Re: NOT A CHILD
It's a "current" project, but I really doubt it will be finished any time soon... I'm going to use KT33s for output valves since I already have them.
The 6BX6 is a B9A pentode, so it won't sub easily for 6AU6s - I haven't checked the specs for EF86s recently, but I think they'd be closer to them than 6AU6s.
I think the European designation is EF80. A little OT, but on the subject of 'undiscovered' valves, I recently stumbled on the specs for the 9002 (a miniature substitute for a 1930s acorn valve I think). The curves are possibly linear enough to give the 6SN7 family a run for their money!
jeff mai said:
It's a "current" project, but I really doubt it will be finished any time soon... I'm going to use KT33s for output valves since I already have them.
The 6BX6 is a B9A pentode, so it won't sub easily for 6AU6s - I haven't checked the specs for EF86s recently, but I think they'd be closer to them than 6AU6s.
I think the European designation is EF80. A little OT, but on the subject of 'undiscovered' valves, I recently stumbled on the specs for the 9002 (a miniature substitute for a 1930s acorn valve I think). The curves are possibly linear enough to give the 6SN7 family a run for their money!
I asked Mr Kitic's advise on modifying the RH807 (see My looking for KT88 Circuits thread), he was rather dismissive and far from helpful. His attitude seems to be "build it as is" or don't build at all. Its a shame his input has become so dismissive as he obviously has great talent and good ideas. He seems to be advocating build it by numbers approach, rather than any real experimentation.
Shoog
Shoog
I've been thinking about the problem of too little transconductance from the driver stage at the low current operating point used when the current path is through the output stage feedback resistors. I have a couple of suggestions:
1) The 47s in Gary's amp do not appear to run at max voltage. (It looks like they are around 200V between plate and cathode.) A higher B+ would allow more current to run through Rfb without lowering the driver stage B+. Pushing this to the max plate voltage of 250V only gets you an extra 0.5ma of driver stage current, but you get significantly more transconductance in return.
2) Reduce the value of Rfb to allow more current through the driver stage at the same operating voltages. Of course, this will require more current swing from the driver stage. Again, the benefit is increased transconductance from the driver stage because of its better operating point.
The TubeCad article says "If too low, Rfb becomes a heavy in itself" but then never clarifies what is meant by this. It's obvious if it's too small, the current requirements increase (both the idle current and the current swing required for full output) and maybe this is what is meant. A lower value doesn't seem to adversely affect the Zin and Zout calculations.
A combination of the two suggestions should improve the input sensitivity - whether this adversely affects other things is another issue, but I can't see why it would. Perhaps I'm overlooking something?
As an example, reducing Rfb to 100Kohms and increasing the 47's Vp-k to the max of 250V would allow the driver stage to run at 2.6ma plate current (per tube). With the 6AU6 @ 100Vscrn, this increases its Gm from 1.6ma/V to over 2.5ma/V. This increase more than offsets the reduction in I-V conversion ratio (from 130V/ma to 100V/ma) of the output stage due to the lower value of Rfb. It isn't a huge gain overall (maybe 20%) but it might make marginal into adequate.
1) The 47s in Gary's amp do not appear to run at max voltage. (It looks like they are around 200V between plate and cathode.) A higher B+ would allow more current to run through Rfb without lowering the driver stage B+. Pushing this to the max plate voltage of 250V only gets you an extra 0.5ma of driver stage current, but you get significantly more transconductance in return.
2) Reduce the value of Rfb to allow more current through the driver stage at the same operating voltages. Of course, this will require more current swing from the driver stage. Again, the benefit is increased transconductance from the driver stage because of its better operating point.
The TubeCad article says "If too low, Rfb becomes a heavy in itself" but then never clarifies what is meant by this. It's obvious if it's too small, the current requirements increase (both the idle current and the current swing required for full output) and maybe this is what is meant. A lower value doesn't seem to adversely affect the Zin and Zout calculations.
A combination of the two suggestions should improve the input sensitivity - whether this adversely affects other things is another issue, but I can't see why it would. Perhaps I'm overlooking something?
As an example, reducing Rfb to 100Kohms and increasing the 47's Vp-k to the max of 250V would allow the driver stage to run at 2.6ma plate current (per tube). With the 6AU6 @ 100Vscrn, this increases its Gm from 1.6ma/V to over 2.5ma/V. This increase more than offsets the reduction in I-V conversion ratio (from 130V/ma to 100V/ma) of the output stage due to the lower value of Rfb. It isn't a huge gain overall (maybe 20%) but it might make marginal into adequate.
NO PASARAN....
Hi,
Duh????
Surely everyone knows that the only part of a P version that's actually different from the E version is the heater?
If that heater is fed the way it should be then why would the tube sound any different?
True, among European made tubes there are some confusing references where the unwary may inadvertedly mistake a PL84 for the P version of the EL84 for instance.... Which it isn't.
But that's where the story ends,really.
Does a PL519 sound any different than an EL519 from the same factory and batch? Of course not.
Sorry but maybe you should get used to it.
People who are into do it yourself should be applauded for it, not looked at with disdain.
Obviously there are an awful lot of people unemployed nowadays.
Last time I checked it still wasn't a crime even though some politicians would like to change that if they could....
The social situation being what it is, people building something for themselves to enjoy are trying to find the cheapest/ best relationship. One they can afford at least.
Much better than hanging out in the streets. Some folk around here are trying to actually learn something too...
Always a good thing in my book.
Anyway.....
Cheers,
Hi,
Duh????
Surely everyone knows that the only part of a P version that's actually different from the E version is the heater?
If that heater is fed the way it should be then why would the tube sound any different?
True, among European made tubes there are some confusing references where the unwary may inadvertedly mistake a PL84 for the P version of the EL84 for instance.... Which it isn't.
But that's where the story ends,really.
Does a PL519 sound any different than an EL519 from the same factory and batch? Of course not.
Sorry but maybe you should get used to it.
People who are into do it yourself should be applauded for it, not looked at with disdain.
Obviously there are an awful lot of people unemployed nowadays.
Last time I checked it still wasn't a crime even though some politicians would like to change that if they could....
The social situation being what it is, people building something for themselves to enjoy are trying to find the cheapest/ best relationship. One they can afford at least.
Much better than hanging out in the streets. Some folk around here are trying to actually learn something too...
Always a good thing in my book.
Anyway.....
Cheers,
One further question - I've searched through RDH4 and found nothing:
Does the output impedance of a pentode stage increase when a high impedance is placed in the cathode circuit as it does when the same is done to a triode? This is ignoring the load impedance as it doesn't affect things the way this circuit is set up.
If so, with the CCS in the tail of the driver stage, maybe high Rp of the pentode itself is not necessary. Selection could be made based on the Gm, rather than Gm and high Rp.
I searched a bit further and the 6AC7 still seems the best candidate if high Rp is part of the selection criteria. I discovered the 6AH6 is a miniature equivalent, but the Rp is lower - about 500K instead of 1M.
Does the output impedance of a pentode stage increase when a high impedance is placed in the cathode circuit as it does when the same is done to a triode? This is ignoring the load impedance as it doesn't affect things the way this circuit is set up.
If so, with the CCS in the tail of the driver stage, maybe high Rp of the pentode itself is not necessary. Selection could be made based on the Gm, rather than Gm and high Rp.
I searched a bit further and the 6AC7 still seems the best candidate if high Rp is part of the selection criteria. I discovered the 6AH6 is a miniature equivalent, but the Rp is lower - about 500K instead of 1M.
jeff mai said:
Yes. You can calculate the m of a pentode (though it is rarely specified nowadays, it is in some old databooks) with the same formula m = gm x ra. Since ra is very large, the m of a pentode is also very large (easily 1000 or so) - a small amount of resistance in the cathode circuit can make a very large difference in the effective anode resistance, which is why pentodes can make stunningly good current sinks.
jeff mai said:
But the problem here is that pentodes are mounted in differential. As with triodes (with the caveat of class A loading) the output impedance is ra||Rl. You can easlily raise the output impedance by placing a small value of resistance at the cathodes of each input pentode as long as it leaves enough voltage for your CCS to be happy.
A related thread has been spawned here...
Used Tube Amps, second life..
http://www.diyaudio.com/forums/showthread.php?s=&threadid=57764
dave
Used Tube Amps, second life..
http://www.diyaudio.com/forums/showthread.php?s=&threadid=57764
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.