Philips 543A problem detector stage

[PJL]

Originally Posted by PJL

Were you using the scope on the secondary to monitor the primary adjustment? A 10pF probe has an impedance of 33K at 470KHz so that loading alone will mess up the primary when driven from 220K.

I was really referring to the problem you had peaking the IFT primary out of circuit.

[Christoffrad]

Yes peter I did conect the scope to the secondary directly when tuning the primary when the IFT was out of circuit.

Now today's findings..... we are certainly getting to the crux of it....Having re-installed the IFT into the set and connected the secondary to V3grid via the shortest possible wire and with no connection at all (on the secondary side) to the FM IF and it's switch, the set now has a sensitivity of 200uV for 300mV of audio (compared to the previous value of 4.6mV)

HOWEVER the set has become VERY unstable. It's almost impossible to hold the cores to a position without oscillation
Is this because the capacitors tuning the secondary are now mounted external to the can? See my thumbnail on post 154
Not surprisingly, if I try receiving a broadcast station it is apparent the the set is now nicely sensitive but howls as a station is tuned in.

[PJL]

That's great progress and I suggest the next step is to track down the fault in the wiring you have disconnected. Try measuring the end-to-end resistance from the disconnected IFT link to the disconnected V3 G1 wire and then to chassis. Most likely it is the waveband switch, have you cleaned it with Servisol or similar and how easy would it be to take the wafer out?

[Christoffrad]

End to end resistance is unmeasureable (< 0.1ohm) This includes the FM IFT by-pass switch.
The resistance between the disconnected circuitry and chassis is > 20M
There is no connection with any wafer switch on this part of the set. The switch designated S17 on the Trader sheet (F3 Philips) (which forms a by-pass of the 10.7 IFT in AM mode) is a completely separate switch that is operated via a chord linked to the mechanics of the FM selector key. (must have been designed by Heath Robinson!).
So this IF loading effect would appear to be not a DC measurable condition.

C

[PJL]

Either that or there was a dry joint, try running solder over all joints in this path and wire the 10.7 IFT back in. If you still have a problem then we can eliminate bit by bit by wiring around them.

[Christoffrad]

By measuring between various parts of the disconnected circuitry and chassis my LCR measures between 0.4pF and 4pF.
Whilst receiving a strong Off-Air signal, if I make any form of connection of the disconnected circuitry (eg touching back either end), reception is killed dead.

[frsimen]

I suspect that the greatly increased sensitivity is due to positive feedback in the IF amplifier increasing the gain and causing the oscillation. It is acting as a Q multiplier. The slight additional loading caused by the external circuitry is stopping the oscillation and the gain is back to normal.

You can prove the point by adding some resistance across the 470kHz secondary winding. 100k to 180k should do the trick, enough to damp the oscillation without wrecking the Q factor. I would expect reception to return to its previous level if the oscillation has stopped.

If the contacts of the switch are not making good contact they could perhaps acting as a diode. That could upset the operation. The IF transformer's secondary will swamp any resistance measurement you make to prove the switch, so disconnect one of the wires to the switch first.

If none of the above demonstrates anything useful, there is a lot to be said for opening up the IF transformer and checking that all the wires are terminated properly. Check the 100pF capacitor while you are there too. You will then be able to mount the new capacitors within the can and that may improve the stability.

Paula

[Christoffrad]

If I load the secondary with 180k the set is barely more stable as you tune through a station (and doesn't really reach stability even when on-station) but the sensitivity of the set is not harmed. Sensitivity (800kHz 30%mod) is something like 100uV for 1.4v p-p across 5ohms at near full volume but there is LF oscillation creeping in at this low RF level.
I just give this figure to indicate that the sensitivity is not harmed by the 180k. So I don't understand how the disconnected FM circuitry kills it if in any way attached.

[Mr 1936]

Hi

I am puzzled by the need to incorporate S17 to short out the FM IFT on AM, an arrangement I haven't seen before. Given that the inductor is about 7 uH, the parallel tuned circuit will be only about +j20 ohms at 470 kHz which should hardly matter. Clearly the designer found it was necessary, and it's hard not to wonder if this was a Heath-Robinson style addition as a result of some last minute panic.

You are not seeing any appreciable DC resistive or capacitive loading from this section which would damp the AM IFT. However, I'm just wondering if there could be any HT leakage from the mixer anode to the IF amplifier grid via the FM IFT windings and their insulating former ? OK, it seems unlikely but it's not impossible. Given 165 volts on the anode and a grid DC resistance of 1 megohm, even 100 megohms leakage resistance could still screw up the grid bias by a couple of volts, akin to the "that capacitor" effect.

I suggest that the way to eliminate this as a possibility is to leave the secondary of the FM IFT (including S17) in circuit but temporarily remove the primary of the FM IFT from the mixer anode and link across, thereby interrupting any leakage current path.

[G6Tanuki]

I wonder if your instability is caused by IF-signals getting onto the HT rail?

In times-past I remember dealing with a radio that had strange IF instability - it turned out that a failed capacitor in the post-detector filter circuitry was allowing IF signals to get onto the grid of the 1st audio stage, which amplified them rather well... and passed them on to the output-stage, which also amplified them rather well... [remember that a typical output-tetrode or pentode of the 6V6 or EL84-class can be used quite successfully as a power-amp at 3.5MHz].

The problem was compounded by the fitment of a 'tone-corrector' capacitor across the output-transformer terminals which happily passed the amplified IF signals onto the HT line. The electrolytic smoothing-capacitors really didn't do a good job of bypassing IF-on-the-HT-rail.

I solved the problem by adding 100pF capacitors between the grids-and-anodes of both AF valves and moving the 'HT' end of the tone-corrector capacitor to earth so it bypassed any IF signals to earth rather than to HT.

"Odd signals cropping up in odd places" is the bane of designers - and restorers!

[ms660]

Originally Posted by Mr 1936

I am puzzled by the need to incorporate S17 to short out the FM IFT on AM, an arrangement I haven't seen before. Given that the inductor is about 7 uH, the parallel tuned circuit will be only about +j20 ohms at 470 kHz which should hardly matter. Clearly the designer found it was necessary, and it's hard not to wonder if this was a Heath-Robinson style addition as a result of some last minute panic.

Shorting out the FM secondary would damp the FM primary which might otherwise turn V2b into a tuned anode tuned grid oscillator when tuning around 10 - 11Mhz on the shortwave band?

Lawrence.

[PJL]

Originally Posted by Mr 1936

I am puzzled by the need to incorporate S17 to short out the FM IFT on AM

Maybe to prevent short wave oscillation (10.7-oscillator frequency).

[PJL]

The loading from the 'wire' comprised 20M ohms and 4pF but the 4pF will simply shift the resonant frequency so 180K ohms should have stopped the oscillation if the underlying gain has not changed. The set is unstable probably because we have tacked on a replacement capacitor for the IFT under the chassis.

I am inclined to restore the internal IFT capacitor as the measured value may have been lower than the service data but we have not been able to verify the service data is correct by looking at the printed value on the capacitor. This might stop the instability. If it does and the gain remains high then we can reinstate the 10.7 IFT and try a shorted wire soldered across the 10.7 IFT secondary.

[Christoffrad]

Thank you very much for all your considered comments.
I will try Mr 1936's suggestion of removing the FM primary.

I can also take another look at the IFT and try to slide the can off in order to inspect the internal capacitor. It is rather stuck and I don't want to force it but I'll give it a go!

A couple of thoughts about the IFT secondary and capacitor....
1) the value in the data sheet corresponds with the calculated value for resonance, having measured the secondary inductance. So the measured capacitor in situ is 65pF too low theoretically.
2) It is strange how those occasions when touching the IFT connection with the scope seemingly restored correct operation for a short time. Also..I wonder why had the V2 anode feed resistor and decoupling cap been changed and for the wrong resistance value.

BTW the leakage resistance of the disconnected 10.7 items is greater than 20M. But my DVM top range is 20M.
Also regarding comment from G6Tanuki.. The instability is not affected by volume control so I suppose the audio amp can be eliminated. I have mainly been monitoring the detected audio on the scope. But I will bear in mind HT decoupling.

C

[frsimen]

Bearing in mind Lawrence's comments in post 192 and the enthusiasm for oscillating at 470kHz when only the 470kHz IF transformer is in circuit, is it possible that the circuit is oscillating at 10.7MHz?

You won't see very much evidence at the AM detector output if it is, but the voltage measured across the ratio detector's electrolytic capacitor (C56- Philip's ref) will be high.

The damping provided by the signal generator is sufficient to tame things, which is why the measurements at the grid of the EBF80 are always fine.

The problem may have been caused by someone tuning up the IF strip without the damping resistors as specified in the alignment instructions. If this is the problem, detuning the EBF80's anode 10.7MHz tuned circuit, while switched to MW or LW may well restore normal AM reception. With so much gain around, your new capacitors may have to be fitted inside the IF transformer can to maintain stability.

Paula

[PJL]

The implication is the IFT secondary switch is not working as intended. How about the simple test of soldering a link across the 10.7 IFT secondary?

The theory goes that disconnecting the 10.7 IFT completely stops oscillation but a single ended connection restores it and this could only be because it is grounded via coupling to the 10.7 IFT primary.

I still suggest the set is put back together by replacing the original AM capacitor and the 10.7 IFT wiring as we need both AM and FM fully working.

We need to consider that this set has had previous work done to fix this and the Heath Robinson switch arrangement may imply it was a Philips production fix too. We need to prove it is SW oscillation and may need to make some design/layout changes to fix it.

[Christoffrad]

Voltage on C56 (Philips) (C49 Trader) is a couple of milivolts.

Today I tried the suggestion of Mr1936. I disconnected the 10.7 primary from the V2 anode circuit completely and connected the anode wire to the AM IFT directly. The set now measured a sensitivity of about 2mV although instability made this very difficult despite IFT retuning. When I returned the set to it's previous state,(10.7 circuitry reconnected on the primary side) instead of being unstable it was now quite deaf.
With the 10.7 seconday circuitry still isolated I tried re-instating the original (130pf) internal capacitor and also tried supplementing it with 65pf (externally) to achieve resonance. .... all to no avail. The secondary of the AM IFT is still very flat so something seems to be loading it again and this time it is not the 10.7 secondary.
BTW when the IFT was removed I tried determinedly to remove the can, but without success.
Perhaps the fact that it is now so insensitive, even with the 10.7 circuitry removed from the IFT secondary, helps eliminate that element, but raises other questions. For example about the integrity of the IFT itself. Yet when I examined the soldered connections of the litz wire it looked good. I have also been careful to sleeve the disconnected internal capacitor leg.
This set is now getting beyond a joke ! I wish one of you lived near Cambridge.. You could have had a visit!!!

[Christoffrad]

Couple of further observations suggesting RF instability :
If I apply an RF signal (800kHz with 400Hz mod 30%) to the aerial there is a hetrodyne beat as I tune in. If I kill the power whilst tuned to a broadcast signal there is a momentary increase in audio level.

[Mr 1936]

Hi

I have dug into this question of switches on the 10.7 MHz first IF, and found this snippet from the manual for the Ferguson 383A (a typical British design of the same era).

"L11, the primary of the 10.7 Mc/s IF transformer is switched out of circuit by contacts on S1A to prevent any possibility of harmonics generated by the local oscillator reaching V3 control grid"

See attached schematic.

If we take the IF passband at this point as 10.5 to 10.9 MHz, it could pass (as just one example) the 7th harmonic of the LO in the range 1500 to 1557 kHz, corresponding to a medium wave RF tuned frequency of 1030 to 1087 kHz

Not sure if this helps or not, but clearly designers have had breakthrough issues in the past.

[PJL]

You can check the IFT windings by comparing the winding resistances with the 2nd IFT as they appear to be the same.

Put it all back together, remove the added capacitor, and put a shorting link across the 10.7 IF secondary just in case the switch is faulty. Realign the AM IF following the manufacturers 'LF Alignment' instructions again and the 'LF Filter' instructions that is there to help prevent IF instability.

The current theory is the IF is unstable so we need to confirm this and determine the oscillation frequency and cause.