Hi All:
I was reading through an old Radio-Craft book (from 1947) and came across this Direct-Coupled 10 P-P amp design using 6L6s. And, it is fully balanced as well.
All 6 pages of the article are attached.
I have read it through and it is truly direct-coupled - NO caps in the signal path, only the PS.
I have attached the whole article. Maybe one of you design gurus could re-calculate the values to build this with better (read quieter) tubes than the 6L6. How 'bout EL34s or 6550s? Also, the input tubes are 6SJ7 - maybe change them to 6DJ8 or 6H30?
I am pretty well decided to build one of these to see how it sounds, but would love any input from the tube experts out there.
Jess
I was reading through an old Radio-Craft book (from 1947) and came across this Direct-Coupled 10 P-P amp design using 6L6s. And, it is fully balanced as well.
All 6 pages of the article are attached.
I have read it through and it is truly direct-coupled - NO caps in the signal path, only the PS.
I have attached the whole article. Maybe one of you design gurus could re-calculate the values to build this with better (read quieter) tubes than the 6L6. How 'bout EL34s or 6550s? Also, the input tubes are 6SJ7 - maybe change them to 6DJ8 or 6H30?
I am pretty well decided to build one of these to see how it sounds, but would love any input from the tube experts out there.
Jess
Attachments
Jess,
It should work fine, but note it’s going differential, or splitting phase, off the input jack with 1meg between legs. This would probably be hard to drive straight off a CD player and may sound funny so keep that in mind. I’m not sure the feedback scheme will work as drawn unless you are using the output transformer spec’ed in the article.
On a lark I did this direct connect P-P and it sounds better than I expected. It should be clear I had lots of Edcor bits around.
Matt
It should work fine, but note it’s going differential, or splitting phase, off the input jack with 1meg between legs. This would probably be hard to drive straight off a CD player and may sound funny so keep that in mind. I’m not sure the feedback scheme will work as drawn unless you are using the output transformer spec’ed in the article.
On a lark I did this direct connect P-P and it sounds better than I expected. It should be clear I had lots of Edcor bits around.
Matt
Attachments
Thanks for the comments, Matt
I was thinking of driving it with a balanced preamp (that has SE inputs). To some extent I can tweak the preamp output so that this amp's input sees what it needs.
Won't be the first bit of gear I've built that does that...🙄
Jess
Looks to me like this scheme wastes a lot of power. (requires class A output mode even, besides the voltage dropping resistors.) The high value resistors in the output tube cathodes are going to make the output Z rather high too. I don't think that global feedback "option" is going to be optional.
Why not just use the standard approach of two high value series resistors from each of the input stage plates, connected to a B- rail (or use CCS's on them to a - 30V rail). Center connection of the series resistors goes to the output tube grids. Causes a 2X loss of gain (or no loss of gain with the -30V rail CCS's). Could put CCS's in place of the input tube plate resistors as well to B+ for more gain and less distortion yet.
Don
Why not just use the standard approach of two high value series resistors from each of the input stage plates, connected to a B- rail (or use CCS's on them to a - 30V rail). Center connection of the series resistors goes to the output tube grids. Causes a 2X loss of gain (or no loss of gain with the -30V rail CCS's). Could put CCS's in place of the input tube plate resistors as well to B+ for more gain and less distortion yet.
Don
Maybe I’m reading the schematic wrong, but I only see 1.1K in the 6L6 cathodes, even without a feedback scheme, I don’t think output Z would be unreasonable. I think the hard part of implementing that output feedback as drawn is finding an output transformer with those tap options. Using a different transformer would require recalculating feedback, and would move the plate voltage of the 6SJ7 around a bit.
One cannot leave the feedback out because the gain of the 6SJ7 would be higher than the 70 to 90dB the article says it will provide, which is already too much for a modern amp. It’s not easy to swap tubes or gain profiles with a direct-coupled design, there are too many balls in the air. It’s kind of build it exactly as someone worked it out or do your own design from scratch.
Matt
One cannot leave the feedback out because the gain of the 6SJ7 would be higher than the 70 to 90dB the article says it will provide, which is already too much for a modern amp. It’s not easy to swap tubes or gain profiles with a direct-coupled design, there are too many balls in the air. It’s kind of build it exactly as someone worked it out or do your own design from scratch.
Matt
Saving electrons was not a high priority in 1947 😀
Still, the total power consumption will not be much more than 200 watts, no?
I certainly agree with that observation - and I don't have enough design smarts to tackle one from scratch. I guess I'll just have to build it and see 😉
As for the 500 ohm tap used for feedback - will this provide a particular ratio of power for the feedback that cannot be derived from a 16 ohm tap, adjusted with different R & C values?
Jess
Could you talk more about this amp - new thread perhaps?
I have a boatload of 6aq5s - which make me interested in the schematic you posted. I have many questions - may I start a new thread? thx.
I have a boatload of 6aq5s - which make me interested in the schematic you posted. I have many questions - may I start a new thread? thx.
These aren't amplifiers, they're IMD generators. If you build a pentode without NFB, you had better be planning to use it with a guitar. 
Jess,
Sure you can recalculate feedback for the transformer you use, but it still may move the plate/grid voltage around. As long as you have about three millimeters clip leaded in your circuit and kids calling out values as you fire it up you should be fine. Tubes have a sense of humor that allows for slight mistakes as you get things working.
The French amplifier posted looks simpler then the one at the start of this thread. And its much lower gain is more in line with modern amplifiers, so it should be easier to work with. But you do have to have the tubes.
Moonbird,
There are lots of designs using 6005s, 6AQ5, EL90, and etc. so I’m sure if you start a new thread there will be lots to think about. What I learned with that design is to use ceramic sockets under the 6005 for reliability, those little tubes can pass enough current to burn up sockets. I have no experience with the tube beyond the schematic I provided.
Matt
Matt
Sure you can recalculate feedback for the transformer you use, but it still may move the plate/grid voltage around. As long as you have about three millimeters clip leaded in your circuit and kids calling out values as you fire it up you should be fine. Tubes have a sense of humor that allows for slight mistakes as you get things working.
The French amplifier posted looks simpler then the one at the start of this thread. And its much lower gain is more in line with modern amplifiers, so it should be easier to work with. But you do have to have the tubes.
Moonbird,
There are lots of designs using 6005s, 6AQ5, EL90, and etc. so I’m sure if you start a new thread there will be lots to think about. What I learned with that design is to use ceramic sockets under the 6005 for reliability, those little tubes can pass enough current to burn up sockets. I have no experience with the tube beyond the schematic I provided.
Matt
Matt
Why not use a voltage divider to -V? It works for digital, and it works for analog. Tektronix did it!
Tim
Tim
D*#n, I wish I could read French. The Lafont does look simpler.
As the text of my original article indicates very high gain (accepting high-impedance microphone input directly) I was thinking about needing to tune the input tube gain down a bit, or attenuating the input signal.
I am not quite sure where in the circuit you are referring, Tim. Can you elaborate a bit for us punters? 😉
Jess
Thanks for the response - in this case I am more interested in the 6GM8. What makes it "good for direct-coupling" as the datasheet states. Do you know? Given that both cathodes of the 6GM8 are connected - could a 6J6 (shared cathode) be substituted (with the necessary socket wiring and component changes of course) or is there something special about the 6GM8? Is it a gain issue (6GM8 mu=14 vs 30+ for the 6J6) thx again.
Here is an article about a DC-coupled el84-PP amp.
If somebody builds it, please let me know how it sounds 🙂
If somebody builds it, please let me know how it sounds 🙂
Jess,
If you google voltage divider you should be able to figure out how to attenuate your signal enough to deal with the excess gain of that amp. You can also have a look at this by Voltsecond,
S5 Attenuator with Sweet Whispers Advise
He shows many schemes for padding down the input. You can put your divider right after the input jack and keep the rest of the schematic unchanged. Remember to use insulated/isolated input jacks.
Moonbird,
I never noticed that the data sheet calls out the 6GM8’s suitability for good direct coupling. I chose it because it works at low volts, allowing me to directly couple without the need to pull it down with negative voltage under the cathode or stand the output tube high. But any tube that works at low volts is fine. I actually use a 6N6P when I saw how much 6GM8s go for now. I just plugged it in with no changes to that schematic. Any tube with a common cathode but separate grid and plate will work in that drive position.
Matt
If you google voltage divider you should be able to figure out how to attenuate your signal enough to deal with the excess gain of that amp. You can also have a look at this by Voltsecond,
S5 Attenuator with Sweet Whispers Advise
He shows many schemes for padding down the input. You can put your divider right after the input jack and keep the rest of the schematic unchanged. Remember to use insulated/isolated input jacks.
Moonbird,
I never noticed that the data sheet calls out the 6GM8’s suitability for good direct coupling. I chose it because it works at low volts, allowing me to directly couple without the need to pull it down with negative voltage under the cathode or stand the output tube high. But any tube that works at low volts is fine. I actually use a 6N6P when I saw how much 6GM8s go for now. I just plugged it in with no changes to that schematic. Any tube with a common cathode but separate grid and plate will work in that drive position.
Matt
Matt -- 1 last question - I am looking at the bias inductor thing you have going 😱 -- very unique and facinating (please explain how you arrived at the proper inductor here) and the balance pot - could you comment on that? Why use it? What effect on some and bias - is this to account for unmatched sections?? To dial out (or in) 2nd harmonic distortion - very cool -- but this would require a split cathod to work right?
Jess -- Thanks much for letting me share Mark! 🙂
Jess -- Thanks much for letting me share Mark! 🙂
Moonbird,
I chose the inductor because I had it, and it’s DCR worked for cathode bias. With a twin triode like I used it helps to balance AC between halves. The pot helps balance DC through the outputs by allowing me to tweak DC through the triode halves. It’s a fairly common thing to do, though most around here would use a CCS. The pot and the inductor would serve no purpose with a common cathode driver.
Matt
I chose the inductor because I had it, and it’s DCR worked for cathode bias. With a twin triode like I used it helps to balance AC between halves. The pot helps balance DC through the outputs by allowing me to tweak DC through the triode halves. It’s a fairly common thing to do, though most around here would use a CCS. The pot and the inductor would serve no purpose with a common cathode driver.
Matt
I don't have any Tektronix drawings to reference offhand (if you were so tempted, a perusal of BAMA for any of the older 500-series Tek scopes or plugins will turn up entire manuals for you), but I do have an example in my class D tube amp, which is DC coupled.
An externally hosted image should be here but it was not working when we last tested it.
For example, where the PWM signal enters the driver at the lower left, instead of a grid-leak resistor to ground, a voltage divider to -320V is used. Same thing at the output, 100k on top, 220 + 100k on bottom.
Although this circuit is intended for saturating operation, which means operating points are looser, the same principle applies to linear design. The trick is to use global DC feedback, like an operational amplifier circuit, to keep all operating points balanced.
There is never a need for wasting tens of watts or more across a cathode resistor. 🙂
Tim
I have, Oh about 400 of them too. Does that qualify as a boatload? I also have about 1000 6AU6's and maybe a zillion 6AG5's. So about 2 years ago I wired a few of these things together. And, yes it was a true direct coupled P-P amp. It sounded real nice too. I still have it around here somewhere, and I plan to dig it out after the Simple P-P stuff is done.
It has 4 tubes, not unlike Pete Millett's "engineers amp". I remember using some sand though.
Yes, those little black plastic sockets will MELT! The 6AQ5 gets hotter than some sweep tubes.
6aq5s
Ah - yazzur thatza boatload fer sur -- an oil tanker -- vs my bayliner of about 100 or so. I have a bunch of JAN Tung sols (6005s) and some Millards I want to try.
I have been collecting schematics for a while. It would be great to see what you could come up with Tubelab.
One of the most interesting designs I have found so far (aside from Matts) is one on Fred Nachbaur's site - called the Multimedia amp here:
Paraphase Splitter Experiments
It has a paraphase PI that uses a triode/pentrode tube (which is delightfully out of the box). He says it makes lotza 2nd order harmonics due to the mis-matched PI gain - wanna build that one.
I am also very interested in guitar amps and found an old Selmer amp that used PP 6aq5s - very chimey sounding like a early 60 brit invasion amp should sound like. There are a couple Gibson/Epiphone amps that used them too.
From what I read they sound best in PP triode mode.
Being a purist at heart - makes me want to use other 7-pin tubes - problem is the seven pins means there are no split cathode twin triodes - just the lil old shared cathode 6J6 - which can't really be cascoded. That means many tubes to get the gain of novals/octals.
Ah - yazzur thatza boatload fer sur -- an oil tanker -- vs my bayliner of about 100 or so. I have a bunch of JAN Tung sols (6005s) and some Millards I want to try.
I have been collecting schematics for a while. It would be great to see what you could come up with Tubelab.
One of the most interesting designs I have found so far (aside from Matts) is one on Fred Nachbaur's site - called the Multimedia amp here:
Paraphase Splitter Experiments
It has a paraphase PI that uses a triode/pentrode tube (which is delightfully out of the box). He says it makes lotza 2nd order harmonics due to the mis-matched PI gain - wanna build that one.
I am also very interested in guitar amps and found an old Selmer amp that used PP 6aq5s - very chimey sounding like a early 60 brit invasion amp should sound like. There are a couple Gibson/Epiphone amps that used them too.
From what I read they sound best in PP triode mode.
Being a purist at heart - makes me want to use other 7-pin tubes - problem is the seven pins means there are no split cathode twin triodes - just the lil old shared cathode 6J6 - which can't really be cascoded. That means many tubes to get the gain of novals/octals.
- Status
- Not open for further replies.