Strange Hifi Amp For Your Scrutiny

[rp]

Been slow here since the malware made TAG Leper's Island, seems like the supply boat comes in once every three months now.

Here is a tangent hifi topic for you all. This is a friend's stereo power amp, '90s vintage from a boutique USA company that at the time had a small but world-wide dealer network. I'll leave out the name and model, there's a clue if you look close, the anonymity of the web lets people get carried away, and this way posters are freer to pee on it from a high place or to compliment the strange alchemy w/o hurting anyone's business. The company's amps were very well reviewed, known for their warm, relaxed 3D sound and their traditional builds eschewing the silicon regulated everything that was and still is more popular.

Friend said an original Mullard EL34 failed catastrophically, fuse blew, he put in a new quad of NOS Teslas, worked a few hours then the fuse went again. I had him send me a pict to see if I could spot a burnt screen grid or power string resistor. I was expecting the typical Dynaco, Mullard 5-20, or Eico HF-87 variant and was pretty damn surprised at all the bewildering stuff in here, most of which I don't understand, either the theory or the implementation. My intention was to send my friend to a Seattle area guitar amp tech for a quick fix and avoid the specialty hifi retailers which would surely ream him, or the risky and costly shipping back and forth to the maker. But the amp is weird enough that I told him to send it to the builder. It’s 25 years old now and might benefit from some newer alchemical updates/discoveries along with a re-cap (and some HexFreds if it was up to me).

Tubes are 12AX7 - 6DJ8 (PI) - 4X EL34, ultra-linear. Here's what I've noticed and questions:

Input stage & PI

V1a/b biased using diodes - for fully bypassed without the negative effects of using crappy electrolytics? I guess if you can use a LED you can use an any old diode. LED lighting up would be more fun though.

How is V1 getting it's plate voltage? Which leads to the way oddball phase inverter:

If I got it right, V1 plates share a resistor that comes off the filter cap after the choke and goes to the 6DJ8 grids. The V1 plates appear in parallel with V2 grids, the ones with the oddball Christmas lights or neon bulbs between the grids and cathodes on the 6DJ8. Woah!! WTF??? Some sort of current limiting trick, why? This is my main puzzle, the bulbs and how the PI seems strapped rather than coupled to V1 - is this what Direct Coupled means? How does it work?

Also, how does one dual triode PI drive the two channels?

Power tubes:

A cathode resistor for each pair but why only one shared 20mf/100V bypass cap? This forces you to string a too long wire across the whole amp. In a stereo amp isn't it best to decouple both sides as much as possible? Why not just stick a $1 cap across the other cathode resistor and avoid stringing up all the cathodes?

Big, black 5W Dale resistors right off the OTs. Part of the NFB I'm guessing but why 5W wirewound of all things, there isn’t a lot of current here. Are the wire-wounds there for their transparency?

Lots of small caps off the plates of the input and the output tubes. Tone shaping, cure for oscillations?

Heaters:

I think I got it: 12V heaters? Amp's got each side and the PI using half a wind and the CT that's not grounded. The 12AX7 gets both winds to 4 & 5 w/ pin 9 to ground? Looks like that. Builder feels 12AX7 runs better at 12V? Less noise? Builder plain hates twisting heater wires and soldering up the string? Builder wrongly spec'd 12V filaments on a huge order of PTs and had to deal with it? In a way it's pretty cool, if odd and rare, I too hate running heater strings. Anyone think this is a good way to do things?

Also, how do the heaters here go to ground? Don’t they have to reference somewhere?

To my sophomore eyes it looks like the builder knows enough to be dangerous, like me, but with way more chutzpah. I’ll let the savants here explain and make judgements and try to learn. I’m pretty suspicious of the esoteric, most successful things of any sort get attention and are repeated over and over.

My criticisms, up to my pay grade, are over the excessive, needless use of wire, especially the inputs, I absolutely would have moved the RCAs to the front, right next to V1. Would have used a cathode bypass cap across both power tube resistors, and run a normal twisted heater string. If a was using a molded bridge rectifier I'd source the kind that bolts to the chassis. I would have been consistent in my use of resistors and cap brands just so it doesn't look like a parts box build. I'd have mounted the f'ing choke straight. Oh yeah, and now I know why I use a different colour wire for each part of the circuit instead of basic black for everything.

Hopefully someone will find this interesting and take the time to look it over and explain and elaborate on the oddball stuff (please explain the PI light bulbs). I’ve never heard the amp so I’m reserving judgement on sonic results of all the alchemy. A lot were sold to a lot of people with good ears.

Cool detail: the blow-out hole for the filter cans are punched in the chassis and then lined up!

I can post some more in higher-res if requested.

[Structo]

Looks to have a real mixed bag of components?

Are those domino caps I see in there?

Can you show a picture of the exterior of the amp?

[rp]

Friend didn't send me an outside shot, but here's a google:

Yeh, the parts box quality for production amp is odd, certainly for maintaining consistency, but maybe it's part of the magic. I like techno geeks and I like shamans both can do good work, but I don't like sloppy.

The transformers were supposedly very very good. They look to be.

[pdf64]

The fuseholder looks to be at a funny angle; is its retaining nut loose?
The power cable doesn't look to be suitably secured to the chassis, and it's only 2 core, not grounded.
But so was my hifi; how do they manage to argue the safety case for that?

[jazbo8]

the oddball stuff (please explain the PI light bulbs).

It's rather odd design from QS, the neon bulb is used as a DC level shifter to couple the input stage to the PI - just to get rid of a coupling capacitor from the signal path. This particular model is well known to have noise and hum issues as well, looking at how the unit is wired, perhaps we should not be surprised...

[edit] Taking a closer look at the photo, I think the neon bulb is not used as a DC level shift but it's there to clamp the grid-to-cathode voltage - to protect the tube during initial power on, since it is direct-coupled to the first stage. So it is not really in the signal path and should not impact the sound during normal operation.

[Malcolm Irving]

.... I'd have mounted the *** choke straight.....

The funny angle of the choke might be the result of experimentation to find an angle which picks up minimum hum.

[rp]

The fuseholder looks to be at a funny angle; is its retaining nut loose?

I will email my bud about this, if it's loose I don't want that thing shorting to the chassis. If he does send it in for service I'll tell him to insist on a grounded ac mains.

This particular model is well known to have noise and hum issues as well, looking at how the unit is wired, perhaps we should not be surprised...

Friend said this ones hummy too, not sure how bad though, everyone's tolerance is different for such things.

[JMFahey]

the oddball stuff (please explain the PI light bulbs).

It's rather odd design from QS, the neon bulb is used as a DC level shifter to couple the input stage to the PI - just to get rid of a coupling capacitor from the signal path. This particular model is well known to have noise and hum issues as well, looking at how the unit is wired, perhaps we should not be surprised...

So for the theoretical pleasure of getting rid of 1 (one) capacitor (remember transformers anyway can not pass DC and to boot shift phase big time below the audio band) they added a device which is used to *generate* noise in the audio path?

How silly can they get?

[Structo]

I also noticed that fuse holder but figure you had spotted it at an odd angle.

But the design has me curious.

I am intrigued now.

Isn't it odd to have a neon bulb in that position, how can that sound very good.

I know for a fact a lot of those old mica domino caps were very leaky and sounded terrible.

Please educate?

Can someone PM me what amp this is?
Thanx

[trobbins]

Also suggest:

Rectifier negative should go to the cap, not the star ground.

The cathode bias resistors and bypass 0V ends should go to the main cap 0V, or at least the star, and not use chassis returns.

[jazbo8]

Not sure how to read the resistor code on these... Any idea?

[frankdrebin]

the oddball stuff (please explain the PI light bulbs).

It's rather odd design from QS, the neon bulb is used as a DC level shifter to couple the input stage to the PI - just to get rid of a coupling capacitor from the signal path. This particular model is well known to have noise and hum issues as well, looking at how the unit is wired, perhaps we should not be surprised...

So for the theoretical pleasure of getting rid of 1 (one) capacitor (remember transformers anyway can not pass DC and to boot shift phase big time below the audio band) they added a device which is used to *generate* noise in the audio path?

How silly can they get?

its not theoretical,its called "audiophile",i call it "idiotphile".

[frankdrebin]

do i see right?
is that a 2 wire power cable with a knot?

[frankdrebin]

12V on heaters its not a bad thing in general,i have a commercial guitar power amp which is like that,2 tubes in series,but you can't use just one side,you have to run al 4 to get sound.
Feedback resistors can be so big when you implement current driving,could you draw the schematic?

[rp]

Wish I had the amp in front of me I'd happily draw it out at least a layout, even post it on DIYA, QS schematics are rare and as it's an out of production amp I wouldn't have a problem, people have a right to fix their own stuff and play and learn.

Non-electronic friend took the picts which is why they are poor. First batch was w/ a cell and I made him go find a camera. I had scaled them down to save server space, here's a link to full res. Not much better but a bit.

EDIT: link removed see next post

25 years and I still can't read metal films, an other plus for carbon comps.

is that a 2 wire power cable with a knot?

I'm sure there's a strainer in front of the knot. This was a small production but still widely marketed amp with distribution world wide - and very well reviewed by the esoteric and establishment press and well loved by customers. Look close some of the couplers have the logo on them.