was open circuit, so I cracked off the cement at the ends, crimped the solder at the base of the wires and it measured 1795, as written on it. I put fresh solder at the base of the wires and it was still OK, so it went in to replace the 1K with dubious 750 Ohms in series.

I disconnected the two main HT caps in one can and rechecked them; they are perfect. The four section main dropper is well within tolerances.

So now getting 6.46 V LT, vs the trader sheet 6.75. Except the 1R5/DK91 LT is at 1.45V!
1T4/DF91: 1.27
1S5/DAF91: 1.21
3V4/DL94: 2.48, with 1.31V across the 470 Ω, earlier I mis-labelled that as a DL92. It's a 3V4 in listing and fitted.

I took out the valves after unplugging isolation transformer. The 470 Ω is 522, which is 11.1% high. The 220 resistors are within 4.5% and the 180 Ω across the 1R5 is only 5.6% high.
Perhaps some sockets or pins are more tarnished than others. It is normal to run the mixer/osc filament a little higher than the IF amp and detector/preamp so as to get the very last out of the battery, but it's too high on the 1R5, especially if the LT was at a higher voltage.

I'll need to think about this. Suggestions welcome.

[Murphyv310]

Hi.
C5A is most likely for safety reasons and possibly to reduce hum effects on mains operation. I've seen it somewhere before on mains battery valve portables. The shaft "could" notice could become live if a fault occurred. Seems they want to isolate the band switch totally from any direct path to chassis etc through L3.

[Mike Watterson]

The tuning rotor, the shaft, IS potentially live as it goes direct to the chassis.
The "RF live" bits are the stator. Trimmers too are usually wired with the slot head/moving part to OV and the fixed part to RF, built-in on tuning gangs or standalone.

The switch shaft being live is maybe more likely than the volume control, but not particularly due to the chassis path of L9.

The switch itself also switches one side of the mains OR the -HT battery to chassis. The +LT battery or the ballast resistor is switched by another part of that switch to the filament chain.

The power switch on the volume control is disconnecting the LT- of the battery and the -ve HT from the band switch, which is either the -ve from the battery or direct to the mains cable.

So no matter which way you wire the mains lead, the band switch has a direct connection to mains on two pins and the other side of the mains via a rectifier HT dropper to 85V and ballast resistor to two contacts on Sw1c. I measured 320V DC between the R12 ballast switch contacts and the end of the mains lead that goes to SW1D pair of mains contacts. No doubt 230V AC peak via the rectifier and stray capacitance to the 20 MΩ input impedance DMM.

So while an interesting idea; the C5a does nothing for safety of the wave change switch shaft or the tuning shaft.

I've quite a few sets with the crazy combined wave change and battery/mains, including some Pye models. The Pye made Invicta I have though has a separate three way switch styled like a toggle switch. Philips and many USA sets had the battery/mains switch on a dock for the mains plug. One USA model has to have the original size and shape of 110V two blade plug as it has a spring loaded drum for the mains cable and it can't retract fully and operate the switch unless everything fits properly.

[Mike Watterson]

I measured the 1R5 filament as well as loads of others. Most were a bit over 7.5 Ω and two 9.5 Ω. The fitted one was 11.5 Ω! filament chain looked better with one of the 7.5 Ω valves. I added labels to the 9.5 Ω filament valves.

Something is shorting just below 200 kHz / 800 KHz approx. However I did the 1500 KHz and 200 kHz. I fine adjusted the LO core for 200 KHz = 1500m on scale as it won't do 600 KHz. Then re-checked 1500 kHz.
It now gets BBC R4 LW and poorer than it was on 252 kHz. It's ominous that the tracker trimmer for LW RF is to be adjusted at 200 kHz, terrible LW RF tracking.

Plenty of static now on MW.

So the only thing left is to clear the short on the tuning when it's about half closed or more. Really I might as well have left it as a display model.

Both the front cover and rear cover hinges slide apart.