The Audiophoolery Thread.

[Hartley118]

Originally Posted by Radio Wrangler

I thought measurements were anathema, and that when a really high end system was delivered, the purchaser got a signed and sealed certificate that none of the equipment had ever been measured in any way. Such an action would have irretrievably polluted it.

The idea of having a measurement device actually built into an amplifier therefore seems odd. Something with a cheap and wonky movement would not be accurate and would therefore do less damage. The bouncing meter in the 125k monster has a power scale which goes beyond this and into the realm of comedy, so it's free from risk of performing damage, it can even enhance the pleasure. A sort of up-market answer to those Amstrad speakers with built-in disco lights.

Opinions vary, but everyone agrees that anything digital in place of moving coil would be completely unacceptable.

David

Not a lot of people know that the internationally standardised VU meter, with its diode bridge rectifier hung directly across a '600 ohm' line, adds significant distortion to the signal which may approach 1%, depending on the source impedance driving the line. Look closely at a manufacturer's distortion performance measurements for a studio mixing desk with 'proper' mechanical VU meters and you'll see in the small print that the meters were switched away from the outputs for the measurement.

Maybe the VU meter distortion (generally increasing at high frequencies) adds that 'certain something' for the ears of the audiophool.

Martin

Martin

[knobtwiddler]

Maybe the VU meter distortion (generally increasing at high frequencies) adds that 'certain something' for the ears of the audiophool

Alas, I have seen this for myself. There is a particular one that I could mention that has exactly this artefact, and it's highly regarded sound-wise. It uses O/P transformers that I suspect come from Alibaba.com. These introduce all manner of harmonics (not just 3rd), and the meters put the icing on the cake. I did think of putting a video on Youtube to show how the distortion goes up and down as you disengage the meters...but that would be mean.

The irony is that, if you have the knowledge to make a decent mixer (i.e. understanding of ground current flows and PCB / signal routing to get decent crosstalk and fader cut), then by default you should know how to isolate the ground currents from the meter circuitry to the point where they are virtually impossible to see, even with modern analysers. Having 'invisible' meters takes mutliple strategies (just running a dedicated return alone to the PSU won't do it), but anyone who considers themselves a designer ought to know how. Isn't it odd that people are still marketing products today that have the same fault that you identified in the 70s?

And yet you can still hear/identify a good amp even though its distortion might be 100 times lower than a pair of speakers.

I think all you can do is identify an amplifier you like, not a "better" one (well better to you) it's all down to frequency response, not distortion with any modern properly designed amplifier. I would bet the RIAA stage is most to blame/thank for this.

[stevehertz]

Originally Posted by merlinmaxwell

I think all you can do is identify an amplifier you like, not a "better" one (well better to you) it's all down to frequency response, not distortion with any modern properly designed amplifier. I would bet the RIAA stage is most to blame/thank for this.

Well I recall from the days of Hifi Choice - the lab test report books not the mag - that generally the better sounding amps correlated with the ones that turned up better performance specs. It's not all down to frequency response, no way, although that for sure that will affect a listener's judgment if the issue is not addressed properly for the purposes of the test.

[tritone]

Seen a website where a guy is selling a power cable for $1500 to power Hi-Fi sound equipment.
He swears and reckons that it sounds better.

H'mmmm

Doe's batteries, mains, wind generated, or solar panels electric sound different?


Confused.


Dave? what have you done to us all ?

The question needs addressed.

OK, I have a thing called perfect pitch, but A=440 anoys me, for me, its a wee bit lower and a nicer resonance at about 436-438. out of concert tune I know.

Tritone or dissonant collisions are stuff that we can all hear, but I cannot hear in my ears the difference of power sources supplying my equipment.

OK, Now, I can hear differences between class A, AB, G, and D, but not how they are powerd up with different sources of our friend, called electricity.

[Craig Sawyers]

Originally Posted by Radio Wrangler

Blew up twice?

Gives a second's worth of oscillation on turn-on?

Clearly, good engineering is undervalued. Circuitry which does not behave properly during turn-on and turn-off is not competently designed. Full stop.

David

Well I have no idea how long any oscillation lasted, or if that was the cause of silicon death. But mine fritzed twice on power on, and I've seen the same happen to a KSA50 in a high fi demonstration room. And the real monsters - the KMA200 monoblocks I know from the importer blew up while they were being installed.

So although they were beautifully built, their ability to survive switch on was touch and go.

Craig

[Radio Wrangler]

An expensive conflagration.

Class-A amps ought to be pussycats, they avoid a number of nasties of devices that get cut off.

I'd rather use maths and engineering to design an amplifier any day over fashion and religion, but it's fashion and religion that sell the big bucks stuff.

One thing I do in my power amps is not to let the forwards gain go uncontrolled-high towards DC. There is a load resistor and compensation capacitor where there is usually just the capacitor. Most people looking at the diagram think it's unnecessary, some say the amp would perform better at low frequencies without it, but it keeps the behaviour uniform, and it has very good advantages at turn-on and turn-off.

David

[knobtwiddler]

Did the Krell not have soft start? I'm guessing muting relays would be out...audiophools can hear relays you know!

Soft Start wouldn't stop oscillation if the design is on a hair trigger, but I am intrigued as to why an amp would blow up on start. Surely, they could snaffle away a couple of relays and WW resistors (a basic SS) in a little enclosure somewhere without the hifi press realising that they were there. They could be disguised as AC filters. I don't imagine that the transformer would stay mechanically quiet for very long without SS, due to the inrush current from (what I assume would be) massive filter caps on switch-on.

Something audiophools like to do is increase the size of filter caps without thinking too hard about it. Without soft start, this can lead to rectifiers flaming out and flying across the room...

[knobtwiddler]

Class-A amps ought to be pussycats

And pussycats like to curl up around them in winter! (as well as dogs)

[Craig Sawyers]

Originally Posted by knobtwiddler

Did the Krell not have soft start? I'm guessing muting relays would be out...audiophools can hear relays you know!

Soft Start wouldn't stop oscillation if the design is on a hair trigger, but I am intrigued as to why an amp would blow up on start. Surely, they could snaffle away a couple of relays and WW resistors (a basic SS) in a little enclosure somewhere without the hifi press realising that they were there. They could be disguised as AC filters. I don't imagine that the transformer would stay mechanically quiet for very long without SS, due to the inrush current from (what I assume would be) massive filter caps on switch-on.

Something audiophools like to do is increase the size of filter caps without thinking too hard about it. Without soft start, this can lead to rectifiers flaming out and flying across the room...

Yes it had a soft start. A chunky Omron relay with all four contacts in parallel shorting a 5 ohm 50W chassis mounted wirewound after about half a second. And a DC offset protection again via a separate Omron relay in series with the output for each channel. That clearly worked because when the smoke came out the protection relay immediately operated and there was never so much as a click from the speakers. There was also a thermal switch on each fan cooled heatsink chimney.

So it had all the necessary features. But still died twice immediately on power up. All the characteristics of second breakdown.

[knobtwiddler]

Great protection, shame about the amplification circuit itself!

Funnily enough, in another thread here a few days back, I was reminded of when I bought my demagnetiser from a branch of the chain 'Laskys', in what must've been 1984. The more commercial tower systems were on the 1st floor, and the exotic stuff was upstairs. They had a line of TTs as long as the eye could see, starting with Dual and going up in price. Doubt I will ever visit such a store again in my life (maybe they exist in Japan?). The first time I saw a Krell was in this store, but in a dark corner, underneath a shelf - totally hidden away... It was in the most hidden spot in the shop, as if they didn't want you to see it. I found its position highly odd. I wonder if this predisposition towards instability had something to do with its low profile in the store? The sales persons were only too keen to sell other stuff - they'd swoop and whisk you into the demo room. Maybe they weren't keen on demo-ing the Krell?

[Craig Sawyers]

This is a link to someone who stripped one of these Krell's, completely cleaned and reassembled http://home.ca.inter.net/~lloyd.maclean/Krell/Krell.htm

However, if you look at the main amplifier board, the straight heatsink has four mosfets on it. Everything else in the amp is bipolar. But there is no attempt at preventing mosfet oscillation; no grid stoppers, ferrite beads etc. Ugh.

But it at least gives a good impression of the scale of the beast.

The one you saw at the hifi shop hidden away had probably bitten the dust already.

Craig

[Radio Wrangler]

Bipolars can oscillate, too.

That link also has the Krell brochure for it, which makes a big thing about very low feedback - meaning overall feedback. Local degeneration is either invisible or doesn't count.

These things made a big splash back in the day, but maybe they invented the boat-anchor class of amplifier. And not a fitting tribute to 'Forbidden Planet'. Robbie the robot could have done better!

David

[knobtwiddler]

I had no idea that Krell used MOSFETs. I wrongly assumed that, like most American amps, they'd have BJTs all the way. I tend to think of UK OEMs being more into MOSFETs and the US makers opting for bipolars.

In fairness to our American cousins, I ran some tests on Levinson's ML1 preamp and found that its noise spec and FR flatness puts many modern OEMs to shame. Like a lot of gear from the period, they are overpriced today - but at least there is some substance to the claim of quality (volume pot has exceptional L/R matching), and it's not just a nice front panel and a lot of dogma.

If you consider Krell's fans, however quiet they could have been, unless you put the thing in a separate room, then I would've thought that the THD advantage to running the thing in pure class A to be somewhat obviated by the additional low-level fan noise... An A/B design of the day might have had more THD on paper, but if the fan noise is -50dB relative to your mean listening SPL, then the only real beneficiaries of the class A amp will be your pets on a winter evening.

[Craig Sawyers]

The Mosfets are not the output devices; they are the second stage of the input amp. Then after that bipolars for the voltage amplifier, drivers and output stage.

Why mosfets in the middle of the amp? Who can say.

[GrimJosef]

Originally Posted by Radio Wrangler

I thought measurements were anathema, and that when a really high end system was delivered, the purchaser got a signed and sealed certificate that none of the equipment had ever been measured in any way. Such an action would have irretrievably polluted it ...

That might explain an oddity I've just come across. I've just returned an expensive amp to its owner who had complained about hum and 'tizz' which has been worse than he would have liked for some years and has recently got significantly worse still. It's as simple as it can be - a single-ended triode output stage (PX25) driven via an interstage coupling transformer by another single triode (5842WA). There's no global feedback at all. It's a stereo amp, so there are two of everything in the audio stages. Only the HT supply (5U4G rectifier plus 10uF/10H/10uF/10H/10uF CLCLC smoothing filter) is common.

The main problem turned out to be failed smoothing capacitors on the PX25 DC filament supplies. These accounted for the 'tizz' which came from the 200Hz, 300Hz, 400Hz etc harmonics out of the bridge rectifiers.

Once they were replaced the remaining hum turned out to be at 50Hz. There was ~2mV RMS of it across each of the 8ohm speakers. The customer has a very high sensitivity horn speaker system which makes 2mV RMS very audible.

The amp is very nicely laid out and constructed. The chap who designed and built it has long experience and knows a good deal about the subject. So I was surprised that the hum level was so high. My usual experience of 50Hz hum is that it comes either from the small-signal valves' heater supply, or from poor layout of the input stage wiring, or from a signal grounding problem. I spent some time checking all of these out, but none seemed to be the cause of the trouble.

Eventually I tracked it down to 50Hz ripple on the PX25s' HT supply ! It turned out that although the amp designer is very experienced, his experience clearly doesn't cover the fact that directly-heated rectifiers, like the 5U4G, will impose a 50Hz component on their HT output if their filament's fed from a 50Hz supply and the output is taken from one end of the filment, as it was in this amp. The situation is worse if the subsequent filtering is also a bit light in capacitance, again as it was in this amp.

Given the care that's gone into the rest of the design I can only assume that the amp was only ever tested into insensitive speakers, and the actual output was never measured. If it had been then the 50Hz would have shown up and a pair of 'centring' resistors could have been used to connect the HT rail to the rectifier filament. Doing this eliminated the 50Hz almost completely.

Cheers,

GJ

[turretslug]

That does seem a curious oversight in a type and design of amplifier where a need for awareness of safeguards against hum is prominent throughout. I'm accustomed to seeing rectifier heater winding centre-taps in bigger vintage HT transformers where the 5V winding is of around 3A or greater capability (i.e. 5U4 and beyond in filament heating current)- I can't claim to have done the sort of statistical analysis that would let me claim that the majority are so equipped but it's certainly not an unusual feature. Extracting the HT from the rectifier filament winding centre-tap also cuts the HT current gradient along the rectifier's filament. Admittedly, a fixed winding centre-tap can't be adjusted for minimum hum in an overall circuit but it's surprising that even the resistor centre-tapping scheme wasn't provided for. Perhaps it illustrates how some of the science and art of valve amplifier design has faded into the background over the years.

[Valvepower]

Hi,

Ah, maybe I’m different in the world of audio electronics, as I usually do a sweep of the output between 20Hz and 50KHz down to -120dBV on the lookout for noise on the output. This lifts the stone on all sorts of creepy crawlies

A common creepy crawly seen on early spins of a PCB is errant charging pulses emanating from the reservoir capacitor ground layout in the power supply.

When I was doing the EF184 "bits ‘a this and that" amp it showed fairly early on the EF184 heater had to be run off a DC supply!

Regards
Terry

[G6Tanuki]

Part of me thinks this:

https://www.telegraph.co.uk/luxury/technology/hi-fi-system-costs-much-four-ferraris/

is complete and utter audiophoolery.

Another part of me thinks "OK, if that's how you want to waste your money, feel free, and kudos to someone for making a fat profit as a result".

[GrimJosef]

I heard that system, or maybe an earlier version, at an audio show a couple of years ago. It sounded fine. As always though the 'sound' was dominated by the original recording (as it should be), the speakers (they were seriously expensive and also seemed to work well enough) and the room (well, it's a hotel, I guess they did the best they could with it).

I think with a little care it would have been possible to assemble an amp + preamp + CD player using brand new components for about 0.5% of the cost of that system and not be able to find anyone who could have distinguished the two in a properly conducted double-blind test.

At another audio show, nearer 10 years ago now, a group of enthusiasts were challenged to assemble and demonstrate the best system they could using second-hand bits and pieces bought at sales and on eBay. The budget, if I remember rightly, was £250 or so and that had to include everything - source, amplification and speakers. The result sounded very creditable indeed.

Cheers,

GJ