Just thinking at the moment - but I use a tube width between smaller tubes at the moment and orientate the plates so they radiate in directions away from the other tubes. Would the same work for the larger tubes? Given the inverse square law and power dissipation I'd expect more but how much more?
The reason I ask is I have some offcuts of riven blue slate ~300mm wide, ~300mm deep and 30mm thick from a different project. It's possible to mill tube holes out and secure the socket to a metal under chassis.
The design I'm thinking of is basically a sandwich of two offcuts of slate and a 20-30mm sheet of aluminium (CNC). Slate has no issue with heat and the aluminium acts to both provide a structural mount but also to acts as an integral heatsink for the chassis components. The power supply I would split off from the 'stereo' amp within a separate chassis. I'm be tempted by a bigger bottle design after this one and I know nothing about big bottles.
The reason I ask is I have some offcuts of riven blue slate ~300mm wide, ~300mm deep and 30mm thick from a different project. It's possible to mill tube holes out and secure the socket to a metal under chassis.
The design I'm thinking of is basically a sandwich of two offcuts of slate and a 20-30mm sheet of aluminium (CNC). Slate has no issue with heat and the aluminium acts to both provide a structural mount but also to acts as an integral heatsink for the chassis components. The power supply I would split off from the 'stereo' amp within a separate chassis. I'm be tempted by a bigger bottle design after this one and I know nothing about big bottles.
I mount my power tubes 2 1/2 inches centre to centre.
The GEC sheet for KT88 recommends 4 inches though, and shows how the sockets should be mounted.
https://frank.pocnet.net/sheets/084/k/KT88_GEC.pdf
The GEC sheet for KT88 recommends 4 inches though, and shows how the sockets should be mounted.
https://frank.pocnet.net/sheets/084/k/KT88_GEC.pdf
Sounds more like the issue will be the thick plate the tubes stick through, and not so much how close the tubes are to each other...? 30mm slate + 30mm alu and you have 60mm holes for the tubes. This will probably create thermal heat pipe where air is sucked thru, if you make the holes right. The tube-tube distance will likely not matter so much. Or am I picturing this wrong?
It really depends on the heat emitted by each tube. KT88s dissipate rather large amounts. 2A3 much less.
I've seen 'rules of thumb' one tube diameter between tubes. Which is about what you say you already do. Not sure you need to orient the plates.
I've placed 6AS7s so they are about 20mm glass to glass with no problems, but I never felt too comfortable with them that close and have ca one tube dia most of the time. 6AS7s they are in the same power class as 2A3s. (Actually a bit higher with 2*12W vs 15W, but close enough).
It really depends on the heat emitted by each tube. KT88s dissipate rather large amounts. 2A3 much less.
I've seen 'rules of thumb' one tube diameter between tubes. Which is about what you say you already do. Not sure you need to orient the plates.
I've placed 6AS7s so they are about 20mm glass to glass with no problems, but I never felt too comfortable with them that close and have ca one tube dia most of the time. 6AS7s they are in the same power class as 2A3s. (Actually a bit higher with 2*12W vs 15W, but close enough).
The base and thickness would need working out but it looks like it’s possible to get a low power tubes close enough to fit.
Thank you.
Thank you.
My thinking here is that the socket would be mounted on the top side of the all chassis, the CNC then makes slots in the chassis for components/wires and the like. The top slate then mounts flat on the alu with the tubes only needing ~30mm above the slots. The holes milled wider to provide a air gap/convection channel. The nice thing is that the slate left the mill calibrated to 30mm thickness.
The lower stage could be milled slightly to allow some depth below the alu plane.
Food for thought.
Attachments
Last edited:
I have learned from the older tube builders that I should have 1½ tube diameter between them so that they do not heat each other, which otherwise shortens the service life. I have this as a guideline when I build, a tip is to look at Japanese tube amplifiers often they are built that way.
You might want to try using the JJ 2A3-40 someday.
The JJ 2A3-40 diameter is 62mm.
A regular 2A3 tube diameter is only 52.4mm.
That is a 4.8mm difference in the radius (center to outside).
Or, with those 2 tube types, and the same location of the socket centers, the glass of tube to tube is 9.6mm closer.
The JJ 2A3-40 diameter is 62mm.
A regular 2A3 tube diameter is only 52.4mm.
That is a 4.8mm difference in the radius (center to outside).
Or, with those 2 tube types, and the same location of the socket centers, the glass of tube to tube is 9.6mm closer.
Last edited:
Be Adventuresome!
Join the Navy, or . . .
Try different tubes.
By the way, a 2.5V DHT filament versus a 5V DHT filament gives much different results,
if you are using AC power for the filaments.
The 2X Frequency of the Power Mains Frequency causes Intermodulation on each music tone.
That Intermodulation is 6dB less on the 2.5V DHT, versus the larger IM on the 5V DHT.
Even if you are using DC filaments on those DHT tubes, the next problem is the uneven spread of the current across the plate.
The spread of current is less on the 2.5V DHT than it is on the 5V DHT.
Other than the above, I will let others chime in and discuss the other differences of the tubes.
Correct. They are different tubes.
Your Mileage Will Vary.
Join the Navy, or . . .
Try different tubes.
By the way, a 2.5V DHT filament versus a 5V DHT filament gives much different results,
if you are using AC power for the filaments.
The 2X Frequency of the Power Mains Frequency causes Intermodulation on each music tone.
That Intermodulation is 6dB less on the 2.5V DHT, versus the larger IM on the 5V DHT.
Even if you are using DC filaments on those DHT tubes, the next problem is the uneven spread of the current across the plate.
The spread of current is less on the 2.5V DHT than it is on the 5V DHT.
Other than the above, I will let others chime in and discuss the other differences of the tubes.
Correct. They are different tubes.
Your Mileage Will Vary.
Be adventurous. Triode strap sweep tubes and enjoy the music. 12 volts for heaters is easy and cheap. HV with Delon from isolation xmfr is also easy and cheap. Sound is good enough that I don't want to experiment further. MA-1 Modular amplifier.
I've added 6HD5 and *GT5 to the list.
😀
I've added 6HD5 and *GT5 to the list.
😀
Last edited:
On the subject of DHT I did find this interesting piece about hum reduction: Effects of AC Heating Power Applied to Directly Heated Triodes
Injecting PS hum into the grid and using it to cancel more hum than a hum bucked pot alone.
Injecting PS hum into the grid and using it to cancel more hum than a hum bucked pot alone.
There are no free lunch.
Injecting anything into the signal, which is originally not there, generates intermodulation components.
To fix a bug so that another is generated, inadvisable (this is the first thing an engineer is taught).
To eliminate the cause, it is the responsibility of the designer.
Injecting anything into the signal, which is originally not there, generates intermodulation components.
To fix a bug so that another is generated, inadvisable (this is the first thing an engineer is taught).
To eliminate the cause, it is the responsibility of the designer.