Thanks for your reply, Koonw. I wondered what Tim Mellows meant by "matched pair". Are you saying that I should go ahead and buy a job lot of the Svetlana tubes and first match them for DC gain and then select for AC gain? I now see a bunch of 18 of them, brand new in boxes and manufactured in the USSR in different years between 1972 and 1985 quite cheap.
Yes, matched tube pairs is when the tubes have same gm(mA/V), and Ia or Ik is matched within few mA. Usually matching gm has priority over dc current match, but when it's direct coupled dc match become more important as the difference in Ik/Ia will determine the amount of idle DC offset (when the output is 0). Same gm ensured that DC offset can be maintained at higher output. DC offset of course can be adjusted with bias adjustment, but doing so may also altered the gm or gain of driver tubes. You should therefore also watch the DC offset at higher output level after DC offset adjustment.
Likewise different gm of output tube will cause higher DC offset at higher output level, but maybe somewhat compensated using different gm driver tube. But you don't really know which is which, so I think you need at least some known pairs for comparison, say you have matched pairs in front-end and driver, then you only know if the output tubes are matched or not. If not then you can replace the driver tubes see if they match again.
Likewise different gm of output tube will cause higher DC offset at higher output level, but maybe somewhat compensated using different gm driver tube. But you don't really know which is which, so I think you need at least some known pairs for comparison, say you have matched pairs in front-end and driver, then you only know if the output tubes are matched or not. If not then you can replace the driver tubes see if they match again.
Last edited:
Wiring
While I'm waiting to receive my Svetlana tubes from Russia, I took a few close-ups of the wiring. Any comments or criticisms are welcome. Point to point wiring always looks a bit messy compared with printed circuits. I have put a link to these shots on Google Drive.
Close-ups link for Google Drive.
While I'm waiting to receive my Svetlana tubes from Russia, I took a few close-ups of the wiring. Any comments or criticisms are welcome. Point to point wiring always looks a bit messy compared with printed circuits. I have put a link to these shots on Google Drive.
Close-ups link for Google Drive.
Yes, the gas-discharge tube is across the output after the speaker fuse. They have never actually blown a fuse even though the voltage has risen above the 90 V and are glowing brightly. This is most likely because while the output tubes are warming up, they can't supply enough current. Or of course, they may not actually work as well as they are supposed to. Anyway, I hope they don't come into operation while I'm running the amp in service.
Last edited:
Switch-off problem.
I have noticed one other problem now when I switch off the amplifier. With a dummy load of 39 Ohms, the offset rises to almost 2 Volts on both channels for about 10 seconds before it gradually goes down to zero. This would mean that I will get a good bit of static cone displacement on my speakers. It might not be the best thing for expensive Lowthers if I ever buy them. The PSU smoothing capacitors are obviously supplying this current as they discharge.
I have noticed one other problem now when I switch off the amplifier. With a dummy load of 39 Ohms, the offset rises to almost 2 Volts on both channels for about 10 seconds before it gradually goes down to zero. This would mean that I will get a good bit of static cone displacement on my speakers. It might not be the best thing for expensive Lowthers if I ever buy them. The PSU smoothing capacitors are obviously supplying this current as they discharge.
And you tooked it out or shorted it ?
Tim sais: "In the event of a
fault forcing the output stage latch either
up or down, R29(errato R33) provides a means
of limiting the current through the output
stage and loudspeaker. If its value
were too small, either an output tube or
the loudspeaker or both could be damaged.
If its value were too high, a small
offset voltage across the loudspeaker
could cause a significant imbalance in
the supply voltages HT2 and HT4."
I realy trust in every single ward the Tim's article
Tim sais: "In the event of a
fault forcing the output stage latch either
up or down, R29(errato R33) provides a means
of limiting the current through the output
stage and loudspeaker. If its value
were too small, either an output tube or
the loudspeaker or both could be damaged.
If its value were too high, a small
offset voltage across the loudspeaker
could cause a significant imbalance in
the supply voltages HT2 and HT4."
I realy trust in every single ward the Tim's article
I just took it out and wired the centre tap directly, but I think I'll put it back in now. Thanks
If you plan to use a low-power driver like an older Lowther or a PHY, IMO/IME its a good idea to build a fuse holder into the loudspeaker cabinet and fuse the speaker correctly as so many amplifiers can easily destroy such speakers!
My sim shows that EC90X blows 3.15A fuse in about 1mS assumed 8 ohms load and 150V dc out. If 3.15A fuse doesn't blow for some reasons, the 2A fuse will fuse in about 9mS (it takes longer as if ec90x is not functioning, fuse model used has time constant=5mS, use as reference only not actual values.
Last edited:
Measurements
I don't trust my cheap moving-coil milliameters so I'm going to put a small resistor in the anode or cathode connections to one or other of the O/P tubes and take leads to some Hirschmann 2mm measurement probe sockets. This will also make the measurements much more accurate.
I'm thinking of resistors of either 1.0 or 0.1 Ohms. Any suggestions about the value without affecting the performance?
I don't trust my cheap moving-coil milliameters so I'm going to put a small resistor in the anode or cathode connections to one or other of the O/P tubes and take leads to some Hirschmann 2mm measurement probe sockets. This will also make the measurements much more accurate.
I'm thinking of resistors of either 1.0 or 0.1 Ohms. Any suggestions about the value without affecting the performance?