Quad FM3 no stereo

[CRM-114]

Hello everyone, I could do with some help with a recalcitrant FM3.
It's one of the last series with the CA3089 and MC1310 chips.

The short:
Tuner works fine on power up, all lights work, stereo beacon on, good sound through Quad 33.
Then, sound fades, distorts, beacon goes off and very low sound. (switch to Mono on preamp and sound is loud and strong. Also the tuning lights indicate reception).

This occurs at every switch on, and the time 'in stereo' becomes shorter the longer the tuner has operated. by this I mean if the FM3 has been off for a week I might get up to 3 or 4 hours of stereo before it fails, but if turned on the following day it will only give 10 mins. Eventually the time shortens to just a split second at turn on.

The long:
I am at my wits end. But I should add; radios are not really my thing, preferring early CD players and valve amps.

This is what I have done so far, with increasing desperation. Firstly the basics were checked, mains filter caps replaced, audio supply cap replaced, antenna socket thoroughly checked.

Thinking it was some sort of power supply failure, looked to replace resistors in obvious places, such as around MC1310. No change.

Replaced both IC with new ones, no change.

Because tuner works in Mono, started looking 'up the chain' (maybe signal is too weak for decoder to trigger ?), such as R14, 15, 19, 20, R25, 26, 33, 102, and caps C22, 23, 25, 26. no change.

replaced all caps and resitors around MC1310, no change.

So widened scope further 'up the chain', but cautious not to adjust anything lest the set becomes 'unaligned', only changing caps and resistors. no change.

Eventually, all resistors and caps were changed on the 'front end and tuning board". (in some instances the tuning varied with some small value caps, so they were re-instated after observing no change in behaviour). The only resistors not changed are R29, 31, 3232 and 35, all to do with the tuning lights, which operate correctly. And the audio output , of course, as that is not relevant.

I have taken some voltages before and after failure around both chips, nothing changes significantly. (other than the 'beacon on the MC1310, but this is a symptom rather that the cause, I would think).

I even took the F1 filter apart, and resoldered the ceramic crystals , but no change.
Cleaned the variable condenser, no change

AArrgh

Thank you for reading so far.

I hope someone can help.

[PJC58-Hythe]

You haven't mentioned what serial number the unit is.

There are 3 different circuit variants each dependant on the serial number, it's difficult to help you if we don't know which circuit you are working from.

All the best

[Graham G3ZVT]

MC1310 drifting out of lock?

Decoupling cap on pin 14 leaky?

[Craig Sawyers]

Originally Posted by PJC58-Hythe

You haven't mentioned what serial number the unit is.

There are 3 different circuit variants each dependant on the serial number, it's difficult to help you if we don't know which circuit you are working from.

All the best

There is only one version that uses the chips that the OP lists

Craig

[Craig Sawyers]

Originally Posted by CRM-114

I have taken some voltages before and after failure around both chips, nothing changes significantly. (other than the 'beacon on the MC1310, but this is a symptom rather that the cause, I would think).

I even took the F1 filter apart, and resoldered the ceramic crystals , but no change.
Cleaned the variable condenser, no change

AArrgh

Thank you for reading so far.

I hope someone can help.

What you are doing is a process called "shotgunning" - which is to replace parts until it works. Or in this case still doesn't.

What you need to do is use logic and measurement to diagnose what the problem is and find out what the solution is.

You clearly have a meter to measure DC voltages. Do you have an oscilloscope, and means of measuring frequency (like 19kHz)?

Have you downloaded the FM3 service manual and MC1310 decoder datasheet?

Craig

[alanworland]

My FM3 would function fine for a (very) short while then stereo light would go out with resulting mono output.
Changed the decoder for a nos item - still the same! Replaced it for another and it has been fine since.

Alan

[Craig Sawyers]

I suspect that might tolerance on the device parameters. Which is why pin 10 on the decoder is the 19kHz free run signal; you adjust RV100 until a 19kHz square wave is measured.

Craig

[CRM-114]

Thank you all for taking an interest.

Some quick answers:
Peter, SN 21567

Craig, I do own an oscilloscope, though not very competent in it's use. (mostly to check eye patterns on CD players). It's a dual trace TTM 303A, with a final setting of 0.5 uSec. (I can't a reference to max frequency).
I visualised the 19 kHz on pin 10 and also a repeat around VR100. So it seems the chip (s) is ok.

I condensed my initial post to keep it 'readable', sparing you the permutations (I'm on my 3rd MC1310 for example), haha. I'll be more complete in my responses from now on.

Yes, Craig, I agree with you; this is not the way I would operate normally, but having concentrated on the MC1310 to begin with no result I expanded my search up the chain with corresponding desperation ;-).

Graham, C112 was the fist thing I did, but thanks.

After replacing components around the MC1310 and informing myself a bit more on how FM tuners work, I thought the MC1310 wasn't getting enough signal to trigger decoding, and switched into 'low signal mute'. This seemed (to me) confirmed by the FM3 separate feed for MONO that was still present when the Stereo failed.

I then played with VR3, no change, so returned it to initial position.

Replaced TR6 with a BC549, no change.

Then replaced CA3089, no change

I have not touched L6 and L7, but as I see it, when it's working, it's...well, working. so the inductors must be in the right place. When it drops off it does so relatively quickly (a couple of seconds from full on to (almost) nothing.

There will be a slight lag in my responses, as I'm on the other side of the planet ;-)

thanks

Philippe

[CRM-114]

Some pics of the 'state of play'

[CRM-114]

Some measurements
IC1 (CA3089)
Stereo
1 - 2.02
2 - 2.02
3 - 2.02
3 - 0
5 - 0
6 - 5.9
7 - 5.95
8 - 5.45 (audio 'pop' when touched)
9 - 5.45 (dim audio when touched)
10- 5.45
11- 12.34
12- 0
13- 2.83
14- 0
15- 4.48
16- 0

mono
1 - 2
2 - 2
3 - 2
4 - 0
5 - 0
6 - 5.6
7 - 5.6
8 - 5.51
9 - 5.52
10- 5.52
11- 13.8
12- 0
13- 2.7
14- 0
15- 4.6
16- 0

IC 100 (MC1310)

Stereo
1 - 10.7
2 - 2.5
3 - 5.5
4 - 2.25
5 - 7.28
6 - 0.94
7 - 0
8 - 2.36
9 - 2.25
10-1.73
11- 2.36
12- 2.36
13- 2.36
14- 3.36

Mono
1 - 10.46
2 - 5.95/6
3 - 9.63
4 - 6.05
5 - 6.1
6 - 13.1
7 - 0
8 - 2.3
9 - 2.3
10-1.58
11- 2.38
12- 2.38
13- 2.38
14- 3.16

(obviously the 13V on 6 is the beacon 'off', as the stereo light is fed from the negative rail)

Also, for completeness, I did adjust RV100 after each MC1310 substitution. Adjusted by tuning to known strong stereo station and placing VR100 in the centre of the 'on' light. (the service manual suggests this is adequate procedure).


Philippe

[CRM-114]

Allan, 50% fail on new part, this is a worry.
well, waiting for a HA1156 in the mail, so we'll see ;-)

[CRM-114]

oops, I don't know what I'm talking about, 13 volts on pin 6 is just 13 volts

[Craig Sawyers]

The only thing I might suggest is the remaining tantalum capacitor that you have not yet replaced around the decoder - C103. According to the datasheet that is the filter capacitor for the stereo switch level detector.

It may well be OK. But since you already seem to have replaced C100 and C104 with plastic film, it might be a good idea to replace C103 too. If that is leaky or misbehaving in some way it could screw with the level at which the chip detects stereo.

Its value seems to be non-critical. Datasheet says 0.25uF and Quad fitted 0.68uF.

Craig

[CRM-114]

OK, thank you Craig.
I tried the substitution of C103 (the tant by Quad was 1uf), with the same value plastic, and unfortunately, no change in behaviour.

Coincidentally, my HA1156 arrived today, so I plugged it in (Quad conveniently provided a socket) but alas, no change either. Operated in stereo for about 10 seconds....

I have a question or two if you please (mostly to parade my ignorance to the world):

I don't understand why, if the mono - stereo division occurs immediately at pin 2 of IC 100, the mono audio is so clear and strong? If the signal strength is so diminished that stereo decoding is lost, surely this would be perceptible in the mono output? Is there something else in the demodulation that is required by IC100 to work. If so, I wonder if this is being stripped out somewhere.

Reading up on the chips and FM generally, and it seems Quad have a singular implementation of the CA3089. I can see they preferred to keep their tuning light setup (which I like a lot, simple and intuitive) and eschewed the meter and muting in the chip. How is the AGC achieved ? Simply within the CA3089. (pin 15 is not connected) Is Tr6 also an amplifier? I'm looking for where the signal could be diminished.

Also, F1, the 10.7mhz filter, if that is faulty, could that give such a symptom?

Trying some logic and deduction ;-)

Thank you

Philippe

[Ted Kendall]

The FM3 has a dedicated mono output, if memory serves, and it may be this you are listening to when you select mono on the 33. In any case, the degradation in s/n ratio of stereo reception over mono is about 20dB, because stereo requires a wider bandwidth (53 kHz) and the stereo information is amplitude modulated.

[nutteronthebus]

I have weird faults on the FM3 and it turned out to be the earth link between the 2 PCB a dry joint .I may be wrong but it's worth a look

Dave

[Mike Phelan]

The last one I had ended to be the MC1310 itself.

[frsimen]

When the fault is on, the voltages you measured on pins 2 and 3 of the MC1310 are a lot higher than when the decoder is working. Pin 2 is the input to the decoder and pin 3 is the output of the first amplifier stage. Assuming pin 7 is connected correctly, one possible cause of this external to the MC1310 is base-collector leakage in TR101, or something causing such a leakage. I suggest you check around that transistor.

Paula

[PJL]

Terrible design that relies on the bias and base current for TR101 coming from the MC1310 input pin 2.

[Craig Sawyers]

It is not such a bad idea - pin 2 goes high on mono, resulting in about 1.4mA of collector current in TR101. Quad fitted several types of transistor depending on availability, with BC550C being an equivalent. That has a beta of 400 - 800. So it will pull a pretty insignificant 3.5uA to 1.75uA from pin 2 on mono.

Having met the guy who designed the FM3 (the late great Mike Albinson) he would have thought that through long and hard before deciding on that way of doing mono.

Craig