The Audiophoolery Thread.

[greg_simons]

Thanks David for that excellent piece on how amp designers of the day managed their available devices, as an expert here perhaps you could give us some insight on what approach today's designers would take, given a clean sheet and unlimited budget, do the latest devices give us any improvements, or have we reached peak audio performance beyond which the human ear cannot detect any difference.
Greg.

[knobtwiddler]

I haven't had a chance to digest the THX patent that it's based on, but there has been quite a bit of conjecture (example here) that the Benchmark AHB-2 - with its uber-low THD spec - contains a modern take on FFEC. It wouldn't surprise me. The likelihood of a patent examiner having the chops to discern whether it were sufficiently different from the [expired] Quad patent is probably small (and there are many outrageous patents in audio that set precedents).

[Radio Wrangler]

If I remember rightly, the original patent that THX surround sound was based on was another ouvre by an ex HP guy, Siegfried Linkwitz. An R&D-ite who went into management in marketing and spent his relaxation time developing speakers. Someone else hereabouts has a set!

What would a sensible clean-sheet designer do?

Hmm, Difficult.

The easy way out would be true complementary bipolars and would be steered by available devices. Power MOS for audio devices seem to have been and gone. You can make a decent audio amp out of either them or bipolars, and things have been beyond the provable discrimination of hearing for a long time.

Actual current audio amplifier design is ruled by marketing. You have to do something different for them to hang their claims of superiority on. In reality, will anyone be able to notice any difference? Unlikely.

Doug Self's blameless amplifier would be a good starting point.

Other than that, you can have fun!

Bill, a colleague at HP was developing audio amplifiers for fun. He was trying for extreme simplicity with the minimum number of stages. He was reading all the same things in the library that I was, looking for ideas and inspiration. He did marry the librarian! Hugh Walker of wireless world stereo mixer and letters battle with JLH had married the previous librarian!

Bill and myself had a slight bit of competition going, so, me being a total ****-artist, I seto ut to design an amplifier to drive my rather large transmission line speakers and I decided to use as many transistors as possible, subject to each actually doing something useful. Thus was born the amplifier in my lounge getting the epithet 'Ludicrous' long before anyone had heard of Elon Musk.

Low power transistors and resistors are dirt cheap.

My input stage is a diff pair with cascoding common base transistors and there's a constant current source as the 'long tail' This stage is duplicated having both a PNP one and an NPN one driven by the same inputs. These take the signal outwards in voltage close to the +70v rail and to the -70v rail.

The cascode transistors act as common base stages with good bandwidth, protect the actual inout pairs from high voltages and greatly reduce Miller and Early effects. These stages are FAST and also quite linear.

The second stage is rather like the first. Constant current sources, a bipolar diff pair, cascoded. These act from the +70v and -70v bringing the signal into acting with respect to ground. There is a resistor to ground to control the LF gain, and a capacitor to set the HF roll-off. This second stage is also fast. There's a vbe multiplier to set bias for Iq in the output stage and then it's a pair of TO220 devices, PNP and NPN as emitter followers into two pairs of Hitachi MOS power devices.

Iq isn't very well controlled, but it diminishes as things get too hot, so it's safe and doesn't seem to create trouble.

Performance is silly. Harmonics are better than 100dB down with the amp 1dB below clipping. They fall at theoretical rates as the level is reduced. With a two-tone IMD test with PEP 1dB below clipping, intermod products are better than 100dB down.

R-C networks on the input set the LF end -3dB at 6Hz and the HF end -3dB at 56kHz. With +/-15v supplies and NE5534s in the preamp, even a rail-to-rail transient won't cause slew rate limiting in the power amp loop as it slews from saturation at one supply to the other.

I made one small goof. The 56kHz pole capacitor allowed the input transistors to tweet at RF. There really should have been base stopper resistors on the four transistors. It's the only bodge on the board.

The PCBs were laid out as a training exercise for one of our PCB layout people wanting to learn about doing layout from schematic herself rather than needing an engineer present all the time. She did a great job. A small run of the things were built and they're scattered around.

Bill's amp did go into production, with the Linn name on it. He'd left to join them and became head of R&D there. Still good friends.

If I had to do it again, I'd do something very similar, but with complementary bipolars. The stupidity of using many many transistors didn't cost much or use much area, so why not? It's an amplifier designed for a bit of fun. Its performance level is quite unnecessary. No-one's going to tell the difference form something more mundane unles you go an awful lot more mundane.

Back in the day I experimented with bipolars versus mosfet and decided it was a free choice. You can make a great audio amp with either. I engineered a single channel of DC-coupled valve amp purely for the hell of it. Transformerless, wideband and plenty of feedback. All traces of 'valve sound' were engineered out!

We're in an era where good performance doesn't tie you to any particular devices or techniques, so you have freedoms. You can trade them for cost, size or whatever, but I had fun instead.

David

[Craig Sawyers]

Absolutely David.

Personally I'm using Linkwitz's last gasp speakers - the LX521.4. These are fed by his design of analogue crossover, and then 10 channels of Blameless.

Alas SL shuffled off this mortal coil in 2018 in his early 80s after battling prostate cancer for the best part of a decade.

Because of the nightmare of wiring, and the possibility of getting it wrong and shoving the woofer signal up the tweeter, I use colour coding. And to remove the possibility of hum loops when feeding three power amp chassis, I run balanced throughout. Neutrik very kindly make colour coded plates that fit underneath their chassis XLRs with a little recess for a label, and the cable plugs and sockets (normally with black cable grips) can be fitted with their colour coded ones.

So all I do is match the colours.

Then to feed the amp outputs to the speakers - Neutrik Speakons. Once these are wired and checked, is is likewise impossible to wire it wrong.

Now because I have a love/hate relationship with IEC connectors, with their possibility of coming out if you step on a power cable - I use Neutrik Powercons. Lockable the same way as Speakons. Push in and twist.

The one thing that has been found by quite a few LX521 builders is that the lower midrange surround goes progressively hard, eventually becoming rock hard. Because this happens progressively - and only in one of a pair - you don't notice it happening. I actually thought I was going deaf in one ear. Alas SEAS have walked away from the problem, saying the drivers are out of guarantee. In a sense it is fortunate for SEAS that SL is not around - he would have been on the case big time.

Craig

[knobtwiddler]

Glad to know you're a fan of Hitachi Lateral MOSFETs, David. Baxandall reputedly said that he felt that well-designed SS amps were impossible to discern from one another by the late '70s (I presume he meant when operating below clipping levels). When I adjusted the bias on my 1978 Hitachi amp a few months back, I was amazed to see that it gave -106dB THD (0.0005%) when feeding 10W into a 4R7 dummy load. As I don't have the largest listening room and barely pull a Watt peak most of the time, I left it at that and didn't bother exploring more. I made a video as I couldn't believe that a 1978 amp was capable of such performance. Admittedly, I haven't checked it into an inductive load beyond a WW resistor, but I suspect that it would embarrass quite a few current-production amps out there.

IEC connectors, with their possibility of coming out if you step on a power cable

Isn't that better than the amp / CD player being dragged off the side and landing on your foot?

[Craig Sawyers]

Anyone come across Stereophile's challenge to Bob Carver? In the mid 80's he accepted the challenge to make his $700 amp sound identical to a Conrad-Johnson Premier Five (at that time $6k for a pair of monoblocks).

https://www.stereophile.com/content/carver-challenge

Craig

[Craig Sawyers]

Originally Posted by knobtwiddler

G

Isn't that better than the amp / CD player being dragged off the side and landing on your foot?

My power amps are far too heavy to be pulled anywhere by a mains cable. In fact just about everything in my system is too heavy to be pulled anywhere

Craig

[Radio Wrangler]

My power amp is the size and weight of a decent spectrum analyser. By the oddest of coincidences, the cabinet looks a lot like a spectrum analyser.... There's no risk of it being pulled off the small chest of drawers it lives on by an IEC cable.

David

[ScottBouch]

Flip side of this is, regarding immovable items: 2 weeks ago I had to replace a mains socket for a family member, as she had the tumble drier cable pulled across a doorway, at trip height, and you got it, her son tripped over it and wrecked the wall outlet socket!

[Craig Sawyers]

That reminds me of a story. Half a lifetime ago I worked for a technology consultancy company near Cambridge.

I ended up running a project for Granada (yes, the same company) Hospital Products Group in the US.

This was to do with selling the right to use the bedside telephone and TV when you were ill an in a hospital bed. Of course in the US you have to pay for everything.

Now they has already introduced a special remote handset so that when you paid to watch the TV, the orderly used the special handset.

Now the telephone used old technology - a strap with a lock on it. Pay to use the phone and orderly came with a key and took the strap off.

Granada wanted to use the same handset to authorise the phone. Could I design it?

Well yes of course I could. So I flew to the States for the start up meeting. The whole senior management team was there. So I am making a nice presentation of how this will work, when one of them chirps up "How are we going to stop people stealing the phone? They get pretty p***d off having to pay all the time, so there is a risk they steal the phone. Can we just recess the socket?"

"Won't work" say I "phone sockets are designed to be yankable, so if someone trips on the cable, the plug just pulls out the socket"

"Here's a socket with a cable to the telephone, let's test it" he said.

DOINK.

Nothing happened "you have to imagine someone is tripping and give a good yank.

DONK.

Nothing happened "put your back into it - remember someone is tripping on the cable"

DOINK at which point the socket pulled off the wall in a shower of plaster and a bunch of wiring came out.

"Well" say I "I'm glad you did that and not me!"

They took it in good part with quite a bit of laughter.

Anyway it was a good project and worked very well.

Craig

[sexton_mallard]

A bargain at $1495. A pair of old mattresses behind the speakers will not do!

https://www.shakti-innovations.com/hallograph.htm

"....The Hallograph contours the frequency, amplitude and time coefficients of the first reflections you hear, which produces a stunning increase in realism."

The testimonials speak for themselves...

[Craig Sawyers]

Good grief.

[Radio Wrangler]

Chocolate-dipped zig-zag breadsticks!

Flying in formation and endorsed by a mullet teleported in from the 1980s as well. Might attract rats, and not just the audiophool exploiting sort.

The boomerang-shaped mounting brackets exert symbolic influence to return incident sound waves, so any claims of damping are specious. Notice also the orientation of the breadsticks and you'll see that it's polarised. It could only ever work on vertically polarised transversal waves.

David

[Dickie]

Can always hang your washing on them when the tumble dryer breaks down.

[Craig Sawyers]

Originally Posted by Dickie

Can always hang your washing on them when the tumble dryer breaks down.

Ha ha

Probably would actually make a real difference to the sound from the speakers with washing hanging on them!

Craif

[knobtwiddler]

Originally Posted by Radio Wrangler

Chocolate-dipped zig-zag breadsticks!

Flying in formation and endorsed by a mullet teleported in from the 1980s as well.
David

The guy with the mullet runs one the web's most popular audio groups, with 124,000 members. Be careful what you say. Fear the mullet

(this discussion is blocked from being cached in Google - as I found when I tried to find something useful that I remembered being posted in it)

[Ted Kendall]

Utter bks

[Radio Wrangler]

Seeing 'shakti' in the URL was a good indication that it would open up a rich seam of audiophool silliness.

I have this daft idea for a Monty Python sketch where Peter Belt goes into his local post office to send some of his fine products, and the person behind the counter asks what the value of each package is....

David

[Trigon.]

Originally Posted by knobtwiddler

(this discussion is blocked from being cached in Google - as I found when I tried to find something useful that I remembered being posted in it)

Not here it isn't



Cheers

[knobtwiddler]

Originally Posted by Trigon.

Not here it isn't

Attachment 297912

Cheers

Oh... Maybe it was when I tried a search a while back... I was looking for a link in the thread that isn't related to audiophoolery. It'll serve me right for discussing non-quackery items in the thread