Philips 531A Motorboating Problem

[Mach One]

I have been trying to renovate this radio for a couple of years now - slowly as time permits. I originally bought this radio at a Jumble Sale in the mid-Seventies, used it up to about 1980 and then passed it on to my cousin, who used it lots until it was retired to his loft. Thankfully he still had it and has returned it to me, still working but with a very dim magic eye. It is the only radio that I have left from that era of Jumble Sale radio affluence, unfortunately.

I had the chassis out of the case and have put in lots of new resistors and capacitors, but not all - only those that are easily accessible, mainly in the audio stages. The only odd value I found was a resistor in the circuit to the magic eye. I fired the chassis up and the radio seemed to work. Having done all this, I put the radio back in its case, expectantly. However, there was now motorboating on all bands, which changes its character slightly from one end of the tuning range to the other and is on all bands.

The motorboating is coming from the RF section. It is not there with the volume turned down. I first thought that it might be some mechanically instigated issue as it had not been evident up to the point that I put the chassis back in the case, but having taken the chassis out once again it still does it.

I decided to measure all the voltages as laid out in the Philips service manual. These are generally good except for the following, as measured by my Fluke 73, which is of course a very high impedance device, and with no antenna connected:

V1 - (ECH42 in this model)
Vgt+3 (pin 4) = 0.001v (should be -0.5v)
Vg2+4 (pin 5) = 162v (should be 54v)
Vg1 (pin 6) = -7.75v (should be -0.5v)

V2 -
Vg2 (pin 5) = 145v (should be 54v)

I am aware that as V1 g2+4 and V2 g2 are connected together they should be the same. I suspect that the actual value varies over time and may even be going up and down with the motorboating although the meter did not suggest this at the time. Interestingly, the motorboating could be stopped instantly by touching either V1 ah (pin 2) or V2 a (pin 2) or g1 (pin 6). R8, the resistor feeding thos grids measures 55k and the capacitor seems to measure open circuit as you would expect, at least on my fluke with the power turned off, which is probably no kind of measurement at all! The reading at V1 g1 (pin 6) might be the AGC kicking in with the motorboating?

Obviously, I would be grateful for any comments that would point me to the culprit.

[60 oldjohn]

If it worked ok out of case, may be worth swapping speaker leads over. I can't see how this would work but read it somewhere. Hope it's an easy fix.

John.

[60 oldjohn]

I think the Fix I mentioned may have related to the 1930s mains energised speaker set. Just had a look and yours is 1950s, so not for your set.

Good luck with it, John.

[Mach One]

Thank you, John...

It worked originally "out of case" but now when out of case it is motorboating just as it is "in case."

[G6Tanuki]

Does it have a frame- or foil-stuck-to-the-inside-of-the-case antenna?

Reason I ask is that many years ago I came across a radio which suffered from instability when used with its frame-aerial [built into the case] but not when the radio was taken out of the case and used with an external speaker.

Turned out to be a failed decoupling capacitor in the detector - with this failed, enough IF signal was passing through both audio-amplifiers, which had enough gain at IF, and coupled through the self-capacitance of the output-transformer to make the speaker-wiring 'hot' at IF! When this wiring [and the speaker case] were close to the frame-aerial the radiated IF got back into the front-end and it squealed.

Replacing the failed 100pF decoupler in the detector pretty much cured the problem but you could still hear a change in background pitch of the 'sharsh' from the speaker when tuning across the IF's second and third harmonics so there was still some residual feedback which I cured by earthing one side of the speaker-transformer secondary.

[jonnybear]

Have you replaced the smoothing cap C2, another culprit for motor-boating on these philips radios is capacitor C24, a .1uf 400v.
John

[yestertech]

The value of R8 is close to spec. but the voltage at the grids is too high.
It's therefore possible that one or other valve is losing emission, as there is not cathode resistor in either to be suspect.
Try tacking in a 0.1 cap from the valve side of R8 to ground to see if this cures.
Also check ( by substitution ) that the 47nF cap which decouples the AGC line is good.

Andy

[Silicon]

The voltage levels specified in the service manual were taken with a 20k Ohms per Volt meter. It would have an effect when measuring the control grid voltages.
However it won't have much effect on the screen grid circuit.

You might be able to simulate an AVO by putting a resistor across the input of your Fluke meter. The 25v range would have a resistance of 500kOhms.

C24 should be replaced.

[Mach One]

Thank you for all your replies. I will be looking into the suggestions as soon as I can get time to have a look.

With regard to the Avo "impersonation" with a resistor, I did think of that, but decided that the appropriate resistor would depend on the range setting, and as it is 25 years or so since I have used an Avo I had no idea what the ranges were. Thank you at least for the resistance suggested!

[PJL]

Originally Posted by Mach One

The reading at V1 g1 (pin 6) might be the AGC kicking in with the motorboating?

You are absolutely right - your voltage readings can't be trusted when the set is motor-boating from IF instability.

[Harry40]

Hi
Have not seen the circuit for this radio but IF RF instability was frequently caused by a faulty decoupler, usually on the screen. Some models used a common screen dropper and if the decoupler failed there was direct feedback between the stages.

[Mach One]

This is actually the continuation of a thread which I opened 4 years ago and which has been closed. If you are interested in what was previously discussed, here it is:

https://www.vintage-radio.net/forum/...d.php?t=110813

This Philips 531A radio has been sitting waiting for further attention in my loft for these four years. My children are growing up and leaving home and I have finally moved it downstairs into what is now my "office" and am trying to get it fully working again.

Thank you to all the people who replied to the previous requests for help. I replaced the suspect de-coupling capacitors mentioned in the previous post, including the one on the AGC line, but the set still motorboated.

With a piece of wire in the antenna socket I could stop the motorboating by tuning into a strong radio station. But the sound was as though the whole set was about to go into instability - rather like the old TRF sets with the reactance turned up too high. So I was left with one final option, that was the supply decoupling capacitor C2. This was still the original one. Since the set was also exhibiting a fairly prominent 100Hz hum I suspected that one of the two in the can was not working at full strength at least. Paralleling C1 with a 50µF capacitor made little difference to either the hum or the motorboating. However, having tried the same parallel capacitor across C2 stops both the hum and the motorboating. So that looks like success.

There is but one fault left on this set, as far as I can tell. Once it is fully working I might find something else... That fault is the very high screen voltage (about 170v instead of 54v) on V1 and V2 which are both fed via R8, a 56k resistor which is well within tolerance. I have changed the EAF42 to no effect but I have ordered an ECH42 to try it. If that doesn't work, I'm stuck.

[Sideband]

What meter are you using? A digital meter can give higher readings than an Avo which was probably used for the original readings (much has been mentioned about this in many posts here in the past) . What circuit are you referring to for the voltage readings? If it's a Trader sheet it's possible that the readings given on the circuit are wrong....it happens!

A point to bare in mind, it isn't always good practice to just shunt an electrolytic with another. If the original is leaky, shunting it with a good one may do nothing. If it's open circuit then of course it doesn't matter.

[Mach One]

The meter is a Fluke. I am aware of the loading effects of meters, but would not expect that the voltage would be effectively three times the Philips figure (yes, it's the proper service manual). Someone on a previous post said that even the Fluke has a loading effect.

Yes, the paralleling of capacitors was simply to see if there was any effect. It's not easy in this set to simply disconnect the reservoir capacitors and to avoid any unnecessary damage I chose the easy option.

[Nuvistor]

High voltage on the screen grids, is that with a signal tuned in, if so the AGC will reduce the current through the valves and raise the screen grid voltage.
How does the service sheet state the voltage should be measured, under signal or no signal conditions?

[Mach One]

The voltage table says a signal of around 150mv at 857kHz. I would think that 150mv is quite a strong antenna signal so I have compromised on the local BBC Five Live 693kHz transmissions which are very strong here and a length of wire in the aerial socket. The (rather dim) magic eye is fairly near fully closed. The voltage quoted for this condition is 54v. I will try and measure the voltage with the set tuned in and not tuned in to see what the difference is.

[Mach One]

Sorry - just re-read the instructions. The voltage quoted are for no-signal conditions. The other condition is for measuring AVC voltage.

[Nuvistor]

Have to see what the voltages are under no signal conditions, I expect they will be much nearer the service sheet numbers.

[Mach One]

145v... (no signal)

[Nuvistor]

What meter sensitivity is used in the manual for voltage readings? It could well be an AVO 7 with 1000 ohms per volt. If that is correct it will have a much greater loading than a Fluke.

Low emissions valves or high AGC voltage will also have an effect.

Looking back Sideband mentioned this earlier.