Ok, things are moving along. I intend to use the basic voltage doubler circuit reflected in the Fisher 400 schematic (or something very similar). But a few questions:
1) Am I right in assuming that just distal to the voltage doubler (in this case, we have 405 VDC) I can treat that voltage like any other DC voltage? Which is to say, I can knock it down with dropping resisters as one normally would? I can smooth it with reservoir capacitors as one normally would?
2) Just how smooth IS the voltage coming off a voltage doubler? Is it as rough as that coming off a rectifier tube?
3) Can one use a choke in this application? The demonstrations I have seen of an inductor versus a resister are compelling, to say the least.
Now, the 5881 shows a nominal plate voltage of 250VDC, but 500VDC possible when used in pentode push-pull mode. This will work. The 6L6 is showing nominal at 270VDC and maximum at 500VDC. This will work as well. My preferred tube will be the 6n3C. ALL the data I can find shows nominal 250VDC. Is there any reason to believe this tube cannot be driven with at least 500VDC to the plate (as a max value)?
Also, how does one calculate the value of dropping resisters in the PSU when you don't actually know the current draw of the circuit distal to the resister? If I know the source voltage, and I know my target voltage (delta V = my voltage dropped across the resister), I can use Ohms law to calculate my resister value. But I can only do this if I know how much current will be drawn through the resister. And I can never know that. Do you just do this with educated guesses and trial and error?
1) Am I right in assuming that just distal to the voltage doubler (in this case, we have 405 VDC) I can treat that voltage like any other DC voltage? Which is to say, I can knock it down with dropping resisters as one normally would? I can smooth it with reservoir capacitors as one normally would?
2) Just how smooth IS the voltage coming off a voltage doubler? Is it as rough as that coming off a rectifier tube?
3) Can one use a choke in this application? The demonstrations I have seen of an inductor versus a resister are compelling, to say the least.
Now, the 5881 shows a nominal plate voltage of 250VDC, but 500VDC possible when used in pentode push-pull mode. This will work. The 6L6 is showing nominal at 270VDC and maximum at 500VDC. This will work as well. My preferred tube will be the 6n3C. ALL the data I can find shows nominal 250VDC. Is there any reason to believe this tube cannot be driven with at least 500VDC to the plate (as a max value)?
Also, how does one calculate the value of dropping resisters in the PSU when you don't actually know the current draw of the circuit distal to the resister? If I know the source voltage, and I know my target voltage (delta V = my voltage dropped across the resister), I can use Ohms law to calculate my resister value. But I can only do this if I know how much current will be drawn through the resister. And I can never know that. Do you just do this with educated guesses and trial and error?
Agreed, just using a voltage doubler with nice large caps (no need to go nuts here) will be more than sufficient. If it was enough for the Citation II, it will be more than enough for anything that uses Fisher 400 transformers. Any chance you could measure the DC resistance of the Fisher 400 PT 160V winding? That will give a rough idea of how much current we can pull and how tight our regulation can be.
I might be inclined to use a pair of 470uF caps. Something like this:
250VXH470MEFCSN25X30 Rubycon | Capacitors | DigiKey
If you want more filtering, add a choke and another cap after it. An overkill power supply isn't a bad thing, though it is costly.
With big caps, the voltage coming off the voltage doubler is pretty smooth. It doesn't hurt to add more filtering and a choke after it, however.
I am not familiar with the 6n3C, but if it's like a 6L6, it'll take 500 volts. There are a lot of tubes that are all variations on the original 6L6 (7581, 5881, 6L6GC, 6L6G, 807, 1625, 6BG6, a few Soviet tubes I don't remember the names of, etc). Their ratings vary a bit, but they're all similar in how they behave and what they can take. I believe that the plate dissipation varies a little from version to version (from about 25W for the 6L6 and 807 to 30W for the 6L6GC). They also differ in maximum screen voltage as well- for example, the original 6L6 and the 807 have a wimpier screen grid and thus aren't as good a choice for UL operation. Also be aware that the 6L6 is a great transmitting tube- it can oscillate right up to 30MHz and beyond. Grid stoppers and good layout are important to keep it well behaved.
You won't be able to calculate the exact value of that dropping resistor until you've built the amp and determined exactly what each channel draws, but you can get pretty close right off the bat if you know your topology and tubes. What you do is you estimate your idle current draw. If you plan to idle the power tubes at 50 mA (this will depend on the tube and the output transformer), then figure out how many of what tubes will be used in the input and driver stages, and use that as a ballpark estimate for your current draw. You can get pretty close that way, and you may need to go up or down on that resistor value when you actually build and test it.
@ plimpington2
why do not you leave a known schematic? it's a lot simpler to start. you have choise your "family" of power tubes, all you have to do is look for an amp diagram that uses these power tubes and a voltage doubler. it is, I think, the best way to start. @ H713 6p3s and 6p3s-e are very close in the operating points. 6p3s is more porch 6L6 WGB but sounds rather flat. 6p3s-e is closer to 5881 / 6l6wgc. I love this tube. the datasheet is VERY conservative, Wavebourn knows them very well, it will tell you what we can or can not do with it.
why do not you leave a known schematic? it's a lot simpler to start. you have choise your "family" of power tubes, all you have to do is look for an amp diagram that uses these power tubes and a voltage doubler. it is, I think, the best way to start. @ H713 6p3s and 6p3s-e are very close in the operating points. 6p3s is more porch 6L6 WGB but sounds rather flat. 6p3s-e is closer to 5881 / 6l6wgc. I love this tube. the datasheet is VERY conservative, Wavebourn knows them very well, it will tell you what we can or can not do with it.
Last edited:
Yes, I suspect you are right. Regarding ultimate tube choice, I am most familiar with the 12AX7 and 12AU7 small signal double triodes, and I understand that the 6SL7 is basically an octal equivalent of that. Which works perfectly from my perspective. I am beginning to settle around the 6SL7 + the 6N3C PP OR the 6SL7 into 7591's PP. So looking like: Input -->switch-->1/2 6SL7(pre-amp) -->volume pot-->6SL7(driver/phase inverter)-->7591+7591.
Phase inverter would likely be cathodyne, because seems simplest for my first build. The disadvantage, as I am led to understand, is that there will be no gain on that half-stage. I suppose that is fine. I am just looking for about 20 watts per side. PSU will be voltage doubler, followed by choke and then another smoothing capacitor. I am HOPEFUL the inductor I have is of low enough DC resistance to maintain a sufficiently high B+ for my output tube plates. The choke comes off an old power supply from the 40's (I'm guessing) and I need to study up on how to measure its parameters. If I can't use the choke, I will use a dropping resister before the first B+ take-off. Thereafter, I do some additional smoothing and dropping for subsequent B+ taps. I don't have my circuit yet . . . But I'm throw a few ideas together.
I know you are right. And I just may do that. But I want to have a CHANCE design my own amp. Even if sounds crap, will be good learning experience.
Make sure you keep the volume pot outside the feedback loop. Typically in the power amp section you would have voltage amp->phase inverter/driver->output stage.
we are talking about full wave voltage doubler here, there is such a thing as half wave voltage doubler too...and that is the one to avoid...
Huh?
Let's get on the same page, a historic TungSol 5881 sheet: http://www.mif.pg.gda.pl/homepages/frank/sheets/127/5/5881.pdf
First page-- 5881 was specifically targeted at 6L6 sockets where 6L6 had been failing. The first RCA 6L6 sheet showed 400V but was quickly walked-back to 360V because the early pin-base was not long-term stable at 400V. Nevertheless designers insisted on cheating 6L6 conditions higher than this. Tube death was not the designer's problem.
However the first page TYPICAL conditions *start* at 250V. Why?
Electolytic caps made low-cost radios possible. E-caps naturally have a 450V max. You pick your PT AC voltage around 320V so in the few seconds when the rectifier is hot but the power tube(s) is cold, the e-cap does not burst. So 450V no-load. Then you have 50-100V drop in rectifier at normal load, Pre-WWII you had a speaker field-coil to power, and this was conveniently also the filter choke, and typically had 100V drop. So 450V-100V-100V makes about 250V at your main B+ node. If you look through most of the older radio tubes, they all have a ~~250V condition.
You could design for more. But in a Single Ended amp, power output is limited by Pdiss much more than voltage. A wide range of voltage will work, at not very different power output. Note the 5881 sheet says for 250V-300V-350V you get 10W-11W-13W (that last at high THD; probably <12W @ <10% THD). So there is no strong incentive for high voltage. Especially since this implies a higher load impedance, which means more turns of finer wire and much longer winding time. For many classic audio pentodes, SE, 250V is a fine place to be.
On a clean-sheet design, push-pull allows DC current to rise with signal, instead of being near-constant as it has to be in SE. Then you can raise B+ without exceeding Pdiss.
But there is another "marketing problem". The 5881 does not propose any new conditions, only the ones on the (walked-back) 6L6 sheet. It was pushed as a "rugged 6L6", not "a new tube".
The main difference though was that 6L6 Vg2 Max was 270V, 5881 is 400V. At this new higher Vg2 you could get more current into lower-Z loads. The hot trick was to use 6L6 AB2 conditions, raise Vg2, lower Vg1 to compensate, and get 40-some Watts in AB1 (no grid current).
So 250V is just a starting suggestion. 6L6 begat 5881 and then TV sets begat 6L6GC, which had a 56W condition. G.E. later published a suggestion for around 70W from a pair of 6L6GC in amateur use.
Bear in mind that if you are buying new tubes, I don't think you can even buy 6L6s built to the original spec. Most of what is available is the 6L6GC, which is quite rugged in comparison, able to handle both plate and screen voltages as high as 500 volts- great for UL operation.
If you can take the extra filament current, the KT66 is even tougher, and has similar characteristics.
If you can take the extra filament current, the KT66 is even tougher, and has similar characteristics.
The original 6L6 was a metal tube, not a glass envelope tube. Nobody builds any metal envelope Octal tubes anymore. NOS is all that is available.
All 6L6 types, Except the 6L6GC, have identical specs. You might be able to find some 6L6G, 6L6GA, or 6L6GB NOS, but probably not new production. I have JJ 6L6GC tubes (30Watts plate). I love them.
I also use JJ KT66 tubes (25 Watts plate). I love them too.
Careful . . . If you use a tube in Ultra Linear Mode, please check the tube curves, especially the Screen current curves. If the screen is at 500V, and the UL tap is at 40%, consider when the plate swings down by 300V, (500V - 300V, = 200V. The plate is now at 200V). The screen will be at 500V - (0.4 x 300V); 500V - 120V = 380V Take a look at the screen current when the plate is at 200V and the screen at 380V. Wow!
Lets keep going further when the plate volts swings downward even further. The screen is lagging behind even further, but most all of the cathode current is being taken up by the screen, not the plate. Screens do not make good plates (basketball guards are not good centers).
If you doubt that the plate can go down to 100V from 500V (and the UL screen is at 340V) . . . with a low impedance UL primary of the output transformer, just consider a bass note that is at the bass driver resonance of 50Hz (and is 24 Ohms there, not 8 Ohms at that frequency). At bass resonance, the OPT primary reflects according to the 24 Ohm load, not 8 Ohms load (the primary is 3X its rated impedance in that case). (Swing Low, Sweet Plate).
Do not get me wrong, I do like UL. Just be careful if you design UL. My latest amp is KT66 SE in UL (50% UL). I am considering building SE UL 60%, but do not have a 40% UL primary that I can reverse the plate and B+ leads.
I have also used Triode wired Beam Power and Pentode tubes. But in Triode wired mode, the screen starts at the same voltage as the plate, and swings down to the same voltage as the plate, so the screen current is more controlled than in UL mode, And screen current is more controlled than in Pentode/Beam Power mode.
All 6L6 types, Except the 6L6GC, have identical specs. You might be able to find some 6L6G, 6L6GA, or 6L6GB NOS, but probably not new production. I have JJ 6L6GC tubes (30Watts plate). I love them.
I also use JJ KT66 tubes (25 Watts plate). I love them too.
Careful . . . If you use a tube in Ultra Linear Mode, please check the tube curves, especially the Screen current curves. If the screen is at 500V, and the UL tap is at 40%, consider when the plate swings down by 300V, (500V - 300V, = 200V. The plate is now at 200V). The screen will be at 500V - (0.4 x 300V); 500V - 120V = 380V Take a look at the screen current when the plate is at 200V and the screen at 380V. Wow!
Lets keep going further when the plate volts swings downward even further. The screen is lagging behind even further, but most all of the cathode current is being taken up by the screen, not the plate. Screens do not make good plates (basketball guards are not good centers).
If you doubt that the plate can go down to 100V from 500V (and the UL screen is at 340V) . . . with a low impedance UL primary of the output transformer, just consider a bass note that is at the bass driver resonance of 50Hz (and is 24 Ohms there, not 8 Ohms at that frequency). At bass resonance, the OPT primary reflects according to the 24 Ohm load, not 8 Ohms load (the primary is 3X its rated impedance in that case). (Swing Low, Sweet Plate).
Do not get me wrong, I do like UL. Just be careful if you design UL. My latest amp is KT66 SE in UL (50% UL). I am considering building SE UL 60%, but do not have a 40% UL primary that I can reverse the plate and B+ leads.
I have also used Triode wired Beam Power and Pentode tubes. But in Triode wired mode, the screen starts at the same voltage as the plate, and swings down to the same voltage as the plate, so the screen current is more controlled than in UL mode, And screen current is more controlled than in Pentode/Beam Power mode.
Last edited:
6CA7 is like an EL34, not a 6L6. I would not use a 6CA7 or EL34 with Fisher 400 outputs. EL34s and 6CA7s work well with a lower impedance primary, somewhere in the 3.5-5k range. Dynaco used a 4.3K primary which seemed to work pretty well.
Especially the late-model Fisher 400 transformers with the 10K impedance, you won't be making much power with EL34s.
Not current fashion. Historically considered equivalents.
Both US produced beam tetrodes and European pentodes (Dutch & British) were sold under both names.
http://tdsl.duncanamps.com/dcigna/tubes/sheets/amperex/6ca7-1g.gif
Also RCA and Tung Sol datasheets list both names next to each other.
Ok, but we are in the weeds now. This is a first build and I need simplicity with an emphasis on getting it to work without a whole lot of heavy lifting.
I wonder if I am better off with the 6L6CG rather than the 7591 . . . I DO have a ready made circuit/schematic that uses 6L6’s (and a host of the 6SL7’s small signal octals). But it is a PA amp. Maybe I am prejudiced, but I assume a PA amp schematic cannot be designed to deliver the fidelity I would want in this amplifier. Is that possibly wrong-headed?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.