Tweet
My apologies for the long absence. I now am into my second week after the last cataract surgery, and can finally see clearly without the visual obstruction that was complicating my daily existence. Still having to get assistance in moving the heavy amps up to the bench and then back off for the next month.
I presently have a Fender Super Bassman that came in from a steady client, with the complaint that the Overdrive function didn’t start working until the amp had been on for around 15 min. I never found that to be true, but I did find a RED light on the rear panel’s AutoBias display. I knew one of these days, I’d have to deal with power tubes and that circuit.
After removing the chassis screws to release the T-shaped power amp chassis, the middle rear Groove Tube 6550 had a major crack running half way around the tube, now showing frosty white insides. Yesterday, finally getting back to this project, I pulled the chassis out to see how to deal with the mechanics of preamp harness cable being too short to flip the chassis on end to access the power tube PCB. I’d need extender cables for the six tubes to read the cathodes of each. 1 ohm Cathode resistors are used on each power tube, 470 ohm screen resistors, and no plate resistors that I can see from the bottom of the PCB. I don’t actually have the Super Bassman schematic yet.
I ended up making two extender cables using some Molex 5-pin female connectors, having 0.093” ID connectors, suitable for probing, each connector having a ground.
I tack-soldered the cable wires onto each of the cathode pins and grounds of two of the cathode resistors. I began with the two pairs installed…1A, 1B, & 3A, 3B and took readings off the cathodes.
1A 34.9mA
1B 29.2mA
3A 22.4mA
3B 25.6mA
I then installed the working tube 2A into its’ socket, and, finding these GT 6550’s look like Sovtek’s, I went thru some GT6550 pulls I had. 2A read 20.1mA, with the next three pulls for 2B reading 14.7mA, 15.5mA and 6.8mA. I then found one that read 25.0mA, so I was in the ballpark, perhaps. When I read the set again, I found:
1A 29.5mA
1B 30.5mA
2A 18.4mA
2B 25.0mA
3A 20.5mA
3B 24.0mA
When I swapped 2B with 3A, the results got worse.
1A 25.0mA
1B 29.2mA
2A 15.3mA
2B 26.6mA
3A 14.9mA
3B 23.3mA
I then swapped 2A with 3B, but the Auto-Bias still countered my attempt at grouping.
1A 25.2mA
1B 28.9mA
2A 15.7mA
2B 26.4mA
3A 14.9mA
3B 24.4mA
During these moves, I had pushed both the upper and lower ARROW’s to reset the bias to Factory setting. It still came out as listed above. The Average current for the ‘A’ in this last attempt was 18.6mA, and 26.6mA for the ‘B’ set, a 42.8% difference.
Feeling I wasn’t getting anywhere with these tubes, I tried a quad set of matched J/J KT88 tubes I had on hand. Actually, one of that set had failed, and I got a replacement from Antique of that Apex-Matched set that was the closest they had. This was the first time I had actually tried them out, having set them aside for an Orange AD-200 amp of the same client. They were close, +/- 1mA of 26mA. I tried to see if I could get two other tubes to match them, but no luck.
AND, I found Fender did NOT set their mechanical tube spacing to accommodate KT88 glass bottles! With the J/J KT88’s, the bottles were within 0.050” of the front-rear pairs. Plenty of space side-by-side, but using those is just asking for a complete set of broken tubes!!
I went back to the GT6550’s, and tried a different matching. From the original grouping, I swapped 1A with 1B, left 2A alone, moved 2B to 3B, 3B to 3A and 3A to 2B. That yielded an equal balance in each triad set, though with a wide difference in currents in each set:
1B 27.5mA
1A 31.3mA
2A 23.5mA
3A 22.3mA
3B 22.8mA
2B 20.3mA
The average current of each set is 24.6mA, while having mismatch in current sharing of 54% in the ‘A’ group and 23.3% in the ‘B’ group. Though, after half an hour, the ‘A’ group dropped to an average of 20.4mA, while the ‘B’ group remained about the same, still averaging 24.6mA.
I’ve no idea just how Fender’s Auto-Bias circuit works or arrives at such balances. And, all of this is just at idle. I haven’t made any power measurements under load, which I understand Groove Tubes does in selecting their matching over a wide range of currents & at several frequencies. It just bothers me to have this Auto-Bias circuit counteract my attempts to group tubes for a best plate current balance. I have better luck with the Ampeg SVT amps in that regard.
I also did a brief check on line to see if I could find how to replace a given Groove Tube 6550 power tube having the Matching Code, seeking a replacement to match within the set. I had no luck. I've been able to obtain replacement tubes thru Antique Electronics, with their Apex Matching system. This, of course, was prior to my putting the aged GT 6550 tubes back in, not knowing how far off the plate current balance was to begin with! It was reassuring to find the matched J/J KT88's did come up closely matched...just not usable due to the potential breakage problem on this amp.
I presently have a Fender Super Bassman that came in from a steady client, with the complaint that the Overdrive function didn’t start working until the amp had been on for around 15 min. I never found that to be true, but I did find a RED light on the rear panel’s AutoBias display. I knew one of these days, I’d have to deal with power tubes and that circuit.
After removing the chassis screws to release the T-shaped power amp chassis, the middle rear Groove Tube 6550 had a major crack running half way around the tube, now showing frosty white insides. Yesterday, finally getting back to this project, I pulled the chassis out to see how to deal with the mechanics of preamp harness cable being too short to flip the chassis on end to access the power tube PCB. I’d need extender cables for the six tubes to read the cathodes of each. 1 ohm Cathode resistors are used on each power tube, 470 ohm screen resistors, and no plate resistors that I can see from the bottom of the PCB. I don’t actually have the Super Bassman schematic yet.
I ended up making two extender cables using some Molex 5-pin female connectors, having 0.093” ID connectors, suitable for probing, each connector having a ground.
I tack-soldered the cable wires onto each of the cathode pins and grounds of two of the cathode resistors. I began with the two pairs installed…1A, 1B, & 3A, 3B and took readings off the cathodes.
1A 34.9mA
1B 29.2mA
3A 22.4mA
3B 25.6mA
I then installed the working tube 2A into its’ socket, and, finding these GT 6550’s look like Sovtek’s, I went thru some GT6550 pulls I had. 2A read 20.1mA, with the next three pulls for 2B reading 14.7mA, 15.5mA and 6.8mA. I then found one that read 25.0mA, so I was in the ballpark, perhaps. When I read the set again, I found:
1A 29.5mA
1B 30.5mA
2A 18.4mA
2B 25.0mA
3A 20.5mA
3B 24.0mA
When I swapped 2B with 3A, the results got worse.
1A 25.0mA
1B 29.2mA
2A 15.3mA
2B 26.6mA
3A 14.9mA
3B 23.3mA
I then swapped 2A with 3B, but the Auto-Bias still countered my attempt at grouping.
1A 25.2mA
1B 28.9mA
2A 15.7mA
2B 26.4mA
3A 14.9mA
3B 24.4mA
During these moves, I had pushed both the upper and lower ARROW’s to reset the bias to Factory setting. It still came out as listed above. The Average current for the ‘A’ in this last attempt was 18.6mA, and 26.6mA for the ‘B’ set, a 42.8% difference.
Feeling I wasn’t getting anywhere with these tubes, I tried a quad set of matched J/J KT88 tubes I had on hand. Actually, one of that set had failed, and I got a replacement from Antique of that Apex-Matched set that was the closest they had. This was the first time I had actually tried them out, having set them aside for an Orange AD-200 amp of the same client. They were close, +/- 1mA of 26mA. I tried to see if I could get two other tubes to match them, but no luck.
AND, I found Fender did NOT set their mechanical tube spacing to accommodate KT88 glass bottles! With the J/J KT88’s, the bottles were within 0.050” of the front-rear pairs. Plenty of space side-by-side, but using those is just asking for a complete set of broken tubes!!
I went back to the GT6550’s, and tried a different matching. From the original grouping, I swapped 1A with 1B, left 2A alone, moved 2B to 3B, 3B to 3A and 3A to 2B. That yielded an equal balance in each triad set, though with a wide difference in currents in each set:
1B 27.5mA
1A 31.3mA
2A 23.5mA
3A 22.3mA
3B 22.8mA
2B 20.3mA
The average current of each set is 24.6mA, while having mismatch in current sharing of 54% in the ‘A’ group and 23.3% in the ‘B’ group. Though, after half an hour, the ‘A’ group dropped to an average of 20.4mA, while the ‘B’ group remained about the same, still averaging 24.6mA.
I’ve no idea just how Fender’s Auto-Bias circuit works or arrives at such balances. And, all of this is just at idle. I haven’t made any power measurements under load, which I understand Groove Tubes does in selecting their matching over a wide range of currents & at several frequencies. It just bothers me to have this Auto-Bias circuit counteract my attempts to group tubes for a best plate current balance. I have better luck with the Ampeg SVT amps in that regard.
I also did a brief check on line to see if I could find how to replace a given Groove Tube 6550 power tube having the Matching Code, seeking a replacement to match within the set. I had no luck. I've been able to obtain replacement tubes thru Antique Electronics, with their Apex Matching system. This, of course, was prior to my putting the aged GT 6550 tubes back in, not knowing how far off the plate current balance was to begin with! It was reassuring to find the matched J/J KT88's did come up closely matched...just not usable due to the potential breakage problem on this amp.