If anyone is listening I've done my best to figure this out for myself. Nearest I can figure is that the 12AX7 is drawing about 1.1 ma and the 12AU7 ABOUT 4.6 MA. From there I calculated the B+ for the preamp section is about 331 volts. Since the JJ EL84's are limited to 300 volts I'l build the supply for about 300 and see where that gets me.
The 12AX7 is a bit harder to calculate. I think its because of the shared voltage (DC coupled) from the 12AX7 plate to the 12AU7 grid.
The 12AX7 is a bit harder to calculate. I think its because of the shared voltage (DC coupled) from the 12AX7 plate to the 12AU7 grid.
Last edited:
The voltage limit is for the TUBE, not the power supply voltage applied to the anode resistor. It's easy to get confused here because there is no anode resistor on the power tubes, so we think in terms of B+ being the limit as is the case with power tubes.
Different preamp tubes sound better with more voltages (llike the 6SN7) and some sound better with reduced voltage (like the 6n2p). Feel free to experiment a little bit and go with what your ears tell you. I don't use 12ax7 (they sound like crap), but I seem to recall from my early days of tube experimentation that they sounded better with a B+ of about 250. Obviously, the anode resistor is a factor here. Use Ohm's law and your cathode resistor to figure out current draw and make sure you have the current where you want it. Your cathode resistor is what is going to set the operating parameters for the tube. If the current draw isn't right, that's what you want to adjust to make it stay in the linear range.
the dc resistance of the output transformer is in the 100ohms range so this also limits current and drops the voltage...
in many tube amp repairs i have done, i found that lowering the output stage B+ to be very close to the output tube plate rating does not materially affect the sound, probably clipping is earlier but under normal listening conditions, i did not hear any difference...
in many tube amp repairs i have done, i found that lowering the output stage B+ to be very close to the output tube plate rating does not materially affect the sound, probably clipping is earlier but under normal listening conditions, i did not hear any difference...
I would agree with you regarding the 12ax7 tube. Varying the voltage with them was not nearly as pronounced an effect as with the 6CG7, 6SN7, 6SL7, and also seemed to depend on the circuit topology. I had an amp with either 6CG7 or 6SN7 and the B+ was around 300. I bumped it up to closer to 400 and the sound quality was noticeably better. Enough of an improvement that I modified all of my amps (that had that design) in similar fashion. It may be dependent upon the circuit topology. And, again, the popular tubes being the 12xx7 lineup, I wouldn't expect to notice as big a diff with them, if at all. The guy that makes the Big Dawg guitar amps, he claims the 12au7 needs a pretty stiff voltage to sound decent. That's why I experiment a lot. I also use some weird designs of my own, so I have nothing to guide me and I'm forced to experiment a great deal.
I think you are going to be fine at that level. If the 12au7 were a gain stage, I would say kick it up some, but since it's not, you should be fine, and 300 should be fine for the 12ax7 stage; I have seen that as the B+ for 12ax7 in many, many schematics. I suspect that is because many designers like to "push the envelope" with regard to B+. The 6BQ5 is a pretty sensitive tube and this arrangement should provide plenty of gain. The 12au7 is basically functioning like a cathode follower and should just pass the signal along with no gain and coloration of the tone.
My experiences with power tubes would again force me to say, "depends." My first amp I experimented on was a 6V6 push pull pentode mode. I left the screen voltage alone, removed the NFB, and bumped up the plate voltage. Totally changed the sound. It was clearer, and not quite as warm. I left it that way as it was my midrange treble amp and I wanted the highs accentuated more. I was on a kick where I changed from tube to solid state rectifiers and, with the 6L6GB and GC tubes, it made them "punchier" more dynamic. Again, I was changing the spread between the anodes and screens. I assumed that changing the spread between anode and screen was responsible for the change in character. I think the 6V6 rep for "warmth" comes from the fact that the data sheets recommend a very small difference between screen voltage and plate voltage. Just guessing there.
Then, my latest build, the GU50 push-pull ultralinear. It has an anode limit of 1000 volts, but a screen limit of 300 and triode mode limit of 400 volts. Since I went with ultralinear, I used the 400 volt limit and put a pretty stiff resistor on all the screens. I also went fixed bias this time so that I would get that extra 25 or 30 volts across the tube because the GU50 definitely sounded better with more voltage across it. as the voltage fell from 400 volts, the sound quality really suffered. My next project is going to be GU50 screen drive where that screen voltage limitation won't be a factor.
And then again, I have had amps where changing the plate voltage didn't seem to have much effect at all. Tubes are mysterious, and I think that is why I like them. I try to not make assumptions and instead experiment a lot just to see what happens. Sometimes I get pleasantly surprised. My first unexpected surprise was when I was doing an experiment with a 6L6GB or GC amp and the power supply had a bad connection such that the voltage was seriously reduced (pentode mode amp). The sound was VERY sweet, like triode sweet. Just really pretty. Was it the power tubes being run at reduced voltage or the preamp tubes? ? I'm not sure. That was a long time ago and I keep meaning to get back to that concept and check it out, I just haven't done so yet.
I did notice that ultralinear mode on most amps (except the GU50) seemed a little more forgiving of B+ variations. I have concentrated on inexpensive but decent tubes so I can build more and experiment more within the confines of my budget, and I stay away from lower powered stuff anymore. They just don't sound as good (to me). I found that if the bass isn't solid, just as tight and distortion free as possible, it doesn't matter what u do with the rest of the system, it won't sound right. It may sound "okay," but it isn't going to knock anybody's socks off. Every instrument has it's base or resonant frequency solidly in the bass region (Bob Heil, "The Practical Guide To Concert Sound.")
Last edited:
oh yes, "depends" is the right word, and i am not advocating that what i did is the right thing to do for others, it is just that i got those results and is happy with it....
here is another, a G2 regulated pentode can sound better than ultralinear pentode, again "it depends" is applicable here as well...
here is another, a G2 regulated pentode can sound better than ultralinear pentode, again "it depends" is applicable here as well...
That's what I was talking about, screen drive. That's my next experiment. The catch is you need to really kick the drive voltage up, and the designers say you need to be able to source some current (screen grid current). I bought some tubes for that, but one could just as well use MOSFETS since it would be the FET equivalent of a cathode follower and wouldn't have any effect on the tone. (just a buffer and not a gain stage). I have tubes and FETS both for that project. They normally use TV sweep tubes, but I'm going to use the GU50. It sounds so beautiful ultralinear, if this really is a superior topology, this is going to be an awesome amp, and cheap to build.
If you decide to play with the GU50, Get the RUSSIAN sockets. I broke half my tubes with the chinese ones. very disappointing lesson. Also, you will want to get the LS50 data sheets. They are in German. I looked for English and could not find them, so I spent a lot of time on the internet looking up each and every term and made a cheat sheet so I could understand the German data sheet. The Russians made an excellent copy of the Telefunken LS50 and they call it the GU50. This tube, it seems to me, would be a great candidate for Screen drive (G2 drive, as you call it). If you want to go that route, I believe I have a preamp/driver design that will work better than what you see out there in internet land. I have been using a version of it in my amps of late and it sounds WAY better than any of the common preamp designs. IF you aren't afraid of a little solid state, that is. The GU50 really is an audiophile tube. I built a 6550 amp and the GU50 is much better sounding.
A cool feature of screen drive is you can use the grid (G1) to set the bias or idle current. Some of the designers are a bit lazy and use cathode bias.
I just realized you are in PI? Your English is excellent.
In fact, I think I have it figured out where it can be DC coupled (no capacitors) without making the circuit complicated.
Last edited:
If you add the adjustable fixed bias mod EFB for about $10 in parts by Dave Gillespie you gain longer tube life, better sonics and 2-3W/CH more. You can add more pots to adjust a pair of output tubes or individual if you wish.
It's on the Tronola Electronics site and diytube site under ST-35.
It's on the Tronola Electronics site and diytube site under ST-35.
Thanks very much f0r all the replies! I've been working with the amp for a while now and I'm stumped. I added Dave's EFB and that works just fine. The problem is in the front end. I rebuilt the power supply to give me a 300 volt B+ for the JJ EL84's. But I just don't have any voltage on the preamp tubes? With the B+ lowered the plate and cathode resistors for the ST35 are obviously wrong. Dynaco gives you voltages at the pins measured to ground for the 12DW7. Pin 6 (plate of 12AX7 112 volts) I have 90, pin 1 (plate of 12AU7, 215 left channel and 230 of right channel {Why the difference?]) I have 22?
So I set out to change the plate and cathode resistors for both tubes. Problem is it doesn't seem to have any effect? With a new B+ of 280 volts The cathode resitor on 12AX7 didnt really seem to change much, around 1.3k. 12AU7 went from 33k cathode to 1.5k which should have gotten me 4.5ma with a grid voltage of -6. I ended up with 13.3ma! The only thing I can figure is the direct coupling with 112 on G2 of 12AU7 is putting much more current through the 12AU7 cathode resistor? How do account for this? If I AC couple the two tubes will it be easier to calculate cathode and plate resistors?
So I set out to change the plate and cathode resistors for both tubes. Problem is it doesn't seem to have any effect? With a new B+ of 280 volts The cathode resitor on 12AX7 didnt really seem to change much, around 1.3k. 12AU7 went from 33k cathode to 1.5k which should have gotten me 4.5ma with a grid voltage of -6. I ended up with 13.3ma! The only thing I can figure is the direct coupling with 112 on G2 of 12AU7 is putting much more current through the 12AU7 cathode resistor? How do account for this? If I AC couple the two tubes will it be easier to calculate cathode and plate resistors?
Last edited:
Maybe don't worry about the higher V as it's the dissipation one has to worry about and the fixed bias EFB mod solves that. The JJ's or any other new production EL84/6BQ5 can then run fine.
If you added lots of resistance to lower the B+ it may hurt the sonics also. Copy the Eico HF-81 PS for the best sonics.
If you added lots of resistance to lower the B+ it may hurt the sonics also. Copy the Eico HF-81 PS for the best sonics.
it's the cathode resistors!
Someone correct me if I'm wrong here, but I believe you have the 12ax7 as a gain stage feeding the 12au7 which is acting as the splitter-inverter. If that is the case, you MUST have equal cathode and plate resistors on the 12au7. If you go from DC coupling to AC coupling, you are going to need a resistor between the cathode resistor and the cathode of the 12au7, a small value resistor. At the junction of that small value resistor, you will need a 470K resistor that goes to the grid. LOOK at the Angelfire website FUN WITH TUBES. That guy has a great example you can follow.
IMHO, it was really stupid of Dynaco to make an EL84 amp fixed bias. Kinda pointless and needlessly complicates the circuit.
Anyway, on your preamp tubes, what really matters is that you get the correct current flowing through each. You want to be at about 50% of maximum current, maybe a little greater; your goal is to keep it in the linear range. You will calculate this by measuring the voltage across the cathode resistor (or the plate resistor) and using Ohm's law. When you have that right, you want about half of your voltage dropped by the plate resistor. Tailor the plate resistor to obtain that. You will notice as you study schematics that the circuit designer may use wildly varying plate resistor values for the same voltage on 12ax7s. Generally, the 12ax7 functions best with about 82K to 100K. You CAN get more gain by increasing the resistor value, but if you do that it may look great on your scope but when you do the listening test, it sux.
IF you want to drop the B+ for your power tubes, one "trick" is to put the inductor FIRST after the rectifier tube, (diodes if solid state), then the filter cap. That will knock the voltage down quite a bit. You can also use different rectifier tubes 'cause different tubes will drop different amounts of voltage. If you go solid state, it usually sounds better plus you can use more filter capacitance. ANother thing you can do is make a solid state regulator for your B+. IT's a dirt-simple circuit, copy Nelson Pass's example and make adjustments for your higher voltage. IF you don't try to drop too much voltage, the solid state regulator is fantastic, it totally wipes out the hum, giving you a rock-solid DC, which you want.
The 12au7, if it is acting as splitte,r should have 22K plate and 22K cathode.
It's advantageous to get good isolation of the preamp's B+ from the power tubes B+ to keep the noise level down (power supply hum). Since you are dropping the power secton's B+ down to 300, you could just use Nelson's solid state regulator schematic and eliminate the reference voltage part (he tells you how), and it will, w/o hardly dropping any voltage, give you good elimination of the hum. Otherwise, if you are determined to stick with RC filtering, use at least 1K between power tubes B+ and preamp tubes. If you are using solid state rectification, you can increase filter capacitance w/o destroying the diodes.
If you are really stumped, send me the schematic and I'll figure it out for you.
i think the reason you aren't seeing the voltage you expect to at the plate of the 12ax7 is because it is DC coupled to the grid of the 12au7 which now has a reduced B+. I think if you went to AC coupling, you'd have less trouble. The proper value of that small resistor that needs to be between the cathode and the present cathode resistor is about 1/10 the value. IT is going to function like a normal cathode resistor and will set the grid bias. Hopefully I'm being clear here.
I have seen the plate and cathode resistor of a 12au7 as high as 47K, so your 33k resistors may be okay, but I think with reduced B+, you might want to decrease to 22K. Measure the current through the 12au7 AFTER you have the circuit flaws fixed and see what's what
That DC coupling is nice in theory, but it does complicate circuit design somewhat. I have tried it both ways and frankly I didn't hear a difference.
Someone correct me if I'm wrong here, but I believe you have the 12ax7 as a gain stage feeding the 12au7 which is acting as the splitter-inverter. If that is the case, you MUST have equal cathode and plate resistors on the 12au7. If you go from DC coupling to AC coupling, you are going to need a resistor between the cathode resistor and the cathode of the 12au7, a small value resistor. At the junction of that small value resistor, you will need a 470K resistor that goes to the grid. LOOK at the Angelfire website FUN WITH TUBES. That guy has a great example you can follow.
IMHO, it was really stupid of Dynaco to make an EL84 amp fixed bias. Kinda pointless and needlessly complicates the circuit.
Anyway, on your preamp tubes, what really matters is that you get the correct current flowing through each. You want to be at about 50% of maximum current, maybe a little greater; your goal is to keep it in the linear range. You will calculate this by measuring the voltage across the cathode resistor (or the plate resistor) and using Ohm's law. When you have that right, you want about half of your voltage dropped by the plate resistor. Tailor the plate resistor to obtain that. You will notice as you study schematics that the circuit designer may use wildly varying plate resistor values for the same voltage on 12ax7s. Generally, the 12ax7 functions best with about 82K to 100K. You CAN get more gain by increasing the resistor value, but if you do that it may look great on your scope but when you do the listening test, it sux.
IF you want to drop the B+ for your power tubes, one "trick" is to put the inductor FIRST after the rectifier tube, (diodes if solid state), then the filter cap. That will knock the voltage down quite a bit. You can also use different rectifier tubes 'cause different tubes will drop different amounts of voltage. If you go solid state, it usually sounds better plus you can use more filter capacitance. ANother thing you can do is make a solid state regulator for your B+. IT's a dirt-simple circuit, copy Nelson Pass's example and make adjustments for your higher voltage. IF you don't try to drop too much voltage, the solid state regulator is fantastic, it totally wipes out the hum, giving you a rock-solid DC, which you want.
The 12au7, if it is acting as splitte,r should have 22K plate and 22K cathode.
It's advantageous to get good isolation of the preamp's B+ from the power tubes B+ to keep the noise level down (power supply hum). Since you are dropping the power secton's B+ down to 300, you could just use Nelson's solid state regulator schematic and eliminate the reference voltage part (he tells you how), and it will, w/o hardly dropping any voltage, give you good elimination of the hum. Otherwise, if you are determined to stick with RC filtering, use at least 1K between power tubes B+ and preamp tubes. If you are using solid state rectification, you can increase filter capacitance w/o destroying the diodes.
If you are really stumped, send me the schematic and I'll figure it out for you.
i think the reason you aren't seeing the voltage you expect to at the plate of the 12ax7 is because it is DC coupled to the grid of the 12au7 which now has a reduced B+. I think if you went to AC coupling, you'd have less trouble. The proper value of that small resistor that needs to be between the cathode and the present cathode resistor is about 1/10 the value. IT is going to function like a normal cathode resistor and will set the grid bias. Hopefully I'm being clear here.
I have seen the plate and cathode resistor of a 12au7 as high as 47K, so your 33k resistors may be okay, but I think with reduced B+, you might want to decrease to 22K. Measure the current through the 12au7 AFTER you have the circuit flaws fixed and see what's what
Last edited:
The EFB fixed bias mod is actually a great sonic improvement over the cathode bias and is also better than regular fixed bias as proven by Dave if one reads his documentation on the Tronola Electronics site and follow up from others who have tried. It also low cost ($10) and simple (few parts)
I'm sure I'll take some heat for this, but you would have been better off with the Russian 7189s. THey can handle a pretty hefty B+ without smoke-checking. Guitarist friends of mine who have tried them on my recommendation are very happy with them. The Russian tubes get little respect but I have gotten nothing but good results from their preamp and power tubes, more so than more expensive AMerican tubes
6P14P-EV
6P14P-EV
Last edited:
The voltage difference is because your tubes aren't matched. I wouldn't sweat the variation. The 12au7 is working like a cathode follower. It isn't providing gain. It's just passing the signal along to the next stage.
Your current went way up because you have to keep the the same value resistor on plate and cathode on the 12au7. with the 12au7 now DC coupled to the 12ax7, you have the 12au7 running wide open throttle, conducting as hard as it can. That's why you got way more current than you expected.
It is NOT "putting more current thru the cathode resistor" The grid is incapable of passing jack for current. It would simply melt down before that happened. You are going to burn up the 12au7 if you have a small value cathode resistor.
i have had good experience with Russian tubes also.......
but let us keep mum about this....
might increase their price if we keep talking about how good thy are....
welcome to the real world......
twin triodes do not have the same identical characteristics between sections,
that is why they measure differently in yours,
and there is no big reason for them to be so...
also with tubes, just because the scheme says 100v, so it should be 100 in your circuit,
tubes are never like that, as your experience have shown...
changing plate and cathode resistors changes the gain and output swing of that stage,
as you said you did not hear any different....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.