Philips 543A problem detector stage

[Christoffrad]

Thank you PJL and Paula.

Further investigation on Thursday then

C

[PJL]

A final IF transformer fault does make sense. This set has a simple AGC taken from the audio. If a strong signal overloads the IF valve but the signal is suppressed in the final IFT it will not make it back to the AGC to correct the overload. We will see...

[Radio1950]

Standard Type Domestic Receiver Sensitivity Tests

Short Answers

“Sensitivity” is RF input level for say 50 mW AF output.
“Signal to Noise Test” is RF input level for 10 dB change in AF output with Sig Gen MOD ON/OFF.
“MDS” is Minimum Discernible Signal, or lowest RF level that can be heard with a “tune through”.
“AGC Test” is RF level at which AF output level rises by say 6 dB above that AF output level which was observed for 10 dB S/N.


Long Answers

Not intended to be exhaustive, and steered towards domestic receivers.

By “standard”, I just mean “widely accepted for test and comparison”.

Note that manufacturers may not publish test figures for BC AM sets, but you will get to know what is acceptable for domestic receivers. Sometimes a sensitivity specification is given in uV/Metres, which may have interpretation issues in the workshop.

The following has many versions, and ten times as many opinions.
Apres moi, le deluge.


Sensitivity

This is the input RF signal for a specified audio output, of a receiver in “linear” mode (your receiver).

Connect Sig Gen approx 800 KHz, 100uV P.D., 30% AM Mod 400Hz say, directly to Receiver AM Antenna input terminals, (ie electrically, not via eg inductive coupling to ferrite rod).
You might use a 0.01 uF capacitor to connect the hot lead, Sig Gen coax cold lead to Receiver chassis.
Consider all safety aspects before doing this, especially with live chassis receivers.

Listen to the Sig Gen signal via loudspeaker, ensuring no BC station interference, otherwise shift the Sig Gen Freq slightly.

Radio ON, VOL up a bit, TONE controls adjusted for minimal audio effect, else at mid range, AGC ON (for now), TUNE in Sig Gen and fine tune peak for max audio level on an AC voltmeter across an audio dummy load in lieu of speaker.

Find the Sig Gen RF level for 50 mW AF output.
This is 0.5 Volt AC RMS, or 1.414 V Pk-Pk across an assumed 5 ohm resistive dummy load, and assumes your voltmeter is OK at 400Hz. If you don’t have a suitable meter, 50 mW tone is “moderately loud”.

For comparison, my Classic Australian receiver, a HMV “Little Nipper”, has a sensitivity of about 10 uV.
The Little Nipper has no RF Amp stage, and has the classic signal valve line-up of 6BE6, 6BA6, 6AV6, so should perform similarly, RF wise, to a lot of receivers of this era.

Levels of 5 to 40 uV should be OK for general AM BC listening with close broadcasting stations, not but not for DX.

This RF input level of 10 uV is likely to be below the AGC onset, and so should not affect this test.

If we had AGC active from say 1 uV (for this specific receiver), you would not hear receiver background noise normally, it would be suppressed, and the receiver would not have that listening “tuning feel”.

Some people leave AGC ON, others eg with Coms Receivers, leave it OFF. It should have little effect at very low RF input levels.


Signal to Noise Ratio

(If your Sensitivity Test had a lot of noise, then this test will confirm that a problem may be present).

With setup as above, find the Sig Gen RF level for which you notice a 10 dB power change on the audio output meter, when you switch Sig Gen MOD OFF and ON again.
This is a 3.1 times relative voltage change.

My Little Nipper needs about 20 uV input for 10 dB S/N.

The expression and test for “SINAD” and "S+N/N" is rather irrelevant in this discussion.
Be aware that there variations of test description, but ignore it for now.


Minimal Discernible Signal

With setup as above, find the smallest Sig Gen RF level with AM MOD ON that can be just heard as you “tune through” the Sig Gen Freq.
Tune back and forth to avoid “preconditioned listening effect”.

My Little Nipper can be just heard at 4 uV input.

This is not exactly the classic Radar MDS, (whence it was derived) but similar, and very useful as a quick guide to receiver overall function.

Repairers will do this test instinctively to ascertain a receiver’s general tuning and performance.


AGC Test

With set up as above, set Sig Gen RF level to that level required for 10 dB S/N, AM MOD 30% ON (there is a reason), set the volume so there is a readable meter deflection to start with.

Ensure receiver AGC is not disabled.

Reduce Sig Gen RF level until you can sense by meter needle and listening, that there is no audio level clamping, ie that the audio output level is not affected by substantial changes in Sig Gen RF level.

If in doubt, set Sig Gen level to half that of RF level for 10 dB S/N.

Adjust VOL, and output meter range, so that the meter reads say about one third of FSD, or to 1.0 V on a 5 V AC FSD scale, say.

Ideally, use an olde fashioned VOM with an AC Volts range and meter scale marked in dB.

Increase Sig Gen RF level and record the RF level for which the Audio Output level rises by say 6 dB, or a relative voltage increase of 2.4 times.

If your setup sends your meter off scale, readjust VOL or the meter range.

My Little Nipper seems to have a rise of 6 dB audio for a rise of RF Level of 10 to 1000 uV.
(Leave MOD ON to observe unusual AM detector effects with faulty components).


CRO Monitoring

It would be wise to connect a 10x probe and CRO across the AF dummy load to look for anomalies as you do all the above tests.
And monitor the AGC DC Volts with a CRO on Vert DC Coupling, using a 10X probe (some slight loading effects may be seen).

Speaker Monitoring

It would be wise to connect some form of test speaker at reduced volume, for all the above tests so that you may hear funnies, whistles, distortions, interference, and other anomalies.
.

[Radio1950]

Just for completeness.

In my post #105, I asked for a "dummy PDF" of a recording.
This is just an audio file, but with edited extension, for a file type eg PDF, for which the forum has max file size allowance.
But you have zipped it in post #107 with the same result, and it works.
.

[Radio1950]

I had a listen to the audio in post #107, and it may be due to overload somewhere, as Peter_Sol says.

I'm suffering from grand kid overload here today.
Back soon.

[Radio1950]

In post #117, you do some tests on AGC, and thereby sensitivity.

Those input levels are all very high, unless you mean "microvolts" when you write "mV"?
If you actually meant millivolts, then there is a definite problem.
We need to quantify the problem.

Most domestic receivers will have input RF levels around 10-20 microvolts to start to generate some AGC.

I would expect AGC to start at 15 microvolts and fully flatten out by 5 millivolts.

This is sounding like a deaf receiver.

People are imploring you to carry out an AM RF and IF alignment.

Irrespective of its present condition, please carry out the alignment as per Trader or Philips instructions, so that we all know where we stand.
While you are setup for this, you may as well also check the IF sensitivity injecting at pin2 (G1) of the AM mixer V2.

I can understand that you may not want to do this just yet, as it seems illogical to you, before faults are fixed, but we are viewing and surmising at a distance, so for us, it is essential.

After alignment, please report back with RF and IF sensitivity, and AGC, test results.

This receiver seems to me to be a high end domestic receiver, and even without a RF amp on AM, it should have good test results.
(The IF amp chain will contribute most to overall AM reception gain anyway).

In the meantime, we can again read your posts and pore over the circuit.

Don't give up.
You can fix this.

Good luck.

[Christoffrad]

Thanks for detailed notes R1950. I have done sensitivity and SINAD measurements when designing receivers for Philips but these were narrow band FM with sensitivities around 0.7uV for 20dB SINAD but have little familiarity of what to expect of AM broadcast Rx so this is helpful. The reason I've been reluctant to 'twiddle' is because of those instances that I got it to work after just touching the 1st IF connections suggesting a component fault rather than alignment as such. Also knowing the delicacy of IF cores on some Philips sets. (eg 141U)

Also thanks Paula for your tests. Its a pity the cap is buried inside the IF can!

C

[PJL]

The test to see if a faulty cap could recreate the symptoms was only a possible fault condition as other faults could result in losses in the IF amp. You need to do the realignment to see what's going on looking out for adjustments that don't peak.

[Christoffrad]

Before I start this tune up here are some initial conditions according to the notes from Radio1950:

Generator set for 800kHz with 400Hz tone 30% mod
4.5ohm resistor as dummy load. Volume control set to mid point.
measurements taken with oscilloscope.
Sensitivity : For 1.4V P-P measured across dummy load ...generator must be set for an output = 90mV emf
Signal to Noise : with generator disconnected there is no measurable noise ie <1mV even at full volume setting.

This radio is as deaf as a post. (hence my original reported Low distorted audio!)

Now I'll start the tuning...

C

[peter_sol]

Measure the resistance across the AM if transformers with the fault present. If the cap is S/C that will show up against the other windings. If the cap is O/C you could just tack another across it and monitor the results.

[peter_sol]

110pf and 195pf are typical Philip's odd values it 100a d 200pfwill prove the point.

[PJL]

No...

Follow the IF alignment instructions in the Trader or Philips service data (they are the same). The instructions must be followed exactly. You can put the scope across the speaker as that will help with detecting the peak. Adjust the signal generator output to obtain a quiet listening level and repeat the output adjustment to keep it that level.

[PJL]

Originally Posted by peter_sol

110pf and 195pf are typical Philip's odd values it 100a d 200pfwill prove the point.

There is no evidence that the capacitor is faulty, it was just a test to simulate the distortion. Lets see the results of the alignment first.

[PJL]

Just to explain, R1950 is proposing a series of tests but they are not for IF alignment. The service data has a very specific IF alignment procedure that will be designed to set the right bandwidth. If the IF is repeatedly peaked the bandwidth will be too narrow and the set may become unstable. You only need to do the 'AM IF Stages' (Trader) alignment at this stage and we would not be surprised if it fails with an adjustment that does not peak as there may be a fault in the IF amplifier. If the IF aligns OK with very much improved sensitivity then you can move on to the remaining AM alignment steps.

[Radio1950]

I am actually just suggesting a receiver alignment as per Trader/ Philips, and two tests only, an RF and a IF sensitivity test, to start with.

This gives us a rough idea of a few things, and something may turn up during alignment.

[Mr 1936]

Hi

I have been thinking about this (again !).

If there is a fault giving significant attenuation mid-way down the signal chain, the following will happen:

Weak signals will be attenuated into the noise and produce little or no audio.

Medium strength signals will overload the IF amplifier stage, because the AGC action is effectively disabled

Strong signals will begin to produce a little bit of AGC action, but not enough to bring the IF amplifier stage out of overload.

When an active stage overloads, the non-linearity could be enough to change the bias conditions, i.e. no longer working in Class A. When carrying an amplitude modulated signal, the overload will be modulation dependent. The R/C screen and anode bias decoupling (mostly intended for IF decoupling) may be only partially effective at audio frequencies, and its effectiveness will tend to be less in the lower audio range. I haven't yet looked at the time constants involved. This might cause the bias to vary at an audio rate, and explain the observed effect of bass frequencies appearing to modulate the treble.

The finger of suspicion points to a fault causing appreciable attenuation in the area of the IF amplifier anode and the following IFT.

[Radio1950]

Here's a systematic generalised approach for this specific receiver, which should get close to the problem area.

Philips 543A With Low Sens and AF Distn on AM.

General systematic approach for analysis is to pre check a few components, remove any valves which are not used for AM IF or Det, and to disable FM IFTs as reqd, Inject a test signal to IF amp, measure Det AF OP, then work backwards to the Antenna.

Using Philips Service circuit references.

Preliminaries

(The following procedures ignores any switching contacts; you may have to repeat some of the following, after switch cleaning and "contact exercising")

Disconnect any possible external test antenna, and other cabling.
Remove V1, V4, V5, V6.
Disconnect the “detector” end of C66 (coupler, AM DET to AF), and R58 (coupler to tuning indicator) if fitted.

Tack temporary 0.01 mf caps, or DC shorts, across all FM IFT windings, including S30 and S31 (FM Det), as you may not know which winding is Pri or Sec.
Take care here.

Measure and check if both ends are correctly wired – (DET, AF, AGC ) R17, R18, R39, C47, C49, C76, (AGC Mix bias) R5, C18.
Unsolder one leg and measure immediately; eg If you can, slightly heat each component just before test.

Take your time; ignore 10% variation of values. Measure caps for leakage and capacitance.
Measure cathode resistance to ground of mixer and IF Amp/Det valves; should all be zero.

Leave AGC line alone for now.
You may have to disable it for later tests.

Power Up.

Check DC Supplies.

Monitor with CRO and 10X probe at junction of R17, R39 (AM Det AF OP).

Ensure no received signals, tune RX between stations, confirm with CRO.

Measure cathode volts of mixer V2 and IF Amp/Det V3 valves, first for AC, then DC; should all be zero.

Measure standing bias on G1 of mixer and IF amp valves.
Should be about negative 0.4 to negative 1.0 V DC, with no other signal observed; check with high resistance DVM (or CRO with 10X).

AGC polarity is critical.
Be very sure about this measurement.
Investigate if not as expected.

Signal Checks

IF AMP and DET

Monitor with CRO and 10X probe at junction of R17, R39 (AM Det AF OP).

Inject Sig Gen 470 KHz, 50 millivolts PD (100mV EMF) AM MOD 30%, via 0.01 ceramic to G1 of V3 (IF Amp).

This should give a clean AF signal on CRO on DET AF OP, and some rising negative DC AGC volts at the same monitor point (eg switch CRO Vert to DC coupling and observe).

Rough tune AM IFT if required; you will have to realign later anyway.

If not, Power Down, remove V3, solder or jumper a temp earth on V3 pin 8 (DET cathode).

With Power Off, inject Sig Gen “max avbl” RF into V3 valve socket pin 6 (anode), and CRO probe junction of C47, R39, to observe IF signal of perhaps 75% of that injected.

Investigate and resolve if different result.

Remove temp jumper, insert V3, before continuing.

Repeat injection into V3 to get clean AF OP at junction R17, R39, and with AGC DC Volts as before.

Now remove any temp jumper shorts across the FM IFTs in this part of the circuit.
Confirm IF sensitivity test or fault find.

If OK here, continue.

MIXER

Monitor with CRO and 10X probe at junction of R17, R39 (AM Det AF OP).

Inject Sig Gen 470 KHz, 50 microvolts PD (100uV EMF) AM MOD 30%, via 0.01 ceramic to G1 of V2 (mixer).

This should give a clean AF signal on CRO on DET AF OP.

Rough tune AM IFT if required; you will have to realign later anyway.

If set tends to oscillate, slightly detune one IF winding of AM IFT1 eg S28, or dampen with 10K resistor.

AGC volts should be able to be measured (High Z) at V2 grid, and should vary with increasing Sig Gen level, going more negative, and staying negative (important) with a higher RF level.

This last measurement can be a mite troublesome, as you have both DC and RF signals, and your measuring device will have some loading effect.
Just ensure that you measure at the grid pin, and get an approximate result.


If you don’t get this mixer sensitivity result, remove V2.

Ensure you have a 0.01 uF capacitor in the Sig Gen hot lead.

Inject Sig Gen into plate socket pin 6 of V2.
You may have to increase the Sig Gen a bit here.

Now you have to investigate this part of the circuit, looking for faults.

Expect to have a sensitivity at mixer input of 10-30 microvolts or so, and a clean AF signal with just a small amount of noise maybe.

And keep in mind that with these tests, we are just brute force injecting, without regard for any matching, so variations will result. Ignore this for now.

Now remove any temp jumper shorts across the FM IFTs in this part of the circuit.
Confirm IF sensitivity test or fault find.

If OK here, continue.

ANTENNA INPUT and FILTERS

If the mixer and IF and Detector now work OK, follow the service sheet instruction.
Fault find as required.

If you really must have a dummy antenna, as per some instructions, just use a 470R resistor and 0.01 uF cap in series with your Sig Gen hot lead.
Not elegant, not perfect, does the job.

FINALs

After all faults are found, realign.
.

[Radio1950]

If my post #137 doesn't assist in pinpointing the area of the fault, I think you should still concentrate on the AM IF area, and any components here.
Everything you have reported points to this area.
Most of your previous AF CRO results from Sig Gen Tests look OK.

Others have made good suggestions re overloading.

My high suspect areas for now would be IFTs, AGC, switch wiring, switch contacts, high value resistors, funny capacitors acting not like capacitors.

You can make up a test cap with 0.01mF 500V ceramic cap and two aligator clips.
Use it across to test eg screen and plate bypasses, IFT windings etc.
Very handy. Not 100% conclusive , but points to possible problems.
Test all valve cathode earth jumpers to earth with this "test cap"; they may be dry joints to RF etc.


You will have carried most of the simple tests so far.
So we can only suggest, based on your tests and observations.

Good luck.

[Radio1950]

In my post #137, I forgot to include that at the plate of the IF Amp V3, with say 50 microvolts into G1 of V2 mixer, you should expect a few volts of modulated IF, a classic looking AM modulated wave on a CRO.

For higher level IF signals, the AGC loop will keep the signal presented to the detector stabilised to a few volts IF.

At the detector itself, you should then see a few volts of what looks like "half a classic AM wave" at IF freq.
This is because the AM detector is only a half wave peak detector.

Your CRO probe will detune the IFT signal a little.

[PJL]

I thought we had all agreed that IF alignment using the Trader (or Philips) instructions was the next step? It is possible that alignment is all that is wrong but we won't know until it is tried.