I haven't bought any in a while but I do recall that there are always some listings on eBay that are outrageously priced. Sellers who're looking for suckers and just keep re-listing the same tubes, literally for years.
If you're patient and watchful, you can find them at more reasonable prices. I don't think I ever paid more than ~$35 but I think prices have gone up a bit.
The big dealer near me Vacuum Tubes prices. Best prices every day! lists them for $30 but they may not have any in stock.
I'll have to check out the 6AV5GA, which they have listed for $4.
These tubes (6336) don't want positive Volts on the grid. Obeying the 400V max (200V B+) and staying within the 400 mA per plate. With an 825 Ohm SE OT, using both triode sections in parallel, should get you around 13 Watts output. The 6336, Mu 2.7, takes a -LOT- of drive voltage (like 300V p to p).
At 135 mA idle (per section) (200V B+), the Rp for the combined sections is about 150 Ohms. So 825/150 = 5.5, a fairly typical ratio for lower dist. One could go for a higher primary Z to get the distortion down more. And I suspect the 400V peak V could be pushed a little, since this is not continuous regulator duty. A 6528 tube could get the drive V cut in half.
If the 2nd Harmonic distortion is not low enough, then one can go for a higher primary Z, but fewer Watts (unless the B+ is pushed up accordingly). Or one could go for an UnSet drive setup (use 6528, higher Mu factor available), which would lower distortion and increase power output. Or simplest, just put some very light NFB back to the driver stage cathode.
P-P class A with 2 tubes should give about 2X the power. (1650 Ohm P-P OT, 200V B+) (Edcor CXPP25-1.6K) With very reduced 2nd Harmonic dist.
P-P class AB1 should be able to reach as much as 4X the power output. (maybe 1200 Ohm P-P OT)
Last edited:
I should mention that I actually HAVE a Hammond 1640SE 1250 Ohm (200 mA max) 25 Watt OT. Not the best fit for current with the 2x 6336/6528 triode sections, but I plan to push the B+ up some to compensate for the lower current with a 1250 Ohm SE load.
A 6336 section is at about 360 Ohms at 100 mA (250V B+), or a 6528 section is at about 440 Ohm at 100 mA. So a 6.94x Ohms factor for the 6336 and a 5.7x Ohms factor for the 6528 using the 1250 Ohm Hammond SE OT. Should keep dist. under control, but a small amount of NFB could be called for.
A 6336 section is at about 360 Ohms at 100 mA (250V B+), or a 6528 section is at about 440 Ohm at 100 mA. So a 6.94x Ohms factor for the 6336 and a 5.7x Ohms factor for the 6528 using the 1250 Ohm Hammond SE OT. Should keep dist. under control, but a small amount of NFB could be called for.
Last edited:
Emission Labs makes a tube close to what you are searching for, but you better have deep pockets. The EML 520B-4 has a 55W plate dissipation rating and is available with a center tapped filament that is used as a cathode. It takes an octal socket and does not follow the 300B plate curves precisely as it has a plate resistance of 480r vs. 700r for the 300B.
Thank you.
250V is conservative. Very. No arcing happens with 700V on a cold plate. ~350V on the warm plate of first position, 140V on the cathode in second position (which exceeds the 100V spec for Vhk, but in years of operation this has never been a problem). Same as how I treat 6N8S as 6SN7GTB for most purposes (voltages, not power).
250V is conservative. Very. No arcing happens with 700V on a cold plate. ~350V on the warm plate of first position, 140V on the cathode in second position (which exceeds the 100V spec for Vhk, but in years of operation this has never been a problem). Same as how I treat 6N8S as 6SN7GTB for most purposes (voltages, not power).
(6336/6528)
This should get your attention even more:
7241/7242
7241 has a 67 Ohm Rp, both tubes using 7.5 Amp heater. $$$
but they are just 3 section versions of the 2 section 6336 and 6528. Might as well just parallel two of the 6336/6528 types, better performance for way less $$ (well except 5A + 5A heater then, but 50 Ohm output!! OTL).
Attachments
Last edited:
6336 push pull
Recent build 6336 in push-pull Class A, 8K plate-to-plate, no negative feedback delivered 10W at <1%THD, 19W at 5%THD.
De Haan and Heuvleman achieved even better measurements (Glass Audio, May 2005).
Recent build 6336 in push-pull Class A, 8K plate-to-plate, no negative feedback delivered 10W at <1%THD, 19W at 5%THD.
De Haan and Heuvleman achieved even better measurements (Glass Audio, May 2005).
http://educypedia.karadimov.info/library/ax-glass_audio_special05.pdf
Wow, they used 450V B+ with an 8K P-P OT, in class A, to get 10 Watts out with 0.1% distortion.
Pretty good distortion figure, but lowish Watts (for two 60 Watt 6336 tubes) with the high Z primary OT.
I think I would use a lower Zpri OT and lower B+, with some local NFdbk to the driver stage, to get the low dist. with 25 Watts output. (class A)
With 10 Amps of heater power, one should make use of the high current capability of the tubes, and leave the distortion fixing to the driver.
Some experimenting called for, no doubt.
One could also use a 12HL7 differential driver stage to control the outputs so they emulate the excellent internal triode in (selected) 12HL7s using 12HL7 screen grid feedbacks (12HL7 triode curves below).
Attachments
Last edited:
As far as I can tell, all big triode amps are room heaters with the potential to make nice music!! Maybe we should be driving 300Bs in Class A2 to get more bang for the power bill buck (I don't think 6336s would be too happy in A2 - let me know if I'm off course here!!)
Anyway, I tried H&deH's approach with 6336s - I found the distortion figures achievable but the circuit and the power supply got so big that I found it difficult to tame.
Casting around, I was intrigued by Lynn Olson's Amity and I reworked the 6336 amp to a simpler 2-stage topology with a less demanding power supply ...
Both topologies sound BIG - and here I am comparing the 6336 amps to very good PP-KT88. However, my ears told me that my build of the H&deH design could sound edgy with my speakers, where the two stage topology did not and was more pleasant to listen to - so far so good ...
On driver valves, I tried 8077s and 12HL7s, of which I am fond. However, I had better results from big-bottle RCA 12HG7s that deliver gain + mo' current - probably better suited to driving the 6336's significant input capacitance.
Anyway, I tried H&deH's approach with 6336s - I found the distortion figures achievable but the circuit and the power supply got so big that I found it difficult to tame.
Casting around, I was intrigued by Lynn Olson's Amity and I reworked the 6336 amp to a simpler 2-stage topology with a less demanding power supply ...
Both topologies sound BIG - and here I am comparing the 6336 amps to very good PP-KT88. However, my ears told me that my build of the H&deH design could sound edgy with my speakers, where the two stage topology did not and was more pleasant to listen to - so far so good ...
On driver valves, I tried 8077s and 12HL7s, of which I am fond. However, I had better results from big-bottle RCA 12HG7s that deliver gain + mo' current - probably better suited to driving the 6336's significant input capacitance.
Attachments
Ohh!
I just realized that the De Haan and Heuvleman P-P class A 6336 amplifier is only using -HALF- a 6336 tube for each P-P side. That would explain the low 10 Watts output. Usually, using higher B+ increases efficiency.
That would explain the 8K OT then also, partly. They are still using a higher than expected primary Z and B+ to get the low distortion figure. With the high 2nd harmonic dist. inherent in the 6336 plate curves (and most all V reg. types), one has to resort to high Z loading to get that down. (without using some local N Fdbk) But peak plate voltages will be reaching at least 900V that way. The datasheet plate curves only go out to 400V, with a 400V absolute maximum spec.
If that 400V absolute maximum spec is to be believed, then a lower Z OT with lower B+ is in order, with some local NFdbk to fix the 2nd Harmonic distortion. On the other hand, that V spec is for V regulator usage, and may not apply to amplifier usage.
I just realized that the De Haan and Heuvleman P-P class A 6336 amplifier is only using -HALF- a 6336 tube for each P-P side. That would explain the low 10 Watts output. Usually, using higher B+ increases efficiency.
That would explain the 8K OT then also, partly. They are still using a higher than expected primary Z and B+ to get the low distortion figure. With the high 2nd harmonic dist. inherent in the 6336 plate curves (and most all V reg. types), one has to resort to high Z loading to get that down. (without using some local N Fdbk) But peak plate voltages will be reaching at least 900V that way. The datasheet plate curves only go out to 400V, with a 400V absolute maximum spec.
If that 400V absolute maximum spec is to be believed, then a lower Z OT with lower B+ is in order, with some local NFdbk to fix the 2nd Harmonic distortion. On the other hand, that V spec is for V regulator usage, and may not apply to amplifier usage.
Last edited:
Looking at some V regulator tube datasheets (6AS7, 6080, 7233, 7236, 6528, 6336, 7241, 7242) all except 7233 and 7236 have plate curves that end abruptly at their Vmax spec. 7233 and 7236 go up another 150V above that.
Looking at the plate current curves plotted, all of them go up to 1x the max current rating, except 7233 and 7236, which go up to 3x the max current for 7233 and 2x the max current for 7236.
The 7236 datasheet states that it is for amplifier use. But it still has exactly the same high 2nd Harmonic laden curves as the V regulator tubes.
When you take into account the max Watt ratings on a load line (in class A), it becomes clear that exceeding just -HALF- the Vmax rating by say 50V or so for the V reg. tubes, you will be leaving behind the high current capability in order to stay within the max Watt rating. So the high heater current and Htr. Watts begins to be wasted.
There is some longevity advantage in de-rating below the max current rating for sure, but going below 50% is just wasting Htr power.
So my take on these V reg. tubes is to use a Zprimary OT that gives expected typical class A output power (rather than prioritizing low dist. at lower Watts or exceeding the Vmax rating) and then use some minimal local NFdbk to get the dist. down to an acceptable level (below hearing).
Or use the triode emulation scheme (with driver NFdbk enforcing) to get high quality triode curves from the outputs. When you have a quality triode emulation
-model-, the NFdbk has relatively little work to do to get a real output triode to conform to it.
Finally, it seems that these high current V reg. type tubes (except 7233 and 7236) have essentially the same Imax DC and Imax peak ratings. So while 300 or 400 mA DC rated (per section) may seem impressive, the TV Sweeps have similar DC ratings at the 30 to 40 Watt rated level too, but they have 3.5x times that for the -peak- current rating. (of course pentode mode with grid 2 helps assist that greatly)
Once one achieves "Nirvana" using triode emulation NFdbk, there seems little reason to use these large expensive triodes, including 7241, 7242 ... and 300B for that matter. Just use a big TV Sweep tube(s) for the output tube, with a driver stage with a high quality internal "model" triode. (driver UL NFbk, make sure to get the phasing correct with crossed N Fdbks)
Using an UnSet drive scheme may still make good sense with a 6528 V reg. tube (mu 9).
I have some 6080, 6528, 6336 and 7233 V reg. tubes to try out, so I'll see what is making the best sense for an amplifier set-up with these.
Looking at the plate current curves plotted, all of them go up to 1x the max current rating, except 7233 and 7236, which go up to 3x the max current for 7233 and 2x the max current for 7236.
The 7236 datasheet states that it is for amplifier use. But it still has exactly the same high 2nd Harmonic laden curves as the V regulator tubes.
When you take into account the max Watt ratings on a load line (in class A), it becomes clear that exceeding just -HALF- the Vmax rating by say 50V or so for the V reg. tubes, you will be leaving behind the high current capability in order to stay within the max Watt rating. So the high heater current and Htr. Watts begins to be wasted.
There is some longevity advantage in de-rating below the max current rating for sure, but going below 50% is just wasting Htr power.
So my take on these V reg. tubes is to use a Zprimary OT that gives expected typical class A output power (rather than prioritizing low dist. at lower Watts or exceeding the Vmax rating) and then use some minimal local NFdbk to get the dist. down to an acceptable level (below hearing).
Or use the triode emulation scheme (with driver NFdbk enforcing) to get high quality triode curves from the outputs. When you have a quality triode emulation
-model-, the NFdbk has relatively little work to do to get a real output triode to conform to it.
Finally, it seems that these high current V reg. type tubes (except 7233 and 7236) have essentially the same Imax DC and Imax peak ratings. So while 300 or 400 mA DC rated (per section) may seem impressive, the TV Sweeps have similar DC ratings at the 30 to 40 Watt rated level too, but they have 3.5x times that for the -peak- current rating. (of course pentode mode with grid 2 helps assist that greatly)
Once one achieves "Nirvana" using triode emulation NFdbk, there seems little reason to use these large expensive triodes, including 7241, 7242 ... and 300B for that matter. Just use a big TV Sweep tube(s) for the output tube, with a driver stage with a high quality internal "model" triode. (driver UL NFbk, make sure to get the phasing correct with crossed N Fdbks)
Using an UnSet drive scheme may still make good sense with a 6528 V reg. tube (mu 9).
I have some 6080, 6528, 6336 and 7233 V reg. tubes to try out, so I'll see what is making the best sense for an amplifier set-up with these.
Last edited: