hey Don, mind sharing your thoughts about using your crazy drive instead of screen grid drive?
the Berning amp runs on low output plate current, 3 to 5mA, this is the selling point of this amp, the hard part is in the drive, and the high voltage swing needed at G2 of the output stage,
mosfets in place of the 6sn7 cathode followers will surely slap output tube G2 with lots of drive currents...
It would certainly make the amp a little simpler .
What mosfet would you use ?
Here are smoking-amps original findings : http://www.diyaudio.com/forums/tubes-valves/211254-those-magnificent-television-tubes-47.html#post4574362
Mickystan's implementation : http://www.diyaudio.com/forums/tubes-valves/211254-those-magnificent-television-tubes-67.html#post4675483
Last edited:
So even the MOSFET's are par for the course... what happened to "no SS" and "classic"? Just saying...
You're right jazbo8 , I think I should stick to valve drive , and maybe , once i get that working and have learnt from it , try mosfets .
Haha, valve are better than mosfets as drivers, 6sn7 in cathode follower can drive anything. With mosfets your distortion level will rise subjectively and objectively on some measurements. It has been done, tried and measured.
Maybe the only place to use a mosfet is in a 300B PSE to drive the crazy tube up to a highly distorted 8 watts! Although a regulated EL34 can be a better driver solution for difficult loads.
Maybe the only place to use a mosfet is in a 300B PSE to drive the crazy tube up to a highly distorted 8 watts! Although a regulated EL34 can be a better driver solution for difficult loads.
"hey Don, mind sharing your thoughts about using your crazy drive instead of screen grid drive?"
Since the EA-230 amp is already developing the HV drive swing for g2 drive, one would just need to add the two resistors from g2 to g1 and g1 to cathode. (see Mickeystan's implementation) Crazy Drive requires about 3/4 of the drive swing of g2 only drive, so the total drive current will be even lower. The 6JN6 output tubes would be fine, some even bigger Sweeps are easier to drive (26LX6 more gm2).
A Mosfet follower could save space over the tube follower. Just use a HV Mosfet with low input and reverse capacitance and keep 25V minimum across it (to avoid capacitance changes). There was a thread about selecting the right Mosfet recently. An IXYS IXTP01N100D or IXTP01N100 should work fine (5 pF Crss). (some low current Fairchild ones too, FDQ1N80 maybe. 3 pF Crss) A resistor in the Drain circuit to B+ can limit the grid drive current, for safety of the output tube. Use a Gate stopper resistor too, of course, like 1K.
There are always theoretical complaints about Mosfet capacitance effects. Using a low current, low capacitance, Mosfet, with a minimum drain voltage, in follower mode, seems to get good results. But I do wonder about such schemes as this one below with a 300 Amp Industrial strength Mosfet driven by a puny tube:
(Fork lift truck motor controllers? 2,000 pF Crss, 150,000 pF Ciss, And they get great magazine reviews?)
-
Since the EA-230 amp is already developing the HV drive swing for g2 drive, one would just need to add the two resistors from g2 to g1 and g1 to cathode. (see Mickeystan's implementation) Crazy Drive requires about 3/4 of the drive swing of g2 only drive, so the total drive current will be even lower. The 6JN6 output tubes would be fine, some even bigger Sweeps are easier to drive (26LX6 more gm2).
A Mosfet follower could save space over the tube follower. Just use a HV Mosfet with low input and reverse capacitance and keep 25V minimum across it (to avoid capacitance changes). There was a thread about selecting the right Mosfet recently. An IXYS IXTP01N100D or IXTP01N100 should work fine (5 pF Crss). (some low current Fairchild ones too, FDQ1N80 maybe. 3 pF Crss) A resistor in the Drain circuit to B+ can limit the grid drive current, for safety of the output tube. Use a Gate stopper resistor too, of course, like 1K.
There are always theoretical complaints about Mosfet capacitance effects. Using a low current, low capacitance, Mosfet, with a minimum drain voltage, in follower mode, seems to get good results. But I do wonder about such schemes as this one below with a 300 Amp Industrial strength Mosfet driven by a puny tube:
(Fork lift truck motor controllers? 2,000 pF Crss, 150,000 pF Ciss, And they get great magazine reviews?)
-
Attachments
Last edited:
Did the OP ever say exactly what his speaker requirements are for an amp? I don't remember seeing it. I'm pretty sure we all just assumed he needs high power because of the initial request for PPP scheme. Is that what he needs?
If all the PPP request comes from is trying to avoid puny power levels in the normal puny power level topologies then PPP with a small power tubes is VERY viable. In that case a concertina PI driving parallel small tubes is a great solution. You can avoid the extra stage of amplification which is a known degrader of sound. And you can probably avoid the use of ANY kind of follower stage because the concertina PI "is" a follower and has low impedance on both loads.
Smoking-amp, I think you possibly dismissed that solution way too early because of your favorite (complicated) solutions to things. The guy is just getting started. Give him a break.
If all the PPP request comes from is trying to avoid puny power levels in the normal puny power level topologies then PPP with a small power tubes is VERY viable. In that case a concertina PI driving parallel small tubes is a great solution. You can avoid the extra stage of amplification which is a known degrader of sound. And you can probably avoid the use of ANY kind of follower stage because the concertina PI "is" a follower and has low impedance on both loads.
Smoking-amp, I think you possibly dismissed that solution way too early because of your favorite (complicated) solutions to things. The guy is just getting started. Give him a break.
mosfets are easy pick if transconductance is the criteria...http://www.diyaudio.com/forums/tubes-valves/107155-mosfet-driver-power-pentodes.html#post4749286
@smoking-amp, thanks for the response...
From post #4 by the Op:
"my aim would be to build a valve amplifier that has the ability to drive difficult loads ( ie. 2 or 3-way speakers that are not efficient , say ... rated at 8 ohm with impedance dips down to 3 or 2 ohm ) ; and that has a high damping factor , for a valve amp . "
and post #1 by the Op:
" examples of classic PPP circuits that are proven designs , and that are as simple as the circuits I've mentioned above . Nothing with solid state current regulation..."
Requirements might have drifted since then, dunno. Tough requirements to meet all that in a classic design without getting expensive or complex. I suppose the easiest solution would be to go for a higher power amp than typically needed.
"my aim would be to build a valve amplifier that has the ability to drive difficult loads ( ie. 2 or 3-way speakers that are not efficient , say ... rated at 8 ohm with impedance dips down to 3 or 2 ohm ) ; and that has a high damping factor , for a valve amp . "
and post #1 by the Op:
" examples of classic PPP circuits that are proven designs , and that are as simple as the circuits I've mentioned above . Nothing with solid state current regulation..."
Requirements might have drifted since then, dunno. Tough requirements to meet all that in a classic design without getting expensive or complex. I suppose the easiest solution would be to go for a higher power amp than typically needed.
Last edited:
I guess I'm a little confused about where the advice to Fredeb is going here. Maybe it's because he never defined how much power he actually needs. I guess it was just that it was a difficult load. Since I'm under the impression that this is his first build, but he actually has knowledge, it seems like the best thing to do is go with his initial desires.
That seems to be low output impedance coming either out of the outputs tubes or out of the OPT. It translates into the same thing but it gets way more complicated which you route you take. The bottom line is that you need to amplify the voltage or amplify the current (really a mix of the two). To me, maybe not to others, it makes way more sense to emphasize current. That's because you won't have to amplify the voltage as much. You just have to lower the voltage anyway in the OPT. This is really what all feedback schemes are ultimately about (plus damping factor). They translate lower the higher than needed voltage output.
Doesn't it make a lot more sense to just not increase the voltage amplification so much? Especially if it will require a whole additional amplification stage? Just increase the current from the output tubes and lower their impedance by doubling up on them. Mismatch of output tubes is a big bugaboo to some but I just don't think it's much of real world problem. And if you're worried about it then CCS the cathodes of the output tubes. Problem solved.
I think advice here is losing the big picture, especially for someone doing their first real build and actually wants success. One additional point. If you use a concertina as the low impedance driver then you do not need a negative voltage. Much simpler to add a high voltage off the existing xfmr for the concertina than probably a whole additional xfmr for the negative supply for a mosfet follower. These ideas really are not about the badness of individual ideas offered from folks here. It's more about people here just losing the big picture and how they are dealing with someone on his first serious build.
That seems to be low output impedance coming either out of the outputs tubes or out of the OPT. It translates into the same thing but it gets way more complicated which you route you take. The bottom line is that you need to amplify the voltage or amplify the current (really a mix of the two). To me, maybe not to others, it makes way more sense to emphasize current. That's because you won't have to amplify the voltage as much. You just have to lower the voltage anyway in the OPT. This is really what all feedback schemes are ultimately about (plus damping factor). They translate lower the higher than needed voltage output.
Doesn't it make a lot more sense to just not increase the voltage amplification so much? Especially if it will require a whole additional amplification stage? Just increase the current from the output tubes and lower their impedance by doubling up on them. Mismatch of output tubes is a big bugaboo to some but I just don't think it's much of real world problem. And if you're worried about it then CCS the cathodes of the output tubes. Problem solved.
I think advice here is losing the big picture, especially for someone doing their first real build and actually wants success. One additional point. If you use a concertina as the low impedance driver then you do not need a negative voltage. Much simpler to add a high voltage off the existing xfmr for the concertina than probably a whole additional xfmr for the negative supply for a mosfet follower. These ideas really are not about the badness of individual ideas offered from folks here. It's more about people here just losing the big picture and how they are dealing with someone on his first serious build.
If one uses a Mosfet for the Concertina splitter, no need for the extra HV or for a floating heater winding. One less tube too. Not exactly classic, but it would drive parallel output tubes or some big output tube(s) easily.
Last edited:
makes a lot sense.
My first idea was to raise the feedback from his best amplifier and see if it improves the grip on the 'hard to drive' speakers. Everyone agrees that Fred has some nice capable amps already. More power or more tubes will not translate into a lower output impedance. Modify his existing amps is easy: more feedback, grid bias to increase the power and better match the difficult load, maybe that is all he needs. I don't believe in the ultimate amplifier
There is also this neat and simple trick (R11):
http://www.diyaudio.com/forums/tubes-valves/291848-4699-sep-unusual-cathode-feedback.html
Puts negative resistance in to lower the output impedance. Would have to be adapted for P-P.
http://www.diyaudio.com/forums/tubes-valves/291848-4699-sep-unusual-cathode-feedback.html
Puts negative resistance in to lower the output impedance. Would have to be adapted for P-P.
Actually it will.🙂 If everything else stays equal, including the OPT, paralleling output tubes will lower impedance going to the speaker. So if you don't actually want an average lower impedance going to the speaker then you decrease the feedback coming from the OPT and get more power. Or conversely, lower the voltage gain, which is generally a better solution. Less feedback generally leads to a more open, less congested sound. Especially if its GNFB one is using. So that means the two things, output power and output impedance are intertwined.
Last edited:
yes. i built a ppp amp using 16 KT88's wired in triode mode,
using topology by Bob Carver, although i made provisions for gnfb,
i found that i can do without it...
Attachments
oh, the amp was sent to the states, and i think is somewhere in california...😀
btw, i have a 300 watt tube power amp design using just a pair of output tubes, not very high B+, around 650 volts.....
- Status
- Not open for further replies.