Pioneer TX-9500II SIGNAL Meter Problem

[MarkyM]

I'm restoring a nice TX-9500II.
All has gone well and I am doing the alignment now. The distortion looks excellent and all appears well with the tuner itself except for the SIGNAL strength meter which shows 2 problems.

1) With a weak signal below the limiter threshold applied (about 10uV - no modulation) and the output of the IF peaked on my scope at Terminal 52 for true center channel, I can set the center-tune meter just fine with T12 to that point but the SIGNAL meter peaks in a different place (with the tuning slightly higher).
2) Also, it drops down when modulation is applied and with mod applied, it also peaks at a lower tuning point! Closer to the center point but off to the left a bit.

This all happens in both WIDE and NARROW IF modes because the SIGNAL meter is driven off IC Q15, a PA3001A, Pin 13 and the NARROW IF feeds this IC after the 4th ceramic filter (out of 5).

Note that I have replaced the last 3 NARROW filters with Murata 180kHz GDT filters, hand matched using a NanoVNA to the IF frequency of WIDE mode, 10.7035MHz.

Have any of the tuner gurus here seen this on one of these?
Is there anything I can do to fix it?

Thanks!
--Mark--

 

[w3jn]

Since the input/output impedance of a ceramic filter is in the range of ~300-500 ohms, did you connect a series resistor with both the input and output of the filters when you were testing them on your VNA, whose impedance is 50 ohms in and out? If you didn't, the capacitance and loading of the VNA and cable will skew your results. When testing filters or the passband of an IF stage you need to be very careful that your generator and detector don't load down the circuit under test, or you'll get completely erroneous readings.

Second, I suspect the L/C portion of the IF before and after the wide/narrow filters is out of alignment, the passband is probably asymetrical and thus the discriminator is aligned on the slope of the passband, rather than the center. Since you have a VNA, lightly couple its output to the 10.7 MHz IF in, through a 1K resistor or so. Then set the VNA for 10.7 MHz center frequency with a sweep of +/- 1 MHz or so - ie, 9.7-11.7 MHz. THen very lightly couple the VNA input to the proximity of the FM detector. Just a clip lead in proximity to the detector should be adequate, if you do a direct electrical connection I'd use a 1-5 pF capacitor or so in series. Take a look at the response curve, it should be symmetrical around 10.7 MHz with perhaps a slight sag in the middle. Do this in narrow and wide bandwidths and note any large differences in the center of the curve. The curve will be narrower with the narrow bandwidth filter.

I'd align the IF transformers for a symmetrical curve using the narrow bandwidth. Then, I'd figure out the center frequency, put the sig gen/VNA to CW (no sweep nor modulation) at that center frequency and adjust the discriminator for a center needle. Now apply modulation at an adequate deviation and see if the center needle changes, and tuning off freq results in the signal strength discrepancies you noted.

I just rediscovered why I hate Japanese equipment schematics so much.... jeez that schematic is a mess to trace through.

 

[MarkyM]

I'm using the MK test fixture with includes all of the proper accoutrements

Thanks for the tip for using the NanoVNA like that.
I'm not sure which IF transformers you are referring to. There is only 1 adjustment in NARROW mode, T16 which is adjusted for minimum distortion.. Everything else is for WIDE mode.
The metering is fed from a point right after the 4th ceramic filer in the narrow IF chain.

 

[w3jn]

1) With a weak signal below the limiter threshold applied (about 10uV - no modulation) and the output of the IF peaked on my scope at Terminal 52 for true center channel, I can set the center-tune meter just fine with T12 to that point but the SIGNAL meter peaks in a different place (with the tuning slightly higher).

Yes, the metering point for the signal strength meter comes from the narrow IF channel, and the PA3001 has its own discriminator and detector. T12 adjusts the center of the meter completely independently of the audio detector, and is not affected by T7-T11. T6 does not affect the meter either. The more I think about this the more I think you have asymmetrical passband or ripple on your ceramic filters, as the center meter as adjusted by T12 does not correspond with max signal strength. Also, the modulated signal is extending to the slope of the filter, as demonstrated by the fact the strength meter is decreasing on modulation peaks.

Easy to check the passband response of the entire narrow channel with your VNA. Just make sure you *lightly* couple both the input and output of the VNA to the IF channel. I think when you were selecting those filters the VNA loaded them down and you got a couple that aren't correctly matched - or, the circuitry in the IF channel has modified their passband somewhat.

Not sure why you replaced the old filters - were they bad? Could also be that your new filters are just too narrow for this.

The complete mess of a schematic aside, the service manual is pretty decent although the simplified block diagrams are also pretty hard to follow. I am stumped as to why they have a whole extra detector IC just to drive the meter which doesn't correspond to the actual center tuning of the audio detector, but I'm sure there was a reason.

 

[MarkyM]

Thank you,
I’m stumped too although that IC is used for the muting function as well.
The original filters were too wide. I’m using 180kHz GDT filters to reduce HD radio sideband self noise and improve selectivity. Maybe I need to match them even closer (although I don’t have very many left). I did try it with the original filters and it does the same thing although not as noticeable due to the wider bandwidth. Also, a couple of those original filters were quite a ways off from the others too.
The tuner is performing very well and sounding good except for this signal meter issue.
I wonder why they chose to pick off the IF signal after the 4th filter out of 5 instead of at the end where Terminal 52 is?

 

[dr*audio]

If it was me I'd use a spectrum analyzer to have a look at the IF passband and see where it's getting detuned, but your VNA should be fine to do that.

 

[w3jn]

They must have had a reason, who knows. I'm not a fan of that alignment procedure either. It puts the center tuning meter potentially at a different center than the audio discriminator.

You could try setting the thing to narrow, then align the audio discriminator using a 10.7 MHz signal at the mixer output. THen align those other IF transformers in the IF that are after both filter channels. THen, and only then, adjust T12 for zero center on the meter (using an actual 10.7 MHz signal instead of noise), and see how that corresponds to max signal.

But the easy thing is to run your VNA thru the channel and see what kind of IF passband curve you're getting, remembering that the circuitry of the IF channel will somewhat modify the passband of those ceramic filters.

Not sure 180 KHz filters will be narrow enough to eliminate adjacent channel HD noise.

 

[dr*audio]

"I am stumped as to why they have a whole extra detector IC just to drive the meter which doesn't correspond to the actual center tuning of the audio detector, but I'm sure there was a reason." I have seen this a few times, where they set up this second detector that is used just to detect the null in the noise to operate the muting or tell if the station is tuned in. It makes it a pain in the butt to do the alignment. Sony used it on the ST-J75 and there isn't even instructions in the alignment procedure for it, but if it drifts, the tuner won't pick up stereo or will mute, especially on loud passages or bass notes.

 

[MarkyM]

Yep, they use it for muting as well. From the service manual circuit description:
The FM control circuit used to control the meter
drive and FM muting circuits, is independent of the
audio signal circuitry. This circuit, too, incorporates
an IC (PA3001) recently developed by
Pioneer. The block diagram is shown in Fig. 8.
The PA3001 is very similar to the previous IC
(HA1137) in function and operation. It differs in
that the FM muting level can be controlled extemal
to the IC.

I have replaced the PA3001A with an HA11225 (the official replacement for it per Pioneer) and while the signal meter calibration procedure now works properly, the signal meter reading is still affected by modulation.
And it still reads lowest on the low-end of a weaker station's passband and highest on the upper side (on a real station with normal program modulation)

I did optimize the narrow mode filters again. I am using the 3 closest original 300kHZ filters for the the first 3 filters. The center-tune IF frequency for the WIDE mode is 10.7050 MHz (+5.0 kHz).
The 3 original filters measure +1.1, +1.2 and +9.4 kHz (the other two are +10.1 and +16.8)
The 2 180kHz GDT filters are +4.8 and +4.0 kHz. That's as close as I could get as I don't have many of these left and can't find any more :-(

My conclusion is that this is an original design choice/issue.

Yes, the 180kHz filters help greatly with HD sideband self noise. The real cure though for any tuner using a "conventional" switching MPX decoder chip would be a pure sinewave "product detector" MPX decoder like Yamaha used in the T-85, TX-1000 and TX-2000. Of course these came along a decade later than this TX-9500II. Even in Super-Wide mode, there is no trace of HD sideband hiss in my T-85, even for stations running in IBOC MP3 mode where the HD sidebands start at just 115kHz from the carrier.

 

[w3jn]

Wonder if you used a 250 ohm or so resistor in series with the input and output of those filters when you ran 'em thru your VNA

The S-meter decrease on modulation is probably a station pushing the modulation limits, I see that sometimes on my McIntosh MR80 in narrow IF, which has a crystal filter. I took a closer look at the stereo demodulator in the MR80, and the chip is interesting in that it takes into account the possible interference from the 57KHz data subcarrier.

Any R/C VCO will have a bunch of phase noise. It looks like the T85 has carefully avoided that by using a crystal oscillator for the demodulator chip oscillator, as well as careful filtering of the oscillator signal going to the demodulator. I'm not sure that a product detector per se will reduce that IBOC garbage, as I have a few other tuners that have diode product detectors that are somewhat affected by IBOC sidebands. I think the careful design choice on the PLL oscillator to reduce phase noise is more of a factor on that tuner,, but of course I could be mistaken.

 

[amptramp]

"I am stumped as to why they have a whole extra detector IC just to drive the meter which doesn't correspond to the actual center tuning of the audio detector, but I'm sure there was a reason." I have seen this a few times, where they set up this second detector that is used just to detect the null in the noise to operate the muting or tell if the station is tuned in. It makes it a pain in the butt to do the alignment. Sony used it on the ST-J75 and there isn't even instructions in the alignment procedure for it, but if it drifts, the tuner won't pick up stereo or will mute, especially on loud passages or bass notes.

If you have a sensitive IF strip with excellent limiters, by the time the signal reaches the detector, there are no amplitude variations, so you cannot run a meter or a mute circuit. You have to get the meter signal from earlier stages and have it operate the mute as well because you can be in a position to limit on cosmic noise. If you look at the schematic for the Marantz 10B, it has voltage doubler detectors on the diode mixer output, each of the six IF stages plus both limiter driver stages. This gives a signal they use to drive the analog level on the CRT. Only the limiter detectors are used for muting. If you don't have access to the individual IF stages, you need a separate signal path without limiting for the meter and mute circuits and that is what has been done here.

 

[w3jn]

Understood, and there are many other ways to detect the signal strength before the limiters. What's puzzling about this is why all this, to put the center tune meter off a completely separate detector that doesn't necessarily correspond with the actual discriminator?