MX-110 MPX Matrix Issues

[Geek]

Hi Folks,

Gotta MX-110 on the bench here (Ser. 188Z6) with a unique issue I have not encountered before.

Problem:
All (including FM Mono) works awesome. MPX position delivers a mono L-R (left minus right, the info contained on the 38KHz DSB subcarrier) to the speakers. There appears to have no matrix decode of the L+R/L-R.
(frankly from the schematic, I don't see a differential amp doing this at all)

What has been checked:
- Tubes, all tested and the creme of the crop in there.
- Parts checked, replaced as needed. Crystal diodes surpressing the 38KHz local carrier very well.
- Aligned as per service manual. Tuning coils do what their supposed to do (used a scope rather than a VTVM specified. Reasonable Q-factor on the coils).

Did a forum and Google search and it seems an issue here and there with these, but no resolution other than "Take it to a shop", LOL!

Cheers!

 

[dcgillespie]

Because this is a time-division MPX decoder design, it is virtually impossible to have only a decoded L-R portion of the composite signal appear at the outputs of the diode de-modulators. If it were even possible, it would actually present as a L-R signal (only) in the left channel, and a -L+R signal (only) in the right channel. So if this is all that's being output from the decoder, then shorting the two outputs of the decoder together should yield zero sound -- this because you are surmising that the main channel L+R information is already missing from the output, so shorting the two opposite phase sub-channel signals together will cancel them out, leaving zero output from the decoder. I would seriously doubt this is what's happening.

As with any properly operating decoder, shorting the two outputs together produces mono operation by canceling out the two sub-channels, killing the stereo noise associated with them, and leaving only the L+R modulation in both outputs. In the MX-110 however, selecting the FM position takes a de-emphasized signal directly from the discriminator output, so using that position to test for mono operation of the decoder is not possible. As a first test then, I would select the MPX setting, and then short the two outputs from the decoder together -- this to test your basic assumption that the L+R is missing. If it is true, then it should result in no sound output. Most likely however, you will still hear the main channel L+R modulation.

In the MX-110, the phase adjusted pilot signal is separated out of the composite signal, and appears at the plate of V501A. It is then applied via T1 to the trigger stage (V502A), that then acts to lock/synchronize the 38 kHz osc stage (V502B) into the correct lockstep phase with that of the original suppressed sub-carrier signal back at the transmitter. The remaining L+R and L-R portions of the composite signal are filtered to remove any supersonic SCA components, and applied to the CT of T502's secondary. The L+R signal passes right on through the secondary and diode demodulators to appear at the decoder's output. The 38 kHz suppressed carrier L-R subchannel however is detected, and then automatically matrixed with the L+R signal it is riding along with, so that 2L and 2R appear at the output of the decoder. It is because this type of injection method is used for insertion of the composite signal into the demodulators, that it would be virtually impossible to have only a L-R or -L+R signal appear at the decoder's output.

Finally, if you are using the MPX indicator lamp to assess the correct tuning of the 19 kHz circuits, that would be a mistake. It is entirely possible with the Mac decoder to peak T501 so that the MPX lamp triggers faithfully as it should, but while it is glowing brightly, the sound produced is absolutely mono from both outputs. The usual cause of this is that the phase of the 19 kHz pilot signal applied to the trigger stage (adjusted by L501) is out of line far enough to prevent any significant separation from taking place.

If you are certain that the four demodulator diodes are good and installed correctly, and the MPX lamp is triggering correctly as it should, then check that the L501 resonant network is properly adjusted to produce the proper cancellation between the L+R and L-R signals.

I hope this helps!

Dave

 

[Geek]

Hi Dave,

That's exactly what's happening. If I switch the stereo mode to "Mono L+R", the output is null.

I've ran the core of both L4 and L5 top to bottom of the coil form, no change other than lose the signal entirely. Did this as part of the pre-alignment "Q" test to see if they were actually working. L5 of course with 67KHz fed into TP2.

if you are using the MPX indicator lamp to assess the correct tuning of the 19 kHz circuits,

Nope! I used the oscilloscope.

I verified the history of this unit and two other highly experienced techs just said, "I dunno?" on this particular unit.

See why I are stumped?

 

[Geek]

Was confused by your term L501, here is the schematic of the section in question.
Cheers!

 

[dcgillespie]

Yeah, you're working on a Z model. I was looking at the M model schematic -- it's the same MPX unit in both of them, but in the M model L501 is called just that, whereas in the Z model, they label it as "Phase". Same coil, just a different label.

So you are leaving the input selector switch set to MPX so you know the unit is sourcing its signal from that unit. OK. With that understanding then, to prove out your theory, set the Mode switch to Stereo, and remove V12 (the 12AU7 trigger/oscillator tube in the MPX section). With it removed, there can be no detection of the L-R sub-channel, so if it is all that is being reproduced in the stereo setting, then all sound should disappear with V12 removed.

Also, is thee any chance that the Phase Reverse switch has been activated?

A two pipe problem as Sherlock would say.........

Dave

 

[Geek]

Got it!

I unhooked L5 and the problem is solved.

Despite it tuning 67KHz with a decent null in the trap, it was shorting the signal otherwise (flopsied turn, probably. Seen this on ham radio 3MHz Heathkit I.F. coils)

Cheers Dave for helping me to un-overthink this as a modern matrix and see it as a ye olde fashioned diode ring product detector

Now to find a sub for L5.... with the amount of vintage TV horiz. osc and convergence coil forms, I may just be able to get er' done tonight!

 

[dcgillespie]

Super job! I've also seen the series cap associated with L5 cause issues as well, requiring replacement to resolve noise issues it was causing. Both really kind of strange issues..........

Dane