Hi,
I´ve been contemplating a bit about power amps with the output tubes configured as cathode followers. The main disadvantage is the enormous input signal required to drive the followers, we´re talking 100Vrms and up.
The plan is to build an amp using triode-wired PL519 (40KG6) and Hammond 1650T output transformers (1,9k plate to plate, or should I say cathode to cathode).
To make life a little easier for the driver stage I´m planning to run the output tubes "low and hot", 170V p-k @ 200mA and use the 16 ohm tap with an 8 ohm load to get 950 ohm c-c.
If I´m reading the curves and loadlines correctly the driver stage would have to deliver 100Vrms (per phase) to drive this output stage into clipping.
The question is: What would a suitable driver stage?
I´ve had some success using a choke loaded 12B4A to drive an SE cathode follower, but I don´t have any centertapped plate chokes and I´m on a budget here (as always...).
Can I use a pair of 12B4A with CCS plate loads (cascoded DN2540) as a differential gain stage with a common cathode resistor? I guess a third CCS in the "tail" would make a mess, but a simple resistor (for cathode bias) connected to ground should work, right?
I guess such a gain stage wouldn´t be a very good phase splitter on its own, but that´s not a problem. I would need a front stage anyway, so why not let that one handle the "phase splitting".
Would it be safe to use DN2540 as plate loads with 400V B+?
With about 200V/15mA across the chips at idle they would need heatsinks of course, but I´m more concerned about the voltage at full signal. They´re speced to 400V, can this be trusted?
(I distrust semiconductor specs in general...)
I´ve been contemplating a bit about power amps with the output tubes configured as cathode followers. The main disadvantage is the enormous input signal required to drive the followers, we´re talking 100Vrms and up.
The plan is to build an amp using triode-wired PL519 (40KG6) and Hammond 1650T output transformers (1,9k plate to plate, or should I say cathode to cathode).
To make life a little easier for the driver stage I´m planning to run the output tubes "low and hot", 170V p-k @ 200mA and use the 16 ohm tap with an 8 ohm load to get 950 ohm c-c.
If I´m reading the curves and loadlines correctly the driver stage would have to deliver 100Vrms (per phase) to drive this output stage into clipping.
The question is: What would a suitable driver stage?
I´ve had some success using a choke loaded 12B4A to drive an SE cathode follower, but I don´t have any centertapped plate chokes and I´m on a budget here (as always...).
Can I use a pair of 12B4A with CCS plate loads (cascoded DN2540) as a differential gain stage with a common cathode resistor? I guess a third CCS in the "tail" would make a mess, but a simple resistor (for cathode bias) connected to ground should work, right?
I guess such a gain stage wouldn´t be a very good phase splitter on its own, but that´s not a problem. I would need a front stage anyway, so why not let that one handle the "phase splitting".
Would it be safe to use DN2540 as plate loads with 400V B+?
With about 200V/15mA across the chips at idle they would need heatsinks of course, but I´m more concerned about the voltage at full signal. They´re speced to 400V, can this be trusted?
(I distrust semiconductor specs in general...)
Hi Fuling,
unless you are willing to invest in a separate driver stage B+ feed of considerable higher tension than the output stage, you always can consider choke-loaded drivers. Choke-loading will almost double the swing compared to simple resistive loading (where you drop much of the available B+ across the plate resistor).
Also, using pentodes (in pentode mode of course) for the drivers often helps a lot with such demands, because they usually can swing the plate much closer to zero volts across the tube than a triode can.
And there always is the path of engaging interstage transformers (even including doing the phase splitting job), but this usually will prevent the usage of global FB.
Nothing comes for free, eh.
Regards,
Tom
unless you are willing to invest in a separate driver stage B+ feed of considerable higher tension than the output stage, you always can consider choke-loaded drivers. Choke-loading will almost double the swing compared to simple resistive loading (where you drop much of the available B+ across the plate resistor).
Also, using pentodes (in pentode mode of course) for the drivers often helps a lot with such demands, because they usually can swing the plate much closer to zero volts across the tube than a triode can.
And there always is the path of engaging interstage transformers (even including doing the phase splitting job), but this usually will prevent the usage of global FB.
Nothing comes for free, eh.
Regards,
Tom
Would be great, thanks!
No problem, I have several transformers with suitable HV windings available.
I love choke loaded gain stages, but unfortunately I have no suitable plate chokes. I could of course order a pair of Lundahl LL1667 or 1668 with a small airgap (just enough to allow some DC imbalance) but I´m looking for a cheaper solution if possible.
I´ve thought about that, but it seems that the 12B4A is capable of swingning even "closer to zero" than any pentode I´ve seen.
When I was experimenting with an SE follower amp I got over 500Vp-p out of a choke loaded 12B4A with about 290V plate-cathode.
CCS plate loads should allow almost the same results, though limited by the 400V that the DN2540 can handle. With 400V B+ it should still be possible to get something like 330Vp-p out, which is quite a bit more than I need. IF the whole idea works as intended, of course...
A pair of LL1660S interstage transformers would be really cool in this application, serving as both plate loads, coupling elements and even phase splitters.
I actually have a pair of LL1671/50mA transformers already but they are not quite as versatile as the 1660S and the airgap is too big. To get an 1:1+1 ratio I would have to wire the primaries in // which would result in a very low primary inductance, it would probably take another PL519 to drive it...
If you're not going to do something like Lynn's Amity, maybe with 2A3's as drivers, then a pentode diff front end works superbly. I drove my 813PP triodes with 12HG7 and 12HL7. Superb.
http://www.pmillett.com/813_se_triode_amps.htm
http://www.pmillett.com/pentodes.htm
http://www.fortunecity.com/rivendell/xentar/1179/projects/legacy/Legacy.html
http://www.pmillett.com/813_se_triode_amps.htm
http://www.pmillett.com/pentodes.htm
http://www.fortunecity.com/rivendell/xentar/1179/projects/legacy/Legacy.html
I had a pair of decent quality 6BQ5 PP transformers but unfortunately I traded them for some rare tubes that I needed for another project.
If I can find anything suitable for a good price it would be worth a try. I guess mains transformers with 2x115V primaries cannot handle 2x100Vrms at 20Hz.
I decided to take another look at some interesting pentodes to see if I could find something suitable.
I found this one:
http://www.tubes.mynetcologne.de/roehren/daten/el83pentode_150v.pdf
I happen to have a whole box full of NOS Philips PL83...
With a 10k plate resistor and about 500V B+ this tube seems capable of swinging over 400Vpp, not bad!
I found this one:
http://www.tubes.mynetcologne.de/roehren/daten/el83pentode_150v.pdf
I happen to have a whole box full of NOS Philips PL83...
With a 10k plate resistor and about 500V B+ this tube seems capable of swinging over 400Vpp, not bad!
You can look at Tubecad for more idea of driver stage like this
http://www.tubecad.com/2006/09/blog0080.htm
or here
http://www.tubecad.com/2006/09/blog0079.htm
http://www.tubecad.com/2006/09/blog0080.htm
or here
http://www.tubecad.com/2006/09/blog0079.htm
This idea sounds similar to an Allan Kimmel OTL I saw some time ago....
That was a circlotron, and while not what you're looking for, used the PL519 as a cathode-follower from a 150V rail....
He drove the 4 PL519's on each side from a 5687 into the control-grids, But he also drove the Screens using a MOSFET for what he called 'Super-Triode' mode--He claimed a greater efficiency and linearity from what I recall....
I do remember the driver stages were supplied by a 350V rail, while the MOSFET was used as a bootstrap supply for the driver-stage, which also had a negative rail of 150V
The schematic used to be around on the 'web but has now disappeared...
That was a circlotron, and while not what you're looking for, used the PL519 as a cathode-follower from a 150V rail....
He drove the 4 PL519's on each side from a 5687 into the control-grids, But he also drove the Screens using a MOSFET for what he called 'Super-Triode' mode--He claimed a greater efficiency and linearity from what I recall....
I do remember the driver stages were supplied by a 350V rail, while the MOSFET was used as a bootstrap supply for the driver-stage, which also had a negative rail of 150V
The schematic used to be around on the 'web but has now disappeared...
Yes, the screen grids opens a world of possibilities in a design like this. I´ve thought about using µ-followers as drivers with the low Z-outputs direct coupled to the screen grids and then cap coupled to the control grids.
Sort of a "super pentode connection". This *should* give a very high transconductance and thus be suitable in a cathode follower circuit.
Regular pentode mode (screen voltage constant WRT the cathode) would probably be great too, but it would require a separate Mosfet follower referenced to each tube´s cathode.
Anyway, I printed the PL83 plate curves and went to work with a ruler and a calculator. With an 8,8k plate load and 480V B+ it can deliver 135Vrms (190Vp) with more or less perfect symmetry!
Sort of a "super pentode connection". This *should* give a very high transconductance and thus be suitable in a cathode follower circuit.
Regular pentode mode (screen voltage constant WRT the cathode) would probably be great too, but it would require a separate Mosfet follower referenced to each tube´s cathode.
Anyway, I printed the PL83 plate curves and went to work with a ruler and a calculator. With an 8,8k plate load and 480V B+ it can deliver 135Vrms (190Vp) with more or less perfect symmetry!
You mean that gridchoke-thingie? I guess it would work fine with the right choke, I tried it once with a regular toroidal transformer and got nothing above 8kHz.
I believe an LTP with PL83 pentodes will be the first thing to try.
Can I expect troubles caused by the screen grid currents when I use pentodes in an LTP?
When I built my 300Beast (P-P 300B) and the 6AV5 screen drive amp I used a twin triode LTP as the first stage and phase splitter. The 300B amp used a 5751 and the 6AV5 amp used a 6BQ7. The following stage is the driver which uses two identical amplifier circuits. The 300B used 6FQ7's in SRPP since they were in vogue at the time. The 6AV5 amp used a conventional common cathode driver circuit using 5965's. In both cases the cathodes of the triodes are tied together without bypassing as in a diff amp. This configuration evens out any differences between the two stages. Both amplifiers sound great, and provide plenty of voltage swing. The 6AV5 amp is screen driven so lots of drive voltage is required. I run the 5965's from a 550 volt supply. I have the schematic somewhere but I am at work and I don't have it here.
Perhaps you could use an LTP phase splitter for your first stage, followed by pentode drivers with their cathodes sharing a common unbypassed resistor.
Perhaps you could use an LTP phase splitter for your first stage, followed by pentode drivers with their cathodes sharing a common unbypassed resistor.
Yes, a separate phase splitter might be a good idea.
Not that I need the extra gain (not much at least) but it would probably make life easier for the driver stage.
How about using Edcor 10k/10k interstage transformers as phase splitter driven by a low µ parafeed input stage (12B4A for example)?
Not that I need the extra gain (not much at least) but it would probably make life easier for the driver stage.
How about using Edcor 10k/10k interstage transformers as phase splitter driven by a low µ parafeed input stage (12B4A for example)?
Push-Pull PL519s
I have used these tubes for an SSB transmitter where the aim is the same as in an audio amplifier - minimum distortion. I find yoiu get minimum distortion not in class-A but in class-B - actually class AB, as there is an optimum bias. It was a while ago that I did this, so I am not sure how I did it , but I think I just read the points off the curve of anode current versus g1 voltage and typed it into a file. Then I wrote a program to plot the curve. After that, I did the push-pull curve by tuning one curve around (mirri image in the X and Y axes) and adding it to the other, with an X-shift representing the intended static bias. I then varied the shift until the sum of the two lines was as staright as possible. I subtracted the mean slope and magnified the difference to increase the sensitivity. I seem to recall I came up with the need for something like -44 volts standing grid bias, which I did with two 22 volt zener diodes in the cathode circuit.
How much grid drive voltage you need depends on how much peak current you want out. I wanted 400mA, which occurs at -20 volts g1, so that means 24 volts peak, or 48 volts peak-to-peak on each grid, or 96 volts peak-to-peak between grids (34vrms)
I have a favorite tube audio circuit which is extremely linear, and works like this: Use a small power supply transformer that has a 120+120 primary (suitable for 120v or 240v input) and a 12-012 volt output. Connect the cathode of one tube through the 44v zener to one secondary terminal together with the grid of the other tube, and vice versa. Ground the secondary center tap.
For the driver, I use two 6V6's connected to the primary in push pull. EL84's work fine. The drivers are biassed the same - in class-AB with zener diodes in the cathode circuit. You need 22 volt zeners for 6V6s.
Because there is huge negative feedback in this half-grounded cathode, half grounded grid circuit, using a power supply transformer as the driver transfortmer had no noticeable problems with quality. When I figure out how to post diagrams, I'll post the schematic.
One more thing - PL519s will not work as well in triode mode. They like to have g2 at a fixed 170 volts.
P.S. I collected my PL519s and PL509s from old TV sets on garbage dumps in Sweden in the 1990's - a great source of tubes, bases, capacitors etc.
I have used these tubes for an SSB transmitter where the aim is the same as in an audio amplifier - minimum distortion. I find yoiu get minimum distortion not in class-A but in class-B - actually class AB, as there is an optimum bias. It was a while ago that I did this, so I am not sure how I did it , but I think I just read the points off the curve of anode current versus g1 voltage and typed it into a file. Then I wrote a program to plot the curve. After that, I did the push-pull curve by tuning one curve around (mirri image in the X and Y axes) and adding it to the other, with an X-shift representing the intended static bias. I then varied the shift until the sum of the two lines was as staright as possible. I subtracted the mean slope and magnified the difference to increase the sensitivity. I seem to recall I came up with the need for something like -44 volts standing grid bias, which I did with two 22 volt zener diodes in the cathode circuit.
How much grid drive voltage you need depends on how much peak current you want out. I wanted 400mA, which occurs at -20 volts g1, so that means 24 volts peak, or 48 volts peak-to-peak on each grid, or 96 volts peak-to-peak between grids (34vrms)
I have a favorite tube audio circuit which is extremely linear, and works like this: Use a small power supply transformer that has a 120+120 primary (suitable for 120v or 240v input) and a 12-012 volt output. Connect the cathode of one tube through the 44v zener to one secondary terminal together with the grid of the other tube, and vice versa. Ground the secondary center tap.
For the driver, I use two 6V6's connected to the primary in push pull. EL84's work fine. The drivers are biassed the same - in class-AB with zener diodes in the cathode circuit. You need 22 volt zeners for 6V6s.
Because there is huge negative feedback in this half-grounded cathode, half grounded grid circuit, using a power supply transformer as the driver transfortmer had no noticeable problems with quality. When I figure out how to post diagrams, I'll post the schematic.
One more thing - PL519s will not work as well in triode mode. They like to have g2 at a fixed 170 volts.
P.S. I collected my PL519s and PL509s from old TV sets on garbage dumps in Sweden in the 1990's - a great source of tubes, bases, capacitors etc.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.