Got in a pair of Dynaco Mk III amps for a refresh. Obviously, they had not been used for a while (one was missing output tubes)...
From first glance, someone had obviously been meddling... er, modifying these amps in the past. Whenever you see a jumper and/or open solder pins (components removed) on a Mk III board, you know something is NOT stock.
The other thing- and this was a surprise- was that this was the first Mk III set I've personally encountered, where an output transformer was actually BAD- one plate lead read about double the impedance of the other ones (in this amp, and both of them in the other amp- those all basically matched, with this one lead being way out of spec).
As I got further into them, I discovered just how much had been changed- and how ill-advised (and dangerous) these mods could be. Here's a schematic that I reverse-engineered from it, and below that, the actual CORRECT factory schematic:
Coupling caps of FOUR TIMES the value of the original- which pretty much tanks the LF stability...
The feedback cap on the screen lead completely REMOVED (no 390pf, or any other value, cap where it was supposed to be), so that most of the phase and amplitude compensation that made the high frequency response stable, wasn't even there any more. Apparently, someone had increased the size of the compensation cap on the plate of the plate of the 6AN8 pentode to try and get back some semblance of stability... but there really is no way, that that one cap can do both its original job, AND what the 390pf cap was supposed to do.
BTW- my speculation, is that the instability that was likely happening here, due to the above changes, was probably the direct cause of the bad/damaged output transformer... motorboating and or HF oscillation can wreak havoc on a transformer, sometimes...
Speaking of caps- the 1uf output coupling caps? 400v. That's a problem. The phase inverter typically operates at about 365v on the plate- and possibly more, if the wall line voltage is above 120v at all- and there's about -50v of BIAS on the other side of the output coupling cap that's connected to the plate. So, that cap is IDLING at about 415V DC across it!! Not good, for a 400v-rated cap...
Last, but not least- the two most bizarre mods, were resistor substitutions. The 470K grid bias reference/input load resistor on the input was replaced by a 4700 ohm! This meant that the input impedance of the amp was only 4.7K- making the amp pretty much useless on anything but a low-output-impedance solid state preamp- no tube preamp need apply, pretty much, And, to top everything off- the 680 ohm cathode bias resistor on the 6AN8 pentode (which provides the majority of the bias for that tube), was simply REMOVED. I mean, a JUMPER soldered in place of it! The net effect of that, is that the 6AN8 pentode was drawing inordinately too much current- dropping the plate voltage of that section- which, since it's direct coupled to the grid of the phase inverter- almost completely biasing the phase inverter at cutoff. So much for any voltage headroom in those stages!
Needless to say, I didn't even think about attempting to run the amp like that. Installed a Triode Electronics A431S to replace the bad output transformer, and back to stock it goes!
I did, for the sake of capacitor safety, install one slight mod after returning everything else to stock- a 330K 2W bleeder resistor, across the phase inverter supply. This both acted as a bleeder to drain the caps at shutdown (a good thing if the amp is to be left with the top cover off at any time), and to slightly reduce the startup and working voltage on the inverter power supply cap (with modern line voltage of over 120V AC from the wall, it could go a bit over rated cap voltage at startup). A good thing in terms of cap life...
Regards,
Gordon.
From first glance, someone had obviously been meddling... er, modifying these amps in the past. Whenever you see a jumper and/or open solder pins (components removed) on a Mk III board, you know something is NOT stock.
The other thing- and this was a surprise- was that this was the first Mk III set I've personally encountered, where an output transformer was actually BAD- one plate lead read about double the impedance of the other ones (in this amp, and both of them in the other amp- those all basically matched, with this one lead being way out of spec).
As I got further into them, I discovered just how much had been changed- and how ill-advised (and dangerous) these mods could be. Here's a schematic that I reverse-engineered from it, and below that, the actual CORRECT factory schematic:
Coupling caps of FOUR TIMES the value of the original- which pretty much tanks the LF stability...
The feedback cap on the screen lead completely REMOVED (no 390pf, or any other value, cap where it was supposed to be), so that most of the phase and amplitude compensation that made the high frequency response stable, wasn't even there any more. Apparently, someone had increased the size of the compensation cap on the plate of the plate of the 6AN8 pentode to try and get back some semblance of stability... but there really is no way, that that one cap can do both its original job, AND what the 390pf cap was supposed to do.
BTW- my speculation, is that the instability that was likely happening here, due to the above changes, was probably the direct cause of the bad/damaged output transformer... motorboating and or HF oscillation can wreak havoc on a transformer, sometimes...
Speaking of caps- the 1uf output coupling caps? 400v. That's a problem. The phase inverter typically operates at about 365v on the plate- and possibly more, if the wall line voltage is above 120v at all- and there's about -50v of BIAS on the other side of the output coupling cap that's connected to the plate. So, that cap is IDLING at about 415V DC across it!! Not good, for a 400v-rated cap...
Last, but not least- the two most bizarre mods, were resistor substitutions. The 470K grid bias reference/input load resistor on the input was replaced by a 4700 ohm! This meant that the input impedance of the amp was only 4.7K- making the amp pretty much useless on anything but a low-output-impedance solid state preamp- no tube preamp need apply, pretty much, And, to top everything off- the 680 ohm cathode bias resistor on the 6AN8 pentode (which provides the majority of the bias for that tube), was simply REMOVED. I mean, a JUMPER soldered in place of it! The net effect of that, is that the 6AN8 pentode was drawing inordinately too much current- dropping the plate voltage of that section- which, since it's direct coupled to the grid of the phase inverter- almost completely biasing the phase inverter at cutoff. So much for any voltage headroom in those stages!
Needless to say, I didn't even think about attempting to run the amp like that. Installed a Triode Electronics A431S to replace the bad output transformer, and back to stock it goes!
I did, for the sake of capacitor safety, install one slight mod after returning everything else to stock- a 330K 2W bleeder resistor, across the phase inverter supply. This both acted as a bleeder to drain the caps at shutdown (a good thing if the amp is to be left with the top cover off at any time), and to slightly reduce the startup and working voltage on the inverter power supply cap (with modern line voltage of over 120V AC from the wall, it could go a bit over rated cap voltage at startup). A good thing in terms of cap life...
Regards,
Gordon.
