And that's precisely the conundrum NJ: Why wasn't it addressed before being released? I mean, I'm (truly) hardly trying to put myself up as some sort of giant killer by being someone whose "so" capable, while Johnson and company were dufuses. The fix being as simple as it is would have added just a few extra bucks to the cost of building the unit -- which in terms of what these amplifiers sold for, would have been nothing. Other thoughts:
1. I know that back in 1983, ARC was still a pretty small company -- but obviously one well on its way. At that time, to my knowledge, all of the power amplifier designs up to at least and including the D115 era used the same basic balanced paraphase phase inverter with cross coupled driver stage concept. Oh they played with it a bit here and there, but the concept was the same -- and it is that basic concept that unchecked, produces the thunderous whallop in the speakers at shut down. The point I'm driving at, is that there were other low production marques (think Fairchild or Acrosound for example) that are still highly sought after today, due to the high intrinsic value they retain from the reputation they made for themselves. Against that, I don't believe there is the demand or hunt for early ARC power amplifiers like there is for these other marques, because the reputation of early ARC power amplifiers was real. In other words, there really was a problem.
2. I mentioned my studies so long ago of output tubes failing in the overall types of circuits that the early ARC designs represent. That wasn't just some amateurish, casual study of the problem. It was a very detailed study that involved about 5 dozen + tubes with detailed records and notes regarding configuration, number of power cycles, total time in use, and measured power output at numerous intervals for each tube in the study -- over the hundreds of hours that the study entailed. One of the out-growths of that study was the development of my power output tube tester, because I rather quickly found out that Gm testing simply didn't show up the kinds of damage being done by the destructive kinds of power cycling as being discussed here -- until a tube was so far gone that it was dangerous to use (i.e., prone to arcing, which very well spent tubes can be prone to do). Power output testing on the other hand is brutally honest, because it is a high level dynamic test (whcih Gm testing is NOT) that puts a spotlight on cathode condition. In comparison, Gm testing is merely a test for "characteristic", not condition.
What I truly didn't know 40 years ago -- that I am really only piecing together now because of this project -- is that all of the push and attention (at that time) to regulate this and regulate that, add hundreds if not thousands of µF in reserve power supply capacitance, and raise coupling caps to extremely large values -- was really the direct result of Johnson and ARC, and his new approach to designing vacuum tube amplifiers. As an avid "Audio Dimensions" fan of that day (whose associated audio salon in CA was an authorized ARC dealer), that approach was all the rage and discussion, as in Ike Eisenson's Book "Tu-be Or Not Tu-be", and in his monthly updates that followed -- that the few of us left still working with vacuum tubes, were working with. I was certainly aware of ARC at the time, and that their designs employed such techniques, but did not realize that they were the driving source behind it all.
Of course, when I started experimenting with such concepts, I immediately started having problems with output tube arcing and way-to-early failures, which I didn't understand, but wrote off as my lack of any fundamental knowledge as to what was going on at the time. I thought it was me doing something wrong. It was those events however that ultimately lead me on what amounted to a 10 year study (in between getting married, having kids, buying a house, cars, you know -- life...) to truly find out what was going on. Two things became eminently crystal clear and absolute to me from those studies that culminated in 1992: (1) Installing an "Instant Off" AND delayed make on power up output stage B+ control was absolutely imperative to such designs, to eliminate and protect the tubes from all the high powered bumps and thumps and resulting potential output tube arcing such designs produce -- and (2) When installing new lower ESR power supply caps (let alone increasing their value or adding regulation) into an existing design (i.e., modifying your Dynaco MK III, or Eico HF-60, or, or.....), then it became absolutely imperative to install what I originally coined as "Screen Stability" resistors. Yes, they are screen stopper resistors by any other name -- BUT -- what identifies them as Screen Stability resistors, is that they are screen stopper resistors added to designs that never originally included them -- which the MK III and HF-60 amplifiers did not. I mean, who would have thought that by simply increasing and/or improving the capacitance of the power supply, it would then require the addition of screen stopper resistors to prevent the tubes from arcing? Not me, nor apparently anyone else -- and most specifically, the wizards (said affectionately) at ARC, either!! Look at their designs throughout most of the 70s -- big power supply capacitance, their early regulation circuits, and the classic ARC small signal section with all its bumps and thumps -- yet in their two output tube models (think D76 or D51 series for example) -- not a screen stopper to be found! Their push-pull-parallel designs of the 70s (were there any?) would have had screen stoppers out of default, because they are required to keep such configurations stable. Now by the 80s, their two output tube models did include screen stoppers (think D70 -- the little brother of the D115), which was a great first step to guard against output tube arcing from power supplies that were regulated or had massive amounts of capacitance -- except I was hardly following detail changes in ARC design because remember, I had no idea that they were the focal point of all this at the time. Point being, I didn't know to install them just because ARC started doing so. So I independently determined on my own that screen stopper resistors were required in all new designs of the highly regulated/high P/S capacitance type, and that Screen Stoppers were required for so modified equipment that never included them in the original design. But of course, that was only half of the problem. The other half -- and every bit as important -- was the output tube B+ control -- which to my knowledge, ARC never implemented, retrofitted, or even by all appearances, had a clue to do.
When the internet came on the scene and AK became a reality, I started reading of all these folks who upgraded (or restored) their amplifiers with new and/or greater amounts of power supply capacitance (or both), and also usually did the "recommended" increase in coupling capacitance size (adding insult to injury), but damn, now my output tubes are arcing -- and I don't know why -- help! There are a number of factors involved including the type of output tubes used, how they are operated, how they are biased, etc., but where there had been no arcing before and now there was, I knew exactly why, and started answering to install Screen Stability resistors, which to everybody's amazement (at the time) in fact solved the problem. Now today, they are almost universally installed matter of factly, and the whole issue of output tube arcing has generally gone away unless a tube itself went south, or someone keeps trying to use a tube that arced -- "but still tests good"........do NOT do that, as once a tube has arced, it is forever prone to do so no matter the precautions taken.
Somewhere along the line, the powers that be at AudioXpress became aware of my work, and asked me to write up an article regarding all of this for their digital edition, which to date and my knowledge -- is STILL the ONLY written and published source on this topic to this day. I hardly mention any of this in an effort to impress, but only to emphasize this fact: Throughout the 70s and 80s, nobody seemed to really know what was going on regarding these output tube failures, and given all the hindsight now, it is quite possible if not probable that neither did ARC -- who started the whole"movement"! Maybe it just became an acceptable result of collateral damage to be accepted for the improved sound produced -- and make no mistake, there was and is still, universal recognition as to the importance of power supply performance -- courtesy of W. Johnson and company.
So I was (apparently and unknowingly at the time) on a parallel path with ARC in trying to find out the causes involved. We both ultimately recognized the need for screen stopper resistors when "new age" type power supplies were used, so that ARC started including them where they weren't originally included in their earlier, similar designs, and I independently started recommending the installation of Screen Stability resistors where stoppers weren't originally included. But where as I then went on to identify a simple/workable resolve for the second problem of all the output tube pulsing and speaker thumping at shutdown.............
3. ARC ultimately just quietly moved away from the balanced paraphase phase inverter/cross-coupled driver stage, and moved on to multiple cascade connect constant current diff amps, that achieved all the advantages of the original cross-coupled design, but didn't need the up-front single ended paraphase inverter to assist it. Such a design therefore won't have the bumps and thumps at shutdown. The VT100 is a good example of this which I was looking at for another AKer. Oh the older type design was still around in the 90's, but all the input stages and balanced paraphase phase inverter stage had been converted to mosfets by then. And, ARC did resolve the shutdown pulses and thumps, by converting the the B+ regulators for the small signal stages to passive SS designs. That is, being all solid state (no error amplifier or pass tube heaters to cool down), and drawing their operating current from the very source voltage they are ultimately dropping and regulating, the regulators then don't immediately quit at shutdown -- not until the voltage supplying them has dropped 100 volts or so down to their regulation voltage level -- which takes time -- and particularly so given the amount of capacitance ARC uses at the output of the B+ rectifiers. And, ARC had even further increased the already significant bank of capacitance there, making the voltage drop even slower. Of course, the input section mosfets also operate at much lower voltages, too. The collective result is, that by the time the regulators for the small signal stages fall out of regulation, the output tubes have cooled down enough such that any pulse finally created by the SS input stages is of little consequence -- the output tubes are too cool to produce much of anything applied to them by then. Gerrry, your Classic 120s use this latter approach, while still using the old style input configuration.
So the problem was largely the fully vacuum tube (or nearly so) products of the 70s and 80s, which today, the timer and delay relay very effectively resolves for those (few?) units still left in existence. ARC did finally address by the 90s. But as Gordon so aptly described, by then, the damage for newer audiophiles of means that got a taste of the earlier products, was done. Once they dug out of that hole, then they fell into another with their full speed ahead approach in designing all of their products to operate with balanced inputs. And yet still, ARC survived, and thankfully continues to do so. Missteps or missed opportunities aside, their innovation and drive that kept the vacuum tube portion of the high fidelity industry alive when it was otherwise all but dead, and the fact that it's still is a driving force today, is absolutely undeniable.
Dave
