Post by michaelw on Jun 20, 2017 13:49:35 GMT 12
counterpoint was right at the bleeding edge of reliability.
i'm surprised mine lasted for the 3-4 years i had it.
of course snapped up the sp8/2 as soon as it appeared.
from the counterpoint distributor no less !
i still have memories of burnt fingers setting the bias on my old quicksilver monos.
the single adjustment pot was close to the 8417 output tubes.
the trick was to push the phono plug probe in just far enough to get a reading without it locking.
the quicksilver was another great sounding, simple design that was a nightmare to retube.
|
Post by colinf on Jun 20, 2017 14:55:58 GMT 12
Yes Trevor used to bring in the Quicksilver parts from CA and John and I helped build them in Melbourne. It was tedious to bias them and we adjusted them exactly as you mention. The input 12ax7 was critical for gas and gm. The best way to select it was to insert a 10M resistor in series with the control grid, put in a triangle wave of about 200hz and observe the waveform on the output. You'd select the tube for lowest distortion and best clipping symmetry.
AMR-iFi R&D
|
Post by guitardude on Jun 20, 2017 15:18:16 GMT 12
Ha, I too had those quicksilver Mono's with the 8417. Nice amp when it was going but a nightmare for reliability, I was forever buying new tubes for it and getting it serviced. Got sick of all the hassle of it in the end. (shame I didn't know colinf then !)
|
Post by michaelw on Jun 20, 2017 16:25:04 GMT 12
at one time i had the entire stock of 8417s in nz !
alan gore/ariel engineering bought them off the original nz quicksliver distributer, audio excellence.
out of that half dozen, their were no matched sets !
alan modified the quicksilver bias circuit so that each output tube had its own adjustment.
that solved the need for matched sets.
but mild ocd set in with checking bias almost daily.
great sound though.
only other reliability issue was an exploding philips cap.
a friend replaced the lot with fancy wondercaps.
|
Post by colinf on Jun 20, 2017 17:00:10 GMT 12
We tried fiddling with different coupling caps too but in the end we thought the original yellow Quicksilver caps supplied sounded pretty good. The Philips ECG 8417s were best. GEs seemed not to like the Quicksilver circuit and drifted to their deaths more often than not.
AMR-iFi R&D
|
Post by colinf on Jun 20, 2017 17:20:17 GMT 12
Just checked, Roger still has 8417s, for a price! US$150 each.
AMR-iFi R&D
|
Post by michaelw on Jun 20, 2017 18:32:58 GMT 12
he can keep them ! eventually even quicksilver got fed up and switched to el34 and kt88. mine had philips, the dodgy replacements were all ge. i don't remember if we replaced the quicksilver labelled caps or not... still have a non-exploded philips cap...   image hosting api
|
Post by Owen Y on Jun 20, 2017 19:09:49 GMT 12
I'd always thought of those common blue Philips electrolytics were pretty good quality? Maybe not. Or maybe old, or under-rated, or used in a high temperature environment? Was that in the Neg. Bias PSU?
|
Post by michaelw on Jun 20, 2017 20:10:10 GMT 12
dunno to all of the above  one exploded on turn-on leaving the inside of the amp peppered in little bits of papery stuff. repaired by distributor with a different cap !
|
Post by colinf on Jun 20, 2017 21:04:15 GMT 12
They were in the bias supply after the half wave rectifier. So perhaps the diode failed and sent ac through the cap, making it explode.
AMR-iFi R&D
|
Post by guitardude on Jun 21, 2017 8:59:01 GMT 12
I ran the Philips (blue writing ?) in mine, but often went thermonuclear whilst adjusting the bias. Theres goes another pair !
|
Post by Owen Y on Jun 21, 2017 10:11:29 GMT 12
(This deserves its own Quicksilver thread  ......) Here's the Quicksilver Mono schematic I think (some of it) - a Beam Tetrode Push-pull amp wired in 'conventional' Ultralinear OPT config, with Fixed Biased output tubes & Negative Feedback. Why were these reputedly so good sounding?.... Simplicity?...a simple 3-stage cct, Direct Coupled first 2 stages, simple phase-splitter, Fixed Bias/removal of Cathode Bypass cap....? PSU? 
|
Post by michaelw on Jun 21, 2017 10:53:03 GMT 12
and they run in correct phase. unlike counterpoint (and conrad-johnson et al). sorry for deviating from the arc program. maybe split the thread into a new "tube ailments and horror stories" line. 
|
Post by guitardude on Jun 21, 2017 13:57:54 GMT 12
I found it a bit warmer and sweeter than the D70ii, unfortunately I was also expecting similar levels of reliability as was fooled into expecting all high end US valve gear to be ARC like in that regard. The quicksilvers, VTL and CJ amps were to prove anything but....
|
Post by michaelw on Jun 21, 2017 14:09:32 GMT 12
aside from the philips explosion and trying to get decent tube replacements, my quicky was great.
at the time it made mincement of the big heavy duty amps, nz and imports.
and was happy with speakers as diverse as maggies, quads, ar, acoustat, martin logan, proac etc.
|
Post by colinf on Jun 21, 2017 14:53:02 GMT 12
 These amps had a very simple circuit. They sounded good for a few reasons: the circuit was built point to point. Also, the 8417 is a high transconductance output valve so the input stage drive requirements are less. That means you could get away with one gain stage on the input and consequently less negative feedback. There was no resistor on the output screens which makes the 8417 sound lush. The output transformer was wound with many paralleled sections to cut back leakage inductance so the frequency response was reasonably good. This amp also had a few dangerous flaws. If a tube decided to arc over (quite common on this amp, it runs at 480v HT) it would fail spectacularly in meltdown fashion, instead of blowing a fuse or tripping a circuit breaker which should have been designed into the output transformer centre tap HT supply. Also the bias circuit would collapse quickly and charge back up slowly if the power was interrupted for a very short time, giving the still hot output tubes high bias and a hard time. That also stressed out the 5ar4 rectifier tubes, which are running hard anyway given that they have to charge over 3 times the capacitance on the power supply that they were designed for originally.
AMR-iFi R&D
|
Post by Owen Y on Jun 21, 2017 15:56:26 GMT 12
( colinf - thanks for opening a new Quicksilver Mono thread - I have taken the liberty of moving some of the previous related posts from the ARC thread. Cheers.)
|
Post by michaelw on Jun 21, 2017 18:46:45 GMT 12
is that diagram for the original 5 tube version ?
|
Post by Owen Y on Jun 21, 2017 19:08:32 GMT 12
Good question.... That schematic looks like 6 toobs per ch - (1/2) 12AX7 -> (1/2) 12AU7 -> 8417 PP. Plus I believe 2 x GZ34/5AR4 rectifier tubes per ch  . (Did they wire the 12AX7 12AU7 differently on ea mono chassis, so that you could swap these tubes across chasses now & then, in order to use both halves of ea dble triode?)
|
Post by colinf on Jun 21, 2017 20:54:53 GMT 12
The 12ax7 and 12bh7 were wired on one half only, and all the mono amps were made the same.
AMR-iFi R&D
|
Post by Owen Y on Jun 21, 2017 22:05:50 GMT 12
That would seem to be a waste of (1/2 of) the twin triodes, do you think? (Chaps, if you can visualise only one side of the twin triodes being wired up & utilised one half only.) If they were wired mirror-fashion, L-R chs, they could be swapped occasionally, to get maybe twice as much mileage out of ea tube. Also, with a common heater element heating both halves of these twin triodes, one half is sitting there heated without plate voltage or current flow - which (as I recall) can slowly give rise to a phenonemon called 'cathode poisoning', when the unused tube cathode chemically slowly reduces in emission. (I'm no expert but I'm guessing that there must be some discussion about this?)
|
Post by michaelw on Jun 21, 2017 23:07:01 GMT 12
is that the same as tube sleeping disease ?
where were the 5ar4s in that diagram ?
|
Post by colinf on Jun 22, 2017 8:07:08 GMT 12
Only one half of each 12ax7 and 12bh7 heater was wired as well. It gave some observant customers something to ask when only one half of each input tube was glowing. As the amps were sold in pairs we rewired some of them so that you could swap the input tubes from channel to channel and use the better performing triode half. The 5ar4s aren't visible in this circuit diagram. There were two in the 12ax7-12bh7 version. I think the original Quicksilver used a single 12dw7 input tube as gain stage and phase splitter, and used a single 5ar4 rectifier.
AMR-iFi R&D
|
Post by michaelw on Jun 22, 2017 9:24:15 GMT 12
thanks colin. i knew it wasn't time to visit specsavers did you ever deal with the quicksilver full function preamplifier ? i was tempted but the twin volume controls put me off. same reason i skipped the audible illusions modulus preamps.
|
Post by Owen Y on Jun 22, 2017 10:00:03 GMT 12
OTOH, I have liked having twin vol controls - but only if they are stepped/detented. Otherwise, they WOULD be a nuisance. I find it helps on occasion to have separate level/balance control L-R channels.
|
Post by guitardude on Jun 22, 2017 10:17:48 GMT 12
With all these gripes about reliability it's no wonder beeman had a bit of a job selling his !
I too had a preamp once with twin vol controls and no detents: it was a right PITA. Unfortunately I cant quite remember what it was. VTL or Melos I think. The Audible Illusions was supposed to be very good at the time.
|
Post by michaelw on Jun 22, 2017 10:58:43 GMT 12
the quicksilver volume controls were marked and stepped 
|
Post by colinf on Jun 22, 2017 11:59:02 GMT 12
We had a Quicksilver preamp at the shop but I can't remember much about it unfortunately. It's great to see a picture of it again. Twin volume controls are better electronically because it means the crosstalk between the pot sections is eliminated. It's easier to set the volume on stepped ones. But for convenience I like a volume control that works both channels. Adding a balance control means adding another pot in the signal path that you can hear. I also don't mind using concentric volume controls.
AMR-iFi R&D
|
Post by colinf on Jun 23, 2017 10:48:37 GMT 12
It's possibly not helpful but I've been looking at equivalents for the 8417. Although Quicksilver later released a 6550 mono amp, this tube didn't operate quite the same as the 8417. It has half the transconductance and half the triode connected gain which requires higher drive voltage, and consequently more distortion generated by the input stage. And different impedance on the output transformer. The 7591 S is a current tube from JJ and has almost exactly half the transconductance and half the heater current, and the same triode gain, of an 8417. So paralleling two of them would be equivalent to one 8417. Of course it's probably not possible to fit four 7591s into the space of two 8417s on a Quicksilver mono amp. The pin connections are different too. But for a new design a few 7591s in parallel could be nice.
AMR-iFi R&D
|
Post by michaelw on Jun 23, 2017 10:58:53 GMT 12
back in the day i looked for replacements but there was no 8417 equivalent to be found. ariel engineering told me with some fiddling, along the lines that quicksilver eventually took themselves, the monos could be made to work with el34, kt88, 6550 etc. by then i had had enough of tubes and a new krell was on it's way. a saucy jadis snuck in to satisfy those occasional tube pangs. sidebar - did anyone else use the 8417 as an output tube ? afaik the only other audio company using them was dynaco. i believe it was originally a tv tube ?
|
......)
.