The 6SN7 is one of the most popular tubes in audio. It was also intended for use as vertical deflection amplifier. In fact it is capable of 300 mA cathode peak current!
Sweep tubes are not strange it's just that you don't know them.😉
My sweep tube amp runs the 6CB5A across a 5k a-a opt. B+ is 265V and idle is 90 mA/tube. Its g2 is 105V, taken from an IRF820B source follower hung from a voltage divider. It turned out very well.
The 6CB5 is one of my favourites, heater power over plate dissipation is .63...🙂 There may not be any that are higher. EL84 LTP to do the P-I duty. The 866's may or may not have anything to do with how it sounds...LOL
cheers,
Douglas
The 6CB5 is one of my favourites, heater power over plate dissipation is .63...🙂 There may not be any that are higher. EL84 LTP to do the P-I duty. The 866's may or may not have anything to do with how it sounds...LOL
cheers,
Douglas
Hi
I know the 6sn7 but the we are speaking about power amp. And the the 300mA peak are under certain condition as well described on sheet. The maximun dc current, per plate, is 20 mA.
I see the data of 6CB5 and I will find some to test.
I have also OT to check dynamically.
Walter
I know the 6sn7 but the we are speaking about power amp. And the the 300mA peak are under certain condition as well described on sheet. The maximun dc current, per plate, is 20 mA.
I see the data of 6CB5 and I will find some to test.
I have also OT to check dynamically.
Walter
PRR made some interesting points about tubes typically running much hotter than transistors at idle.
I have read that tubes do revert to a more exponential I versus Vg function at very low currents, so theoretically they could gm double down at 1 mA or below maybe. Gm would likely be doubling between evanescent and insignificant down there however. And then the stage would suffer huge variation of gm over the full operating range. Which would produce a deep V shaped gm curve. Lots of 3rd harmonic dist then.
For a general power law device: I = k Vg ^👎
and solving for dI/dVg to get gm: gm = n k Vg^(n-1) = n k^(1/n) I^([n-1]/n)
grid 1 usually looks near to square law over its central operating region, n = 2, so:
gm = 2 k Vg
or gm proportional to Vg
and gm = 2 x (k^.5) I^.5
or gm proportional to SQRT of I
Which fits PRR's comment about 1/4 max current giving 1/2 gm exactly. Giving a nice rough rule for class AB biasing to keep gm fairly leveled in the negative grid1 region.
Very handy rules to keep in mind.
(empirically, TV Sweeps obey roughly I to the .55 power to I to the .65 power for gm at high (like DC max) and very high currents (2 x DC max)
---------------
On that displaced crossover idea: One could use 4 output tubes (parallel sets) with one set displaced to positive signal crossover, and the other set displaced to negative signal crossover. That way one tube set would always be operating with healthy gm while the other set drops out during its (class B) crossover.
Maybe would want to bias each displaced set to look like class aB over the displaced section, so gm wouldn't really "drop out". Instead, a steady class A region from gm overlap. Then arrange the two tube set displacements, so the edges of the class A ranges would meet each other, exactly, or one really would get gm doubling. This is more trouble than it is worth I think.
..
I have read that tubes do revert to a more exponential I versus Vg function at very low currents, so theoretically they could gm double down at 1 mA or below maybe. Gm would likely be doubling between evanescent and insignificant down there however. And then the stage would suffer huge variation of gm over the full operating range. Which would produce a deep V shaped gm curve. Lots of 3rd harmonic dist then.
For a general power law device: I = k Vg ^👎
and solving for dI/dVg to get gm: gm = n k Vg^(n-1) = n k^(1/n) I^([n-1]/n)
grid 1 usually looks near to square law over its central operating region, n = 2, so:
gm = 2 k Vg
or gm proportional to Vg
and gm = 2 x (k^.5) I^.5
or gm proportional to SQRT of I
Which fits PRR's comment about 1/4 max current giving 1/2 gm exactly. Giving a nice rough rule for class AB biasing to keep gm fairly leveled in the negative grid1 region.
Very handy rules to keep in mind.
(empirically, TV Sweeps obey roughly I to the .55 power to I to the .65 power for gm at high (like DC max) and very high currents (2 x DC max)
---------------
On that displaced crossover idea: One could use 4 output tubes (parallel sets) with one set displaced to positive signal crossover, and the other set displaced to negative signal crossover. That way one tube set would always be operating with healthy gm while the other set drops out during its (class B) crossover.
Maybe would want to bias each displaced set to look like class aB over the displaced section, so gm wouldn't really "drop out". Instead, a steady class A region from gm overlap. Then arrange the two tube set displacements, so the edges of the class A ranges would meet each other, exactly, or one really would get gm doubling. This is more trouble than it is worth I think.
..
Last edited:
45:
"If you compare the output level where the switching is happening respect to rated power most class AB solid state amps would be considered pure class B amps in the tube world!"
Indeed!
---------------
OOPs. Late Edit: post #44
Should have been: gm varies as roughly the .65 power at DC max current (n = 3) and .55 power at 2 x DC max current (n = 2.22) for TV Sweeps. Which means TV Sweeps don't reach square law operation (n = 2.0) until over 2 x DC max current. Which explains why running them really hot often makes them get more linear. George (Tubelab) should be happy!
gm proportional to I^[(n-1)/n]
n = 0 for constant gm (at multi-Amperes plate current!)
..
"If you compare the output level where the switching is happening respect to rated power most class AB solid state amps would be considered pure class B amps in the tube world!"
Indeed!
---------------
OOPs. Late Edit: post #44
Should have been: gm varies as roughly the .65 power at DC max current (n = 3) and .55 power at 2 x DC max current (n = 2.22) for TV Sweeps. Which means TV Sweeps don't reach square law operation (n = 2.0) until over 2 x DC max current. Which explains why running them really hot often makes them get more linear. George (Tubelab) should be happy!
gm proportional to I^[(n-1)/n]
n = 0 for constant gm (at multi-Amperes plate current!)
..
Last edited:
Late edit again:
last post, n=1 for constant gm, the real linear case. (not n = 0)
Well, here are some real power tube gm curves (versus grid voltage, Vg):
10/6JA5, page 5 gm curves
http://tubedata.milbert.com/sheets/123/1/10JA5.pdf
6LU8, page 7 gm curves
http://tubedata.milbert.com/sheets/123/6/6LU8.pdf
EL3010, S curve, bottom of page 573
http://tubedata.milbert.com/sheets/128/e/EL3010.pdf
As you can see, these don't generally develop a straight line section (n = 2) in the middle of the gm curve. Have to go to higher current. You certainly won't get any gm doubling in P-P with these at least.
(hard to find gm curves for many power tubes)
E130L, gm curve versus I or plate current, on page 7, showing the approx. square root function of gm versus current.
http://tubedata.milbert.com/sheets/009/e/E130L.pdf
E55L, gm versus Vg on page 5, AND, gm versus I on page 9 (top)
http://tubedata.milbert.com/sheets/009/e/E55L.pdf
12GN7, gm versus Vg, bottom last page (really high gm frame grid types more likely to run past the cathode emission capabilty at higher current, so gm curve bends over up top)
http://tubedata.milbert.com/sheets/135/1/12GN7A.pdf
..
last post, n=1 for constant gm, the real linear case. (not n = 0)
Well, here are some real power tube gm curves (versus grid voltage, Vg):
10/6JA5, page 5 gm curves
http://tubedata.milbert.com/sheets/123/1/10JA5.pdf
6LU8, page 7 gm curves
http://tubedata.milbert.com/sheets/123/6/6LU8.pdf
EL3010, S curve, bottom of page 573
http://tubedata.milbert.com/sheets/128/e/EL3010.pdf
As you can see, these don't generally develop a straight line section (n = 2) in the middle of the gm curve. Have to go to higher current. You certainly won't get any gm doubling in P-P with these at least.
(hard to find gm curves for many power tubes)
E130L, gm curve versus I or plate current, on page 7, showing the approx. square root function of gm versus current.
http://tubedata.milbert.com/sheets/009/e/E130L.pdf
E55L, gm versus Vg on page 5, AND, gm versus I on page 9 (top)
http://tubedata.milbert.com/sheets/009/e/E55L.pdf
12GN7, gm versus Vg, bottom last page (really high gm frame grid types more likely to run past the cathode emission capabilty at higher current, so gm curve bends over up top)
http://tubedata.milbert.com/sheets/135/1/12GN7A.pdf
..
Last edited:
- Status
- Not open for further replies.