Stereo Demodulator issue. Baffled!

[G8HQP Dave]

Nelson Pass is a well-known audio 'guru', but even he can't rewrite the laws of physics. A follower circuit, which is what the buffer is, works by using 100% negative feedback. Some audiophiles try to redefine 'feedback' to mean something else, so they can pretend to themselves that their amp has no 'feedback' - because they have picked up the silly idea that 'feedback' is intrinsically evil. You will find audio websites full of follower circuits, many proudly proclaiming that they have 'zero feedback'. Mostly this is due to ignorance, but Pass is not ignorant.

The uA758 circuit may be 'incomplete', as many datsheets circuits are, as it assumes that people will know about DC blocking, power supply decoupling, pilot tone filtering etc. etc. That is the difference betwen an application note and a magazine article; they are aimed at quite different audiences. It can't tell you what output cap to use, because that depends entirely on what follows.

[LucasAdamson]

Nelson Pass on Negative Feedback, for your perusal:

http://www.6moons.com/industryfeatur...istortion.html

[G8HQP Dave]

Ah yes, that is the article where he seems to imply that emitter degeneration is not feedback. See fig 9 and the accompanying text. He mentions feedback via R2, but somehow forgets to mention that R4 also provides feedback. I guess that is how a follower can have "no feedback" - you just redefine feedback.

Most of the article is OK, an elementary introduction to feedback. What may not be clear is that Pass tends to have an idiosyncratic approach to things, so fact and opinion can get mixed together.

[MrElectronicman]

As with most of the projects the data sheets only give you the general idea of connecting the thing.
I have made a few of these types of circuits and have always added DC blocking capacitors especially when you are going to connect it to a factory built unit, as you never know what may be behind the input socket.

[ms660]

Does the doubler output voltage stay the same when you get the loud hum?

[John_BS]

Originally Posted by G8HQP Dave

you just redefine feedback.

Most of the article is OK, an elementary introduction to feedback. What may not be clear is that Pass tends to have an idiosyncratic approach to things, so fact and opinion can get mixed together.

Well put.

John

[John_BS]

OK, so I have made and installed my new UA758 based Stereo Decoder in my Leak Troughline Stereo, and it works well up to a point. Kind of a big High 5 moment for me! I made it to fit in place of the old solid state de-modulator, in the small box inside the radio. It now runs from the 6.3vAC heater winding, by using a voltage doubler to get 15.8vDC

Can you post details of how precisely this is achieved: the 6.3V winding in the amp has one side grounded...... and voltage doubling will involve quite spiky rectifier currents.

John

[Synchrodyne]

Originally Posted by John_BS

Well put.

John

Indeed so. It seems that Mr. Pass excludes negative feedback that is intrinsic, i.e. that caused by an unbypassed cathode/emitter/source resistor, perhaps because no external and specific feedback connection is required. The inference is that an op amp voltage follower would have feedback though, as there needs to be a connection from output back to the inverting input. His prerogative to use a different definition for negative feedback, but given that it is contrary to well-established and well-documented wisdom, an explanation would be welcome. As a certain Mr. Asimov once said, “extraordinary claims require extraordinary proof.” Or maybe when one reaches “guru” status, that requirement is waived?

Still, however one names what it is that the unbypassed cathode/emitter/source resistor does in a follower circuit, it surely looks like and behaves like negative feedback.

Cheers,

[TuningIndicator]

Yes, I was pondering this too, and thought the same.

Interestingly, Pass cites two legendary audio designers; Peter Baxandall, and John Linsley-Hood; who's reputations are without equal.

I think the difference between Pass and the above two; is they were seeking progress in audio via scientific means; i.e. without a commercial interest. Whereas Pass clearly has a commercial interest, and his writings will always reflect that.

['LIVEWIRE?']

What puzzles me is, if the tuner already had a stereo decoder fitted which was working O.K., why change it in the first place. Admittedly(and I'm not familiar with the Leak Troughline)the original decoder may well be an old discrete component design, but the
uA758 is not exactly state of the art either!!!

[G8HQP Dave]

The original Leak decoder is generally reckoned to be quite poor. It just uses a few transistors. Almost any chip-based PLL decoder will better it.

Emitter (etc.) degeneration is undoubtedly feedback, but it samples the output current (or a proxy for the output current) rather than the output voltage. In the case of a follower, the emitter becomes the output point so there can be no denying that feedback is involved.

[John_BS]

I can't work out why the Leak de-emphasis time-contants appear to be much longer than 50uS? Or have I missed something?

John

[Synchrodyne]

Re stereo decoders, the attached excerpts from one of the Gordon King books show a selection of circuits, including a commercial example using the µA758, and another using an MC1310. The former has passive (L-C) output filters, whereas the latter has active output low-pass filters that use negative feedback!

Cheers,

[Synchrodyne]

And some more from an old Ambit catalogue.

The comment on p.17 (Hitachi HA1196) about oscillation suggests that some care might be needed with power supply arrangements for these PLL stereo decoder ICs.

Cheers,

[Synchrodyne]

And a bit more yet from another old Ambit catalogue.

Cheers,

[G8HQP Dave]

It is a mystery why the original Leak deemphasis is wrong. One theory is that the effect of the error is to deliberately introduce an HF shelf which will reduce treble without entirely eliminating it, thus better suiting FM to the ears of those raised on AM.

[John_BS]

Originally Posted by G8HQP Dave

It is a mystery why the original Leak deemphasis is wrong. One theory is that the effect of the error is to deliberately introduce an HF shelf which will reduce treble without entirely eliminating it, thus better suiting FM to the ears of those raised on AM.

Yes, that was the only explanation I could think of....... it seemed to be a consistent "error" rather than a slip of someone's pen.

John

[Synchrodyne]

Originally Posted by G8HQP Dave

The original Leak decoder is generally reckoned to be quite poor. It just uses a few transistors. Almost any chip-based PLL decoder will better it.

According to Spicer, the stereo decoder modules for the Troughline Stereo were supplied by Mullard/Philips, so they were probably to a standard Mullard design.

Originally Posted by John_BS

I can't work out why the Leak de-emphasis time-contants appear to be much longer than 50uS? Or have I missed something?

Spicer devotes a page or so to de-emphasis for the Troughline Stereo. Evidently the unbuffered outputs – mono and stereo - are much affected by the load that they look into and by cable capacitance. The time de-emphasis capacitor values were chosen on the assumption that the tuner would be used with the Stereo 30, which had 47k radio input impedance. Nevertheless Leak amplifiers varied from 47k to 470k, and Spicer presents a range of capacitance values that would be right for a range various input impedances.

Cheers,


(1) Spicer, Stephen; Firsts in High Fidelity – The Products and History of H.J. Leak & Co. Ltd; Audio Amateur Press, 2000; ISBN 1-882580-31-1.