Philips 650A Low Volume

[Voxophone]

Hi all,

A couple of days ago I finally finished re-wiring my first ‘pre-war’ radio, a Philips 650A, and successfully powered it up after a check through the lamp limiter.

The set works on all bands and from the gram input (which oddly has a dedicated switch), however the volume all round seems quite low – especially on MW. Since I have an almost identical problem with MW on my Marconi 911, which is of a similar vintage, I thought an outdoor aerial might do the trick instead of the indoor one I’ve tried to date for both these sets. I strung up 10+ metres of wire between an upstairs window and a biggish tree in the garden (tethered with string at both ends), but while the volume does improve it still seems well below what I would expect.

As far as the restoration goes, all the wiring has been replaced, except around the wavechange switch where it looks more or less ok, and all the tar caps and electrolytics have been replaced except for two in awkward places below the switch (C4 and C11 on the Trader sheet). The switches and valve sockets (loving the side contacts!) have been cleaned.

All resistors test more or less within tolerance however both the tone and volume pots are faulty. The tone pot is almost o/c and I have therefore replaced it for the time being with a fixed 47k resistor. The volume pot measures up to 1.5M compared with a 350k design value and is still in place. It would be difficult to replace behind the open type mains switch.

I’ve measured the AGC voltage whilst tuning and it does not seem to change regardless of signal strength (it stays at about 16V with respect to chassis). Either the signals aren’t strong enough to activate it or there is an AGC fault. All caps associated with the AGC have been replaced however and all the resistors test more or less within tolerance. The EM1 magic eye also does not move and unfortunately is extremely dim (I’ll have to have a go at a voltage doubler or something to try and brighten it up a bit).

While performing voltage checks I was surprised to see that the control grid of the EBL1 output valve measures positive with respect to chassis. Reading up on this forum however, I found that the measurement should be taken with respect to the cathode, which gives -5V at the grid at initial warm up, rising slowly to -0.8 V and settling eventually back to about -1 V. I’m not sure how normal these values are but at least it’s staying negative.

I often play an MP3 player via the gram sockets on some of my radios, but on the Philips I’ve found that the volume only reaches comfortable listening with both the MP3 and radio volume controls at max. This would seem to suggest low gain in the output stage, but then again, I have nothing to compare it to.

I’d appreciate anyone’s ideas as to what the problem might be, or regarding what sort of performance to expect from a set of this type. (I can’t help but think it should be louder with that massive speaker!) It’s a lovely old design with plenty of interesting features and it would be nice to get some real use out of it.

Cheers

Liam

[Herald1360]

A leaky C4 could be the problem- it's the HT decoupler for the RF amplifier stage. Check the anode voltage on V1.

If you have the trader sheet from up top, the circuit scan (over two pages) is missing a bit in the middle. The manufacturer's info is better but the circuit layout is more confusing.

[Voxophone]

Thanks. I'll try that. C4 is the easier of the two remaining caps to get to if it does need replacing. I hope I never need to change C11...

[Edward Huggins]

It might be low gain in the AF stage, not necessarily the output stage. Edward

[Voxophone]

I've seen this set classified as a 'short' superhet. From what I understand from the Trader sheet, it has the following stages: RF amplifier (V1), mixer/oscillator, (V2) IF amplifier or gram pre-amp (V3) and detector/output amplifier (V4). Am I missing where an AF stage comes in for radio operation other than the pentode section of V4?

[PJL]

It is a short superhet and uses the IF amp as an audio amplifier when in gram mode. I would start by checking the voltages against the trader sheet. The AGC voltage (across C7 on trader 378) should be near 0V and should go -ve with a strong signal. If you are seeing +16V there is something wrong. You can perform a basic V4 condition check by measuring the cathode voltage which I have calculated should be about 11V.

[Voxophone]

Sorry, I think it was actually the grid voltage of the EM1 that I measured during tuning (at the connection between R19 and R20), rather than the AGC voltage. I did measure AGC as a one off, but can't remember the result off the top of my head. I'll try it at C7 as suggested.

I'd estimate the V4 cathode voltage at perhaps 21V as the control grid measures 16V with respect to chassis and -5V with respect to the cathode.

[PJL]

The EM1 rectifies the audio on the grid and sits at V4 cathode potential so unrelated to AGC.

These voltages don't stack up. 21V on the cathode, R27(160)+R28(200)=360, gives cathode current=58mA which is much too high. 21V also doesnot match the R27/R28 divider voltage of 5V*360/160=11.25V. The loading from the meter would make the real cathode current higher than this. If these values are true then check R25, R26, R27, R28. Traders are not always to be trusted but 58mA is well above the 40mA in the EBL1 datasheet.

PS: I just looked at post #1 again and note that the grid to cathode starts at -5V but gradually reduces to -1V. The cause is most likely grid leakage in your EBL1 which will be your problem. Do not use the set like this as the valve will thermally run away and damage the output/mains transformer.

[Voxophone]

Thanks. To be honest I'm working from memory as far as voltages go but I'll take some proper measurements and post them here as soon as I can.

One thing I noticed is that the control grid of the EBL1 is still positive with respect to chassis even when the top cap wire from R25 is disconnected. It's not as positive as when R25 is connected however.

[PJL]

That's another very strange result!

[Voxophone]

Originally Posted by PJL

PS: I just looked at post #1 again and note that the grid to cathode starts at -5V but gradually reduces to -1V. The cause is most likely grid leakage in your EBL1 which will be your problem. Do not use the set like this as the valve will thermally run away and damage the output/mains transformer.

Noted. Fortunately these don't seem to be hard to find for their age so I'll try a replacement. Unfortunately I'm away for work for the next month so I won't have time to try it until I get back.

Out of interest, what is the mechanism that causes the voltage to change over time? Do the internal electrodes move or emit more electrons due to the change in temperature?

[PJL]

My experience is leakage is often due to deposits on the inside of the glass between the pins and this can be cleared using a spark generator. This is probably due to a very slow leakage of air into the valve.

I have also read that thermal runaway causes the grid wire to become splattered with the cathode coating which then starts to emit electrons resulting in a grid current.

[Voxophone]

A quick update on this restoration, which I've been able to pick up again as of yesterday.

A replacement NOS EBL1 has improved the volume considerably. I still think it is much too low on MW compared with the other bands, and the EM1 is not moving on any of them, however the MP3 input is now nice and loud.

Later today I'm going to try measuring the AGC voltage as suggested above and will also probably replace C4.

I'll report back on how this goes.

Cheers

Liam

[Edward Huggins]

If the volume is lower on MW than the other 2 bands this does would not normally suggest there is a problem with the EBL1. Edward

[GeoffK]

From trader sheet 378 C35 in the MW aerial tuned circuit and C39 in the RF tuned circuit are at the top of the large cans and slight adjustment can be made to these. I also have this model and it plays very well on all bands, still with its original black caps, only the electrolytics have been replaced by fitted them on the underside of the chassis, four in total.

[Voxophone]

Originally Posted by PJL

It is a short superhet and uses the IF amp as an audio amplifier when in gram mode. I would start by checking the voltages against the trader sheet. The AGC voltage (across C7 on trader 378) should be near 0V and should go -ve with a strong signal. If you are seeing +16V there is something wrong. You can perform a basic V4 condition check by measuring the cathode voltage which I have calculated should be about 11V.

I have the radio running at the moment and am taking some measurements. With the new EBL1 I get at cathode voltage of 11.6 V with g1 at -4.3 V with respect to it. This seems much more reasonable.

The AGC still appears to be inoperative on MW, measuring 0.104 V at C7 regardless of whether a signal is present. It does however go negative on some LW signals. I haven't checked whether the EM1 is moving as it's barely visible in daylight.

The V1 anode voltage measures 195 V compared with a 245 V design value, whereas the V3 anode voltage is almost spot on at 256 V (258 V design). It looks like C4 is leaking then so this will be replaced.

[Voxophone]

C4 has now been replaced which seems to have increased the volume still further. MW is still quieter than the other bands however and fails to move the EM1 even on the strongest local signals. It does however move the eye when tuned to a patch of loud noise near one end of the scale, with the AGC voltage going negative correspondingly.

With a long wire aerial the set is now quite usable, however I still plan to try and get the EM1 moving properly if I can. I might try the adjustments as suggested however I'm a bit reluctant to touch these as most still have their wax seals. I'll need to check that this is the case with C35 and C39.

Cheers

Liam

[Ed_Dinning]

Hi Liam, if you do try the tuning coils and trimmers do not give them a good screwing in and out, it is sufficient to rock them 30 deg or so max.
Make sure you know where each one is set before moving on to the next.

Ed

[Voxophone]

Thanks. Will do. Does anyone know the correct tool to use when adjusting the trimmers on this set? The adjustments on mine are just thin bits of wire slightly flattened at the ends. I wold imagine a small set of pliers might be needed to turn them, but they look very delicate.

There is also an adjustment beneath the chassis which is a sort of wirewound capacitor (adjusted by adding or removing turns). Is there a particular knack to adusting this or is it best left alone?

Thanks

Liam

[GeoffK]

Don't touch the wire wound capacitors, you will irrevocable mess them up, this is the medium wave C43 and long wave C44 tracker. If you have replaced the black capacitors that are part of the oscillator tracking then the top capacitors C46 medium wave and C47 long wave might need very slight adjustment, this could also account for the noise at one end of the band. The whole top trimmer capacitor turns, a broad plastic trimming tool can rotate the entire top part not just the metal centre part. But all of these trimmers only need slight, if any, adjustment, chances are it would have been better to have left the original black capacitors in these positions unless obviously faulty. And best to not go near the IF transformers or their fixed capacitors, they really should be left alone. Check that stations align on the dial such as BBC radio 5 as a pointer.