Some time ago i was given a lot of small-signal tubes at the local parts store. They couldn't sell them because they didn't have many of each and most of them weren't famous tubes.
I built some grounded cathode gain stages and found them to sound different from each other but without a clear winner. I couldn't also do fair comparisons because it took a lot of time to change the circuit to suit each time.
Once i have found a solid-state output stage that i like, i have had the idea to do "gaincards" (sorry for the joke) with each of the tubes. The cards would use something similar to an isa socket (obviously smaller) and have +300V, +200V, ground, input, output, +12V +9V +7V and +6.3V (If one tube requieres a different voltage it could be stepped down on-board). The two pins at one edge would be shorted and cutted to half the lenght of the socket to get sure they are the last to make contact when the board is attached and the first to disconnect when the board is removed. They should trigger the delayed-turnon (solid state amps are mercyless when driven by cold tubes) and the (obviously immediate) turnoff. This should allow "hotswapping".
The advantages would be no high voltage on the socket when there is nothing on it and the possibility of rolling tubes without touching them, nice for the tube and nice for your fingers. The cards could be boxed in a nice-looking way that made the socket even less accessible for the fingers.
This would also work with more than one tube gain stages, making something like a "moskido" and allow to improve the output stage while keeping all the tube cards.
While standarizing this so everybody uses the same pinout seems pointless since the boards are extraordinarily easy to do, i would like to share the idea and to ask for suggestions before start building the boards, so i don't miss something and realize it once i have already done many of them.
Should i place the delayed-turnon resistor on the boards or i can assume that all the tubes will be ready in a reasonable time, say 10/15 seconds?
I built some grounded cathode gain stages and found them to sound different from each other but without a clear winner. I couldn't also do fair comparisons because it took a lot of time to change the circuit to suit each time.
Once i have found a solid-state output stage that i like, i have had the idea to do "gaincards" (sorry for the joke) with each of the tubes. The cards would use something similar to an isa socket (obviously smaller) and have +300V, +200V, ground, input, output, +12V +9V +7V and +6.3V (If one tube requieres a different voltage it could be stepped down on-board). The two pins at one edge would be shorted and cutted to half the lenght of the socket to get sure they are the last to make contact when the board is attached and the first to disconnect when the board is removed. They should trigger the delayed-turnon (solid state amps are mercyless when driven by cold tubes) and the (obviously immediate) turnoff. This should allow "hotswapping".
The advantages would be no high voltage on the socket when there is nothing on it and the possibility of rolling tubes without touching them, nice for the tube and nice for your fingers. The cards could be boxed in a nice-looking way that made the socket even less accessible for the fingers.
This would also work with more than one tube gain stages, making something like a "moskido" and allow to improve the output stage while keeping all the tube cards.
While standarizing this so everybody uses the same pinout seems pointless since the boards are extraordinarily easy to do, i would like to share the idea and to ask for suggestions before start building the boards, so i don't miss something and realize it once i have already done many of them.
Should i place the delayed-turnon resistor on the boards or i can assume that all the tubes will be ready in a reasonable time, say 10/15 seconds?
This is the pinout i plan to use in the amp:
1. Card detection (short pads)
2. 0V Filament
3. 6.3V 1A
4. 7V 1A
5 8V 1A
6 9V 1A
7 12V 1A
8. OV Cathode
9. Line-level input 1 200K
10. Line-Level input 2 200K
11. 0V
12. B+ 300V
13. B+ 200V
14. CCS 1 [0-50 mA 300V-0V short-circuit proof] [MJE15035]
15. CCS program 1 : I = 0.7/R
16. CCS 2 [0-50 mA 300V-0V short-circuit proof] [MJE15035]
17. CCS program2
18. Output 1 [impedance > 250KOhm 25 pF] [AC coupled]
19. Output 2 [impedance > 250KOhm 25 pF] [AC coupled]
20. Card detection (short pads)
20 will be shorted to 1.
Does somebody see any problem?
1. Card detection (short pads)
2. 0V Filament
3. 6.3V 1A
4. 7V 1A
5 8V 1A
6 9V 1A
7 12V 1A
8. OV Cathode
9. Line-level input 1 200K
10. Line-Level input 2 200K
11. 0V
12. B+ 300V
13. B+ 200V
14. CCS 1 [0-50 mA 300V-0V short-circuit proof] [MJE15035]
15. CCS program 1 : I = 0.7/R
16. CCS 2 [0-50 mA 300V-0V short-circuit proof] [MJE15035]
17. CCS program2
18. Output 1 [impedance > 250KOhm 25 pF] [AC coupled]
19. Output 2 [impedance > 250KOhm 25 pF] [AC coupled]
20. Card detection (short pads)
20 will be shorted to 1.
Does somebody see any problem?
If using AC for filaments, you need only a series capacitor
to react away the excess voltage... If using DC filaments,
you need only the appropriate resistor or zenier across
two terminals to program the power supply.
You need supply only 300V and a bypass cap (to Gnd).
The card need only a series drop resistor between B+
and the bypass cap terminals to set working voltage...
You might wish to attenuate the input to match the gain
of each card to some uniform low figure that all can do...
I am assuming you won't want to attenuate output or
abuse neg feedback in the cathode.
to react away the excess voltage... If using DC filaments,
you need only the appropriate resistor or zenier across
two terminals to program the power supply.
You need supply only 300V and a bypass cap (to Gnd).
The card need only a series drop resistor between B+
and the bypass cap terminals to set working voltage...
You might wish to attenuate the input to match the gain
of each card to some uniform low figure that all can do...
I am assuming you won't want to attenuate output or
abuse neg feedback in the cathode.
kenpeter said:
About the filaments, i have always been biased towards dc and cheap regulators, all on the same heatsink as they will never operate at once. Stepping down 12V to 6.3 starts looking like serious dissipation for a resistor.
I was thinking in using tubes with a mu between 10 and 20. The high gain stages would either be attenuated on board or use feedback (i would like to try feedback at the grid, no idea how it will sound but i have such an amount of useless tubes that i don't mind getting some bad results). The output stage will be something arround 25W, which will need some 10-15Vrms to clip, so all the tubes would do the job.
How high would you set the maximum allowable gain for these cards so i can try most circuits? (Now i'm just about to finish my first one, based on the 7HG8, very unique sound)
EDIT: 7HG8 triode section! I havn't thought in using the pentode section and i expect it to be extraordinarily dissapointing.
EDIT2: Wouldn't the B+ resistor make the bias harder to get stable? I have many tubes that are rated at 150/200Vdc max, would it be dangerous for these to be powered from higher rails if they fail to heat fast enough for some reason? If not i will definitely attach to it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.