Doc. Gets Questions!

Pertinence: QUAD ELS, Gradient Subwoofer SW57 Marantz 8, 8B. Capacitors and dielectric absorption PS circuit impedances: rectifiers and transformers.

In the corridor outside a burger-and-chips restaurant, near the exit, a worn fellow was accosted by a much larger, dumber looking fellow on his way toward the counter. Worn fellow got swung around like he was in a turnstile. Ejected into the arms of a strong Scandinavian man, worn guy, in gratitude told the Scandinavian guy all kinds of fantastic stuff upon learning that Scandinavianer happened to be, like him, an audiophile.

 

Worn: Hi there! Thanky thanky for saving me from the hazards of Mr. Fat Bastard. He's an audiophile, but at times like this, about all he can think about is how much he enjoys the flavour of burger fat.

Dude: You are an audiophile? I am an audiophile. My, that is a coincidence.

Worn: No, I'm not that surprised. You have no ring on your finger. That's got to be because your love of music is greater than your love of women.

Dude: You could not be more wrong. So tell me...

(500,000 word excision for the sake of brevity. With apologies -ed.) *

Worn: ...And that's what makes an electrostatic loudspeaker like the QUAD so difficult to match to a dynamic driver subwoofer. ...

(50,000 word excision for the sake of brevity. With apologies -ed.)*

*Flingpoo reply'd ... No wonder why I stopped writing for The Audion.
Ed. took out all of my best material...

... Excuse me if I retain this viewpoint until convinced otherwise! At least they are better than Vifa, makers of shiny-looking cheap'n' smudged-sounding sonic garbage! I would think the only electro-dynamic woofers that stand a ghost of a chance of sounding good enough to match the QUAD would be a kevlar or glass-fibre composite one from either JM Labs/Focal or B&W of England...

I still think it is an unworthwhile compromise that, once your system becomes transparent enough to really allow the QUADs to do what they do in a way that lesser technology speakers cannot, you will quickly begin to hear the subwoofer cones "lagging behind" and sounding different in character from the rest of the range, and this difference will become increasingly more obnoxious until you will have to disconnect the subwoofer entirely. At which time, you will wonder how you ever put up with the added distortions!

There are a few amplifiers that are, if not the absolute best, at least compelling and enjoyable mates to the QUAD. I would recommend the Marantz 8B as the best amplifier of the vintage types, preferably with an upgrade of the signal-coupling capacitors to polypropylene types, my favourites are WIMA FKP1 (use MKP10 where size is prohibitive...). Teflon types would be better still, but I've never tried them as the cost is prohibitive to the point of insanity! Some audiophiles prefer polystyrene, and these are excellent as well. But they tend to be more expensive than polypropylenes, and I find they tend not to sound as good and are easier to damage with the heat of the soldering iron, as well. They sound "thinner", with less low-level detail than a good polypropylene, and with weaker "image focus" as a result.

You could also describe this effect as a lack of "presence". There is a fad amongst the audio/mystic/fuckup single-ended triode/horn speaker/zero-feedback DHT crowd for capacitors that use paper and oil, such as the Tobias Jensen PIO stuff. I used this GARBAGE years before any of these twits (save the old Japanese tin-eared retards) ever even heard of them, and mostly it was a quick listen in vintage equipment that had been fitted with them originally before they were yanked and tossed in the rubbish bin in place of WIMA, Philips or Siemens polypropylene and polystyrene capacitors that sounded INFINITELY smoother, more transparent, more detailed, more focussed, tighter, vastly cleaner overall, etc. etc.

There is a simple reason why Teflon, polypropylene and polystyrene capacitors sound better than all the other types, such as paper/polyester, paper/oil-impregnated, polycarbonate, electrolytic, tantalum, polysulfone, etc. It is called "dielectric absorption", and it is a nonlinear characteristic that is lowest in the 3 dielectric materials of choice that I've mentioned. It is VASTLY (as in, 30-100 times) higher in PIO's, ceramics, silver-micas, Black-Gate and Cerafine electrolytics, etc. (the latter four types are over 100 times worse...)

There is a science to good audio, despite what those "Audio/Mystic/Fuckups" would like you to believe (which I think is to have audio as a religion, always unknowable, just random and therefore perhaps "more interesting" since you never know what combination of good & lousy & weird components may, all of a sudden, make you have a wonderful audio experience...).

I have given some thought to which circuits are best for the original QUAD, and my favourite for this purpose was the Dynaco Stereo 35 POWER AMP (NOT the SCA-35 INTEGRATED amp...), with the miserable little 100nF input coupling ceramic capacitors jumpered out... Unfortunately, the Marantz 8B would have to be bought, not built! This is because it is an elaborately phase-compensated amplifier, with phase-linearity most closely rivaling an OTL or excellent transistor amplifier...Of course, we don't want that zippy soild-scrape sound, so as nice as a Spectral DMA series or YBA amplifier can sound with the better B&W and JM Lab speakers, it's just not good enough!

So we are left with the dangerous OTL's, like the affordable Rozenblit stereo amp for a grand, or a Marantz 8B (make sure it is NOT a "plain vanilla" 8, as this older variation of the amplifier sounds NOWHERE near as good, and we see that it doesn't have all the phase compensation parts on the secondary of the output transformer, otherwise it is EXACTLY the same amplifier! A more dramatic example of the power of phase-compensation to make or break the sound of an amplifier cannot be found elsewhere!). The 8B would cost you between $1500-2000, the way prices are. And you would still be faced with finding the vital Mullard output valves, as the seller may have removed them for some reason or anotherr.

Because you are in Norway, you will need an isolation transformer to step down the 220/240VAC to the 120VAC the American Marantz was designed to run on exclusively. The added line impedance of the isolation transformer will dull and soften the sound of the amplifier to an unacceptable level, and you will have a pretty mediocre sounding amplifier, even after you upgrade your capacitors!

The proper solution would be have a new transformer made, preferably of better quality for even better sound, and this would be easy enough to do. The best sounding are toroidal types, such as a Lindahl, ILP or Plitron unit. This would have to be custom-made, and the higher the VA (bigger size) rating, the better. The original square power transformer is an E-I core type, and the toroid won't fit well, if at all, into the amplifier case. And if you are going to uprate the VA rating over the original, especially substantially, you will have to put it external, in it's own enclosure and hooked up the amplifier chassis with a potentially dangerous "umbilical" cord. The 8B does have a blanked-off socket on its rear for an 8 or 11-pin valve/relay socket, so you could readily mount a socket there without destroying the cosmetics of the collector's piece by having to drill holes in the chassis!

The toroid would sound fantastic, but could easily cost you between 200-400 Euros to have built. Add the enclosure, wiring and connectors, and you'll have probably added another 100 Euros to the cost. Then, if you do this, you will want to replace the miserable and extremely old Western Electric electrolytic main filter capacitor, along with the tiny electrolytic cans located inside the amplifier. Most of those are in the bias supply, but there is a cathode bypass on each of the 6BH6 input valves. The cathode bypasses for the 6BH6 should merely be removed, and no replacements are necessary. Input sensitivity of the amplifier is very low, and will be lower still as a result. An American or Japanese "high output" (2.5V or so for full output is needed here...) preamp will do the job.

As for the main HT cap sections and the bias capacitors, a QUANTUM improvement in sound can be obtained by replacing them with something like Chateauroux (Solen) polypropylenes. As the dielectric constant of the polypropylenes is much lower than it is for electrolytics, the Solens will be HUGE in size compared to the original electolytics! As a result, again, you will have to mount things externally in another box (perhaps with that new toroid, it would be OK to share the cabinet) to make it all work.

As a result, you will end up with a Marantz 8B connected to a big, heavy box that may well be considerably larger than the amplifier itself! Not to worry, though. At least you will have the best-sounding transformer-coupled amplifier ON THE ENTIRE PLANET, of that I can virtually guarantee you! If you were ambitious enough to replace all the wiring with stranded silver-teflon insulated multistrand (US military "Type E" or "EE" spec wire), and to replace all the old shitty carbon resistors with either bulk-foil (Vishay) or non-inductive wirewound types (Mills or Micro-Ohm) then the transformation to near-perfection would be complete!

Otherwise... I suppose you could do OK by copying the circuit for the Eico HF87 EL34 amplifier (go to Triode Electronics website and look up the schematic there). The newer HF89 is not quite as good. Avoid using rectifier tubes where possible, they degrade the sound considerably over silicon diodes, due to their large internal resistance.

To show you what kind of effect these added impedances in the power supply can have, try this at home: With a component that is reliable and working well, such as a source component or piece of amplification (little may happen if you try this with the QUAD speakers!), replace the mains fuse with the largest amperage one that you can retrofit. With the European-standard 5 x 20mm fuses, I think you are limited to 6,3A ratings.

With the American/Japanese 1/4"x1-1/4" fuses, 15A is the largest, and some automobiles used these fuses as well, so 30A types can even be found! Obviously, once retrofitted (and to both ends of the mains cord, as required...), these large fuses won't really protect much of anything, so if a component malfunctions, it may do so with SPECTACULAR consequences! But if you do this and don't leave the equipment unattended, just do a fairly brief retrofit so that you can audition the difference, things ought to be just fine. Just get ready to jump up and turn off the mains VERY QUICKLY if the need arises! (This is never the case. ed.)

With the larger fuses fitted, you will immediately notice that the sound is much SHARPER. Not in a bad way! The bass will be tighter, but also subjectively DEEPER. The midrange will be cleaner, smoother, with much better separation of instruments and much better focus to the "image". The upper midrange will loose glassiness and glare, sometimes to a shocking degree! Your ears may well express a deep sigh of relief, whispering a loud "THANK YOU!" to your brain for the relief! As well, the extreme treble will seem lighter, sharper, more airier and noticeably FASTER sounding. It will just plain sound like you put a better component in place! Now, the difference in impedance between a 1/4A fuse (as in a CD player) and a 6,3A one is probably no more than about 3 ohms, max.! A GZ34 rectifier tube typically adds about 250 ohms (!!!!), and the difference between the secondary impedance in a stock Marantz 8 power transformer and a custom 500-1000VA toroid is probably over 1000 ohms!!! Need I say more?

I have done several transformer retrofits in Dynaco 70 amplifiers. In one type, I substituted the power transformer used in an Altec PA amplifier, the type 1569A. This amplifier uses 2 pair of EL34 in a typical Williamson circuit, and is a mono "rack mount" piece of shit. But the power transformer in it....! What a monster! And it has the EXACT mounting centres as the PA060 Dynaco transformer, and it is designed for the same type and number of output valves! The bias winding needs to be "padded down" to the 70's voltage, and there is some minor modification of the filament winding, since the 1569A is a mono amplifier with only one 6,3V winding, and the 70 is a stereo amplifier with two windings, one per channel. Why Dynaco bothered with a second winding is beyond me. It's not necessary, and the single 6,3V winding in the Altec doesn't hurt noise or channel separation even ONE IOTA.

The main HT voltage is actually slightly lower in the Altec, so you lose a little bit of power in the 70, but lower voltage also helps the sound (and reliability, of course!). This massive E-I cored substitute in the Dynaco changes the sound of the amplifier even more than changing valve brands can (by a long shot!) or even the otherwise dramatic difference that capacitor and resistor upgrades make! The amplifier INSTANTLY sounds MUCH smoother, much more confident, suave, and relaxed! It sounds more EFFORTLESS, even much more powerful!

The bass tightens right up, the midrange is noticeably clearer and smoother, the "haze" in the upper midrange all but disappears, the fuzzy and smeared "pentode" treble cleans right up. What IS left is a slight heaviness and slowness in the bass, a slight darkening to the highs, a touch of texture in the upper mids. Still, NOTHING like the severity of these exact traits that a "stock" amplifier possesses. Insert an ILP toroid, and it's a different world again! The ILP is no larger than the Altec, perhaps it is a bit smaller and lighter, even. It is rated 300VA, very healthy for a 70 Watt stereo valve amp... With the ILP, you can dispense with the rectifier valve, and use a full-wave bridge rectifier instead of the full-wave centre-tap that the other two transformers require.

I did do one Dynaco Stereo 70 with the Altec transformer, new beefed-up filter caps, and solid-state rectification. It still had the residual colourations of sound that I've described above. So, when I put in the ILP toroid, I did much the same thing, with the EXACT same amount of filter capacitance, and the same diodes, just in a bridge configuration instead of the FWCT pair. Again, the sound of the amplifier was VASTLY different! The bass was near-razor tight, the mids unbelievably open, transparent, TOTALLY lacking in valve colouration, warmth, smear or "funkiness" of ANY kind. This is the midrange of a Spectral DMA or newer Krell amplifier...only with MUCH more low-level detail and PRESENCE!

The "pentode" upper midrange grain/haze/scratchiness in now TOTALLY GONE, leaving a sound that would make you ABSOLUTELY CERTAIN that you were listening to an all-triode amplifier...when you still had a small-signal pentode voltage amplifier (7199) in the input circuit, and you were using Power Pentodes in your output stage (in fairness, I DO connect the EL34's as triodes rather than leave them in the stock "ultra-linear" connected mode...)! The treble becomes incredibly clean and extended, and the amplifier EASILY rivals a fully parts-upgraded Marantz 8B in EVERY way! All this, despite low-quality OPT's, no phase compensation (I remove the badly thought-out garbage that is in the 70...the asymmetrical 390pf screen feedback loop from one of the EL34's, that causes ringing on the negative cycle of the square wave!

I remove the slew-limiting and bandwidth-killing 18k/82pf filter in the input stage) and the simplest possible circuit using a fairly noisy high-gain pentode and crude splitter inverter... When audiophiles talk, especially the pseudo-intellectual so-called "designer" types, they always give you a bunch of blah blah blah about the importance of the OUTPUT transformer. Never so much as a word about the power transformer!

Perhaps you know of the once-legendary reputation of the French valve amplifiers from the company known as Jadis. Jadis, like all over-arrogant French demi-wittes, boasted obnoxiously about their supposedly-unbelievable OPT quality. "The OPT in the Jadis 200 (then the flagship) takes 24 man-hours to wind..." I've had JA200's in my home, I know what they sound like, I know what they can do. And in EL34 guise (as opposed to the later "Americanized" fuckup versions with the horrible 6550/KT88 output valves...although the original European versions had the Genalex KT-88...still pretty lousy-sounding compared to any EL34, even the Russian ones!), the Jadis "does the business".

Anyone who knows anything about these amplifiers knows that they measure quite badly! They have horrible HF characteristics, notably, the slew rate of the amplifier is PATHETIC! HF distortion goes through the roof, the amplifier cannot put out 200W, it only puts out 160W, and it only does it IN THE BASS! In fact, it will do a low-distortion 160W to below 20c/s, which is quite impressive for the OPT! But it only shows that the core size of the OPT is conservative, little more. ANY large OPT can give good bass performance, it has NOTHING to do with winding expertise or quality!!!!

But despite the 160 Watts, the Jadis struggles to put out even 20 Watts by the time it gets to 20,000c/s! Pathetic performance, indeed! But if you've ever heard this amplifier, you won't care about how it measures! You'll be IN LOVE, IN RAPTURE... It's far from perfect sounding, enough to maybe even convince you that what you are really liking is just its particular array of distortions. Certainly, this is what some idiot old-school engineer would have you believe... Only, the things that make this amplifier sound so good are its sense of SUBJECTIVE CLEANLINESS AND EASE, its SUBJECTIVE LACK OF DISTORTION, particularly through its ultra-seductive mid-band, where it measures EXTREMELY WELL, thank you! In fact, there is no measurable distortion in the JA200 of anything above the 3rd harmonic, NOTHING. IF there is any distortion above the 3rd harmonic at all, at any power level substantially below full power, it is AT LEAST -120db into the noise floor!!! Beat that, Spectral, Goldmund, YBA and Sony Esprit!!!

At the end of the day, HF distortion measures horribly in the amplifier, if you try to get ANY power out of it at those frequencies and the bandwidth of the amplifier is not especially wide. An analysis of the bizarre "River" circuit that is used, apparently a French hobbyist-magazine legendary favourite from the very early 1960's, shows that its' phase inverter circuit is the main problem. Poor symmetry, and it uses but a single pair of ECC83's running at only 1mA to try to "push" or "drive" an output stage consisting of no less than FIVE PAIRS of EL34's! That 1mA simply cannot cope with the cumulative grid capacitances of FIVE EL34's, aggravated by the Miller Effect that multiplies it when negative feedback is used around a circuit.

The Jadis uses a fairly significant, if slightly below average, amount of the EVIL negative loop feedback... As a result, this capacitance drags down the EL34 grid impedance as frequency rises, which is what capacitors do. They block DC and low frequencies, gradually turning into conductors, or short-circuits, as the frequency rises. And since this capacitance is IN PARALLEL with the EL34 grids, it creates the effect of a gradual short-circuit for the ECC83 inverter and its paltry 1mA to try to drive... If we accept that the key to the Jadis' brilliant sound is its highly flawed and limited input circuit with this phase inverter (and I do, because it is!), then we understand that good HF measured performance can NEVER be within this amplifiers repertoire. And here's the "punchline":

If good HF performance is NOT a property of a Jadis JA-200, then it DOES NOT need a good OPT!!!! Maybe it does take 24 man-hours to wind this OPT. More likely it takes 10 minutes to wind the hunk of junk, then another 23 hours and 50 minutes to encapsulate it into a big, impressive-looking black casing full of gooey tar compound, and to engrave and wax-fill the engraved letters, and then spray with a clear lacquer the 6mm brass nameplate that is glued to the top of that casing! As large as the OPT is in the JA-200, we know that it is not as large as it appears. This is because the transformer must be smaller than the case it is glued into! How much smaller? We don't know, because we cannot see it due to the potting compound and casing!

Obviously it is still quite large, as the amplifier does measure very well at high power deep into the LF part of the frequency spectrum. But while we are called upon to waste our time committing our visual attention to the considerable size and undoubted beauty of the OPT casework, we are NEVER told to consider the characteristics of the obviously EVEN MORE impressive power transformer! The power transformer is NOT encapsulated. It has a painted laminations, but is otherwise a pretty "run-of-the-mill" looking device of "nothing special" constructional technology.

It DOES have a very pretty chrome-plated top "bell" on it, and that's it. Except for one thing. Yes, this is a very high-powered amplifier, particularly for a valve type. But look again at the power transformer... If the amp wasn't so HUGE, with such a HUGE chassis, we'd more quickly and clearly realize just how MONSTROUS the JA-200 power transformer is! And man! It is a MONSTER! As big as the OPT is, even with the extra size imparted to it by the encapsulation, the naked power transformer is NOTICEABLY LARGER STILL!!!

Think this was just done to give a "nice" impression of quality, and serve no other particular purpose? To think so is not to understand Jadis in particular, and the French in general... First of all, the French are NOT known for the quality of their industrial goods. Food and drink, yes. Clothes? Oh definitely, but only the style! Not for durability, toughness...NO WAY! So suddenly a French hi-fi company, when no other French hi-fi companies before them did truly high-quality build (when Goldmund was still French, along with Lurne, their turntables were so badly made that even when one had paid tens of thousands of dollars for a Reference turntable, the things still had all the reliability of a Renault Dauphine on a bad day, and often didn't work, brand new right out of the box!!!), pulls it off? Not likely!

So if they are cheap and unconcerned about product reliability and long-term customer satisfaction (just like every other French business), and they do fit the SHITTIEST valves to their products now, including cheap Chinese 300B's that would explode within minutes of turn-on when they first marketed a single-ended triode amplifier a number of years ago, despite the $13,000 pr. price tag for a pair of amplifiers that barely put out 10 Watts, if that...why would they put an OVERSIZED power transformer into their flagship amplifier? For that matter, take a look at their lesser JA series amplifiers, and notice that they, too, ALSO have uncharacteristically larger power transformers! I know what's going on here!

Too bad few if any others do, all the so-called "experts"! Regardless of what you decide to build or buy and modify, you can't go wrong with the advice I'm giving you. To summarize: -The larger the power transformer, the better the sound -The better the phase performance, the better the sound -The better the passive parts quality, the better the sound CAN be -NOS valves sound better than current production SHIT -Push-Pull amplifiers sound MUCH better than single-ended JUNK -Feedback is a VERY GOOD THING, but only if you've the brains to know how to use it properly!

Anyone who claims that "they don't like the sound" of feedback has just informed you that they are a stupid "tin ear", and that they are clueless on how to design amplification properly or reliably or both...YOU HAVE BEEN WARNED! HEED THE WARNING!

So Dude, I would look for a good 30-40W valve amplifier, but no larger. At least, not if you plan to either remove the QUAD's protection circuit at some point. And maybe that isn't such a good idea. I've two pairs of the speakers, and I did the same delay-line component upgrades and mechanical mods (I removed the metal screens and rolled-down the cloth grilles...and put them on better stands, but NOT the Arcici ones, which can "twist" the frames and cause panel arcing and failure!) to both.

But on one speaker, I kept the protection circuit in place. It was interesting to hear. Clearly, the protection circuit causes a fairly major and quite obvious degradation to the speaker's sound. But if you just remove it, without doing the delay-line upgrades, the sound is much the same! So, to get the ESL-63 to sound really, really TRANSPARENT, you have to upgrade the delay line (a tedious and delicate job, I put in no less than 150 polypropylene WIMA FKP1 capacitors in series combinations to get the small values required to replace the utterly-horrible sounding but otherwise extremely reliable ceramic originals.

The resistors were easier, I could replace them one for one. But I had to have them custom-made, as samples, from a wirewound resistor company in the US!), AND remove that delay line! And for your trouble, you are now too paranoid to try to extract the best possible sound from them, as an OTL amplifier can fail and cause over 1000 Euros in catastrophic damage, and almost CERTAINLY will over a period of time (depressingly, with OTL's, tube arc-over is more a question of WHEN, than IF!

I think the problem can be licked to a certain extent, but you would have to do a custom-design, and run the valves ULTRA-gently, such that a 100W OTL becomes a 20W one, with the same number and configuration of valves! But to explain why, would take yet ANOTHER long letter describing the intricacies of OTL operation!), and this damage occurs in the blink of an eye, a fraction of a second, because voltage, not heat, destroys the panels.

It takes TIME for a fuse to heat up, or a bi-metallic strip in a breaker to curl and flick itself into operation. But voltage surges are simply a function of their rise times, and typically we speak in terms of milliseconds, and maybe less than hundreds of them!!!!

Skoal! - Joe Rosen