I recently scored a few pieces of 6S4S with the intention of making a P-P amp, but recently some posts about the 6S41S got my attention. This is still a reasonably priced tube, with higher plate dissipation than the 6S4S. Not as high a plate capability as the 300B, but I can buy a whole boatload of 6S41S for the price of even the cheapest 300B. Anyway, I've seen some SE schematics for the 6S41S, but no P-P. Any leads? I'll probably end up doing my own thing, but it's still nice to see what others have done.
Suitable iron is another question. I have some Fisher 500B output transformers, but I kinda doubt that the impedance is all that suitable for a tube with this low of an rp. I was thinking of running lower voltage, higher current, with an 8 ohm load on the 16 ohm tap of the transformer. Is this an an optimal usage of the 500C transformer (I have no idea how the secondaries are wound). I'm also not comfortable with the amount of negative bias is takes to make the 6S41S go "whoa, horsey!" at the higher plate voltages, another reason for a somewhat lower plate supply (though with a custom SMPS, anything is possible).
On another note, I have heard that the tube to tube variability can be a problem. Comments?
Suitable iron is another question. I have some Fisher 500B output transformers, but I kinda doubt that the impedance is all that suitable for a tube with this low of an rp. I was thinking of running lower voltage, higher current, with an 8 ohm load on the 16 ohm tap of the transformer. Is this an an optimal usage of the 500C transformer (I have no idea how the secondaries are wound). I'm also not comfortable with the amount of negative bias is takes to make the 6S41S go "whoa, horsey!" at the higher plate voltages, another reason for a somewhat lower plate supply (though with a custom SMPS, anything is possible).
On another note, I have heard that the tube to tube variability can be a problem. Comments?
I can see the reason why tubes that are not good for SE may be good for PP, but...
Speaking of the tube to tube variability, it is true for all Russian tubes, cars, trucks, etc...
http://datagor.ru/amplifiers/tubes/...rr-na-6s41s-s-pitaniem-ot-impulsnogo-bp..html
Speaking of the tube to tube variability, it is true for all Russian tubes, cars, trucks, etc...
http://datagor.ru/amplifiers/tubes/...rr-na-6s41s-s-pitaniem-ot-impulsnogo-bp..html
Last edited:
I did not say that. I meant an opposite thing: if the tube is good for SE, it is a GOOD TUBE. So-so tube like 6L6 clones may be bad for SE, but it may be still good for PP. The whole your question sounds weird to me.
However, the driver for this tube needs corresponding, but you know better than anybody else how to use MOSFETs.
Not really a weird question, as two people thought the tube was good in SE, and one thought it was awful (or at least, not so good). In the Russian page you referenced (Google gave me a somewhat fanciful translation - I smiled), the author used the tubes with a rather severe transformer impedance (1.7k if I remember correctly). From my calculations so far, I think something in the 3.3-3.5k range will deliver useful power. This also allows me to experiment using some of the iron I have on hand, at least in the preliminary stages of the project. I've decided that I've bought my last output transformer for a while, as I just received another pair of Fisher 500C transformers this week, as well as a pair of transformers that came from a Knight 6BQ5-based amp. I now have quite enough stuff to play with for the time being.
I had an inspiration this afternoon, and I'm going to try a really heretical approach - a switching cathode bias supply. Finally, an audio app where I can use one of my employer's SMPS chips. I won't spill the beans about this circuit until I have it breadboarded, but it will be interesting, and will no doubt elicit howls of protest (or at least a huh?) from the anti-sand purists. It may also prove useful for people biasing low-mu triodes that need both high voltage at the cathode, plus a high quiescent bias current. I really hate blowing 8-9 watts in a resistor (actually twice that if you use a common bias source for both output tubes).
Given that I am using cathode bias, I may back off from using all-sand followers to drive the output tube grids, but instead use a pair of triode/pentode duals in a differential amp/splitter and cathode follower topology. A pentode differential stage followed by cathode followers should give me enough drive to tickle the output tubes.
I had an inspiration this afternoon, and I'm going to try a really heretical approach - a switching cathode bias supply. Finally, an audio app where I can use one of my employer's SMPS chips. I won't spill the beans about this circuit until I have it breadboarded, but it will be interesting, and will no doubt elicit howls of protest (or at least a huh?) from the anti-sand purists. It may also prove useful for people biasing low-mu triodes that need both high voltage at the cathode, plus a high quiescent bias current. I really hate blowing 8-9 watts in a resistor (actually twice that if you use a common bias source for both output tubes).
Given that I am using cathode bias, I may back off from using all-sand followers to drive the output tube grids, but instead use a pair of triode/pentode duals in a differential amp/splitter and cathode follower topology. A pentode differential stage followed by cathode followers should give me enough drive to tickle the output tubes.
Not really - I know where to stop. Wait 'till I actually do the breadboard and post the circuit. It's not all that radical in complexity, but does some fairly radical things in reducing power dissipation for big triodes with cathode bias (Now, if I could get rid of that pesky filament, we'd really have something).
Actually, I have an idea about that as well, but it's much harder to bring to reality. If it worked, it would drastically increase the efficiency of your average tube, but at the same time, demote the tube to the same status of just another part that gets loaded into a PCB.
I don't know about you, but the romance factor of tubes (glass envelopes, glowing filaments, maybe a touch of blue fluorescence for dessert) is one of the things that make them so fun to play with. I'd like to extend their useful lifetime into the next century, but if they look the same as any other lah-di-lah part that gets loaded into a board, where's the fun?
Actually, I have an idea about that as well, but it's much harder to bring to reality. If it worked, it would drastically increase the efficiency of your average tube, but at the same time, demote the tube to the same status of just another part that gets loaded into a PCB.
I don't know about you, but the romance factor of tubes (glass envelopes, glowing filaments, maybe a touch of blue fluorescence for dessert) is one of the things that make them so fun to play with. I'd like to extend their useful lifetime into the next century, but if they look the same as any other lah-di-lah part that gets loaded into a board, where's the fun?
"I'm going to try a really heretical approach - a switching cathode bias supply"
Hmmm,
HF inductor in the cathode path, pulsed grid pos. bias + input signal, continuous analog tube conduction. Or maybe HF inductor in the plate, pulsed grid pos. bias + input signal, discontinous tube conduction but continous plate inductor current. Analog pulse height switchmode instead of PWM? Seems that would have limited dynamic range without pulse width control too. Norman Crowhurst's last article in Glass Audio was for a circuit that combined analog to pulse height and width modulation. He never got to finished part 2 unfortunately. A revolution shorted out.
There are a couple of technologies that could put a tube into a ceramic chip package. One is the nano-rod cold field emission cathode, the other is the titanium thin film tunneling emitter. Graphene grid maybe. Or laser perforated thin film grid. Very good current densities, no filament. Low voltage operation. Hard to get funds though with Mosfets around.
Hmmm,
HF inductor in the cathode path, pulsed grid pos. bias + input signal, continuous analog tube conduction. Or maybe HF inductor in the plate, pulsed grid pos. bias + input signal, discontinous tube conduction but continous plate inductor current. Analog pulse height switchmode instead of PWM? Seems that would have limited dynamic range without pulse width control too. Norman Crowhurst's last article in Glass Audio was for a circuit that combined analog to pulse height and width modulation. He never got to finished part 2 unfortunately. A revolution shorted out.
There are a couple of technologies that could put a tube into a ceramic chip package. One is the nano-rod cold field emission cathode, the other is the titanium thin film tunneling emitter. Graphene grid maybe. Or laser perforated thin film grid. Very good current densities, no filament. Low voltage operation. Hard to get funds though with Mosfets around.
Last edited:
Ive played around with this tube....
A good while ago, I made a Circlotron OTL, loosely based on an Allen Kimmel design with 6C33C-b and while messing round with this, thought it would be fun to try out the smaller 6C41C...
I set it all up,--which was relatively easy, just the heater voltage and pin-out needed altering.
Initially they sounded absolutely terrible--I cant exactly describe why or how--Just BAD!
Anyway, something happened, a friend called, cant remember what, and I just turned vol down and forgot about it for a good couple of hours....
When I came back to it, the sound had transformed--sounded quite good, didn't have the power of the big old 33C (as its only around a third the cathode area) but at normal listening was very pleasant...
They were biassed at 100mA and at 200V +B
Guess these bottles need a Really Good Burn in...
I think 'Bas-Horneman' hereabouts has built a few amps with these, and he likes them as far as I can recall.....
A good while ago, I made a Circlotron OTL, loosely based on an Allen Kimmel design with 6C33C-b and while messing round with this, thought it would be fun to try out the smaller 6C41C...
I set it all up,--which was relatively easy, just the heater voltage and pin-out needed altering.
Initially they sounded absolutely terrible--I cant exactly describe why or how--Just BAD!
Anyway, something happened, a friend called, cant remember what, and I just turned vol down and forgot about it for a good couple of hours....
When I came back to it, the sound had transformed--sounded quite good, didn't have the power of the big old 33C (as its only around a third the cathode area) but at normal listening was very pleasant...
They were biassed at 100mA and at 200V +B
Guess these bottles need a Really Good Burn in...
I think 'Bas-Horneman' hereabouts has built a few amps with these, and he likes them as far as I can recall.....
Last edited:
After my initial experiments, I tried various other currents from 50 to 300mA to see what differences were...
Best ended up at 100mA and 200V. unfortunately, I couldnt alter the +B with that particular machine, but thought they may like a slightly higher +b than I was using....
It was an interesting experiment though...
I may try higher, with the switching cathode bias supply, I can afford to dump more voltage across the cathode bias network without fear of melting resistors. I seem to remember someone saying that these tubes sound better when biased to wiithin an inch of their lives (not George, but a kindred spirit)....
A higher plate voltage with the same dissipation would make the tubes more compatible with the output iron I have on hand. With a 200V anode voltage, I'd be using one of my 6.6k transformers with an 8 ohm load on the 16 ohm output tap. The bias voltage at the cathode would start to get a little crazy, but the switching bias scheme would actually like that.
Last edited:
I tried that several years ago. I called the solid state OPT. Unfortunately I never made one that lasted for more than about 5 minutes. I blew up at least 50 mosfets before punting. I didn't understand as much as I do now, which still isn't much in the SMPS world, but I now understand that using the biggest mosfet that you can find isn't always a good idea especially when it has 10,000 pF of gate capacitance!
I'm still working on the switching bias scheme. As my 6S41S haven't arrived yet, I've been using a 6080 as sacrificial lamb, as it requires similar bias voltages. My first two attempts with my company's SMPS chips struck out - not their fault, they were designed for a different purpose. I'm rolling my own controller using a UC3843 + external mosfet, with additional circuitry tacked in place to transfom it into a volt-second mode controller. We'll see how that works out.
I'm trying to wrap my brain around the idea of a switching shunt regulator for cathode bias. One of the fundamental concepts escapes me. You say:
"I may try higher, with the switching cathode bias supply, I can afford to dump more voltage across the cathode bias network without fear of melting resistors."
It seems to me like if you're not melting resistors, there must be an equivalent amount of heat released somewhere. B+ supplies the entire voltage, the tube dissipates Ia * Vak (ignoring g2 for now)
and the rest of the power input shows up between cathode and B+ return; where does it go?
I do use a switching shunt regulator, to control battery charging from my microhydroelectric generator. It's job is to divert power to a big resistor dummy load. How is your cathode regulator going to dissipate the heat?
Cheers,
Michael
PS Unless the bias is somehow incorporated into the B+ supply such that it doesn't generate the power in the first place... would take some clever solution...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.