This audio cable business is getting out of hand...

[Billy Budapest]

Likewise, people who have very expensive mechanical wristwatches don't believe or claim their watches are any more accurate than a cheap quartz casio.

I remember 20 years ago I was looking at Tag Heuer watches at a jeweler. There were two identical models except one was electric, the other automatic. I asked the salesman what the difference was. He pointed to the quartz model and said “this one tells time better” and then to the mechanical model and said “this one is more expensive.” I bought the quartz.

 

[Krusty09]

Thanks for all the info guys. Great Stuff.

 

[Frank Dernie]

I remember 20 years ago I was looking at Tag Heuer watches at a jeweler. There were two identical models except one was electric, the other automatic. I asked the salesman what the difference was. He pointed to the quartz model and said “this one tells time better” and then to the mechanical model and said “this one is more expensive.” I bought the quartz.

I am an old bloke.
I have several watches. All the older quartz watches don't work any more. All the mechanical ones do. They definitely aren't as accurate but the one I got from my Dad for my 21st almost 50 years ago still keeps acceptably good time and is now rare and valuable.

 

[xr100]

All retired, apart from the 4PR speaker cable now. Previously, the power cables were used on amplifiers and CD players. Russ Andrews' shop was about 30 minutes away from my home and I fell for the whole cable upgrade path he advocated.

Just had a look around Russ Andrews' site--among the goodies is this:

https://www.russandrews.com/star-rf-router/

Yours for just £240! (Not sure if that includes the current "winter promotions" 15% discount.)

And just what might a "Star RF Router" be? No idea, but from the photos on the page, it's a circular shaped piece of wood with (presumably linked) binding posts arranged in a circle. (Suggested use is with an earth rod.)

Here's an antidote:

An Overview of Audio System Grounding and Interfacing (Bill Whitlock of Jensen Transformers.)

 

[DonH56]

Note class G and class H generally mean fundamentally a class AB amplifier with some sort of fancy power supply wrapped around it. I have read papers about applying class G/H to class D "core" amplifiers to provide even more reduction in power and heat. The problem is dealing with the voltage ramp; only recently (I think) have high-voltage output devices gotten fast enough and cheap enough to consider something like that for class D. The output transistors switch very rapidly, typically several hundred kHz, and doing that across lots of voltage with lots of current is hard. My day job deals with circuits switching much, much faster but at a (tiny) fraction of the voltage and current of an audio amplifier output stage.

 

[Wes]

"D" was next in the alphabet, but what happened to Class E, and F?

Why the skip to G, H?

 

[Blumlein 88]

I am an old bloke.
I have several watches. All the older quartz watches don't work any more. All the mechanical ones do. They definitely aren't as accurate but the one I got from my Dad for my 21st almost 50 years ago still keeps acceptably good time and is now rare and valuable.

Well I have one of the early TI quartz watches that still works (late 1970's). I had a couple of Casio's that lasted a few years and got damaged in various sporting accidents. I've a 1984 Seiko quartz that works great and keeps very accurate time. The Seiko was worn every single day for 27 years. It survived many things normally terrible for a watch. I've a Citizen which is light powered. Not quite as accurate as the Seiko, but works nicely.

 

[DonH56]

"D" was next in the alphabet, but what happened to Class E, and F?

Why the skip to G, H?

C, E, and F are all well-known and used extensively in the RF world for oscillators and power amps plus a few specialized applications. The problem is they (e.g. output stages) all conduct less (sometimes far less) than 50% of the time so achieving high linearity, especially at low power levels, is difficult to the point of impossibility in a practical sense. You don't generally care about harmonics in an oscillator, but very much do in an audio amplifier. They also tend to depend upon a resonant tank circuit so have limited bandwidth (single frequency to fraction of an octave) vs. the multiple octaves required for audio amps.

Short answer: they would work, but they sound bad.

So we had A, then B, and so on (C was used in RF amps back in the 1930's, I think), then D, and sometime (much) later E, F, G, H, etc. The sequence is as the amplifiers topologies were discovered and proven, and not all classes are good for all applications. Saving power in a high-power oscillator for a kW (or 100 kW or more) RF transmitter means class C is a good choice -- you use it narrowband, single frequency, and filter out the harmonics. Not so much if you want less than 10% or 30% THD from your audio amp and want to hear the entire audio frequency band.

HTH - Don

 

[Frank Dernie]

Well I have one of the early TI quartz watches that still works (late 1970's). I had a couple of Casio's that lasted a few years and got damaged in various sporting accidents. I've a 1984 Seiko quartz that works great and keeps very accurate time. The Seiko was worn every single day for 27 years. It survived many things normally terrible for a watch. I've a Citizen which is light powered. Not quite as accurate as the Seiko, but works nicely.

I have a citizen and 2 Junhans light powered radio set watches. I used the Citizen when I was travelling and one of the Junghans was my favourite watch but it died and I get no response from their service dept at all about repair. The other, older Junghans still works but the radio aerial is in the leather strap which is coming to pieces.
I enjoy a nice mechanical watch and being to the nearest minute is fine for me nowadays.
I am much more interested by the movement than the case/face styling though, making me a bit weird.

 

[mhardy6647]

I am an old bloke.
I have several watches. All the older quartz watches don't work any more. All the mechanical ones do. They definitely aren't as accurate but the one I got from my Dad for my 21st almost 50 years ago still keeps acceptably good time and is now rare and valuable.

How we tell time at my house...

clockfront by Mark Hardy, on Flickr

 

[mansr]

The problem is dealing with the voltage ramp; only recently (I think) have high-voltage output devices gotten fast enough and cheap enough to consider something like that for class D.

Would it make sense to create a combination DAC/amp with a few milliseconds of lookahead to ramp up the voltages as necessary? Has it already been done (like most good ideas)?

 

[Billy Budapest]

I am an old bloke.
I have several watches. All the older quartz watches don't work any more. All the mechanical ones do. They definitely aren't as accurate but the one I got from my Dad for my 21st almost 50 years ago still keeps acceptably good time and is now rare and valuable.

My Tag finally broke last year. 22 years of service. I’ll try to get it fixed but hopefully at reasonable cost. If it’s more to fix than an Apple Watch, then I’ll go for the Apple Watch. So yes, I get it! An automatic would probably still be running.

 

[Frank Dernie]

My Tag finally broke last year. 22 years of service. I’ll try to get it fixed but hopefully at reasonable cost. If it’s more to fix than an Apple Watch, then I’ll go for the Apple Watch. So yes, I get it! An automatic would probably still be running.

I have an Apple watch 4. I end up wearing a proper watch as well since the display is mainly off. I hate the touch screen, far too often it reacts to a lightest accidental touch when I don't want it to and ignores a firm touch when I want it to - but I find the touch screen is a poorish jack of all trades and only a master of the unintended selection. Detest the things but so cheap they are popping up on everything.
When it stops working I will bin it and carry on with only the proper watch.

 

[mhardy6647]

speaking of hifi wires...

I was readin' the most recent Stereophile (Feb. 2020) last night. There's a new-ish column, I guess, by Herb Reichert called Gramophone Dreams -- but apparently he's focusing on hifi wires for at least a while (as I mentioned here earlier, in fact).

At any rate, while I (who as some of y'all will know am a bit of an agnostic vis-a-vis some of the more esoteric and extravagant wires on the market for hifi nowadays) thoroughly enjoyed the article (and would suggest readin' it unless you have a really deep intrinsic bias re: Stereophile, Reichert, or wires) -- I was really taken with one aside contained therein.

So taken, I'm a-gonna share it with all y'all.

Last year, when I visited Nelson Pass of Pass Labs at his home in Sea Ranch, CA, I was surprised to discover that he connected all the gear in his laboratory listening room with Radio Shack cables. When I asked him why, he told me, "Because I don't want anybody else taking credit for my accomplishments."

I knew I liked that guy.

I know all y'all (well... most of y'all) ain't big fans... but I like his style (and his substance).

 

[DonH56]

Would it make sense to create a combination DAC/amp with a few milliseconds of lookahead to ramp up the voltages as necessary? Has it already been dine (like most good ideas)?

The problem is handling the voltage step at the output without introducing instability or more distortion. PCM/PWM/PFM normally switches between 1 and -1 (equivalent in voltage) so you have to be very careful when you introduce additional steps. With a linear amp, there is rejection of supply noise, but that works differently in class D when the supply rails are essentially the output states. Modulating them can directly modulate the output waveform with power supply noise, very bad normally, and a problem to deal with if you want to do it intentionally without mucking up anything else.

That said, I am not an audio designer by day, and am not up on all the various class D flavors and research going on these days.

 

[Jaimo]

speaking of hifi wires...

I was readin' the most recent Stereophile (Feb. 2020) last night. There's a new-ish column, I guess, by Herb Reichert called Gramophone Dreams -- but apparently he's focusing on hifi wires for at least a while (as I mentioned here earlier, in fact).

At any rate, while I (who as some of y'all will know am a bit of an agnostic vis-a-vis some of the more esoteric and extravagant wires on the market for hifi nowadays) thoroughly enjoyed the article (and would suggest readin' it unless you have a really deep intrinsic bias re: Stereophile, Reichert, or wires) -- I was really taken with one aside contained therein.

So taken, I'm a-gonna share it with all y'all.

​

I knew I liked that guy.

View attachment 45347

I know all y'all (well... most of y'all) ain't big fans... but I like his style (and his substance).

Now here’s a guy I’d like to see participate in this forum.

 

[mansr]

The problem is handling the voltage step at the output without introducing instability or more distortion. PCM/PWM/PFM normally switches between 1 and -1 (equivalent in voltage) so you have to be very careful when you introduce additional steps. With a linear amp, there is rejection of supply noise, but that works differently in class D when the supply rails are essentially the output states. Modulating them can directly modulate the output waveform with power supply noise, very bad normally, and a problem to deal with if you want to do it intentionally without mucking up anything else.

Right, but could any of those problems be mitigated by coordinating the switching with voltage adjustments based on foreknowledge of imminent requirements rather than reacting after the fact? Just a thought, possibly a foolish one.

 

[Wes]

As a general control systems issue, yes. It's been done - dunno if ever done in ampligiers.

 

[Wes]

Dang it! My predictive algorithm failed, despite the foreknowledge of imminent requirements for an 'f' I put a 'g' there...

 

[DonH56]

Right, but could any of those problems be mitigated by coordinating the switching with voltage adjustments based on foreknowledge of imminent requirements rather than reacting after the fact? Just a thought, possibly a foolish one.

No, it is a good thought, and in fact in practice is a requirement, but it is a necessary but not sufficient condition. You would use that "foreknowledge" to control switching but at the output need something that carefully coordinates and controls the transition to prevent adding undesired signals at the output. Feedback can help, and feedforward (more likely for something like this), but it's tricky.

A normal output stage haas decent power-supply rejection plus feedback to make it great, so when you change the power rails it does not get through to the output signal. For class D, changing the rails directly hits the output of the amp, so isolation is a major issue.

For all I know some class D audio designs already do it. I have implemented something like that, feedforward and all, in RF circuits, but for those amps the output modulation is way outside the signal passband so I just had to ensure changing the rails did not generate very high-frequency glitches. For class D, you have a similar situation if you switch the rails rapidly, but any settling "tails" can fall into the audio band so more care is needed.

@Wes -- You can edit your posts, and if you do it quickly enough, nobody will be the wiser. Don't ask how I know this...