I'm planning on building a push-pull amp using Russian EL509s in enhanced triode (screen drive) mode, with an output stage B+ of about 400V. I have a lot of possible options for the preamp/drive stage. One is using a couple of the Russian 6BM8s just like in the Svetlana single ended design. I was also thinking of using a pair of vertical oscillator/amp dual dissimilar triodes like the 6CY7. My question is this - in the enhanced triode mode, what kind of screen current can I expect? Do I really need a husky 8W tube section like in the 6BM8, or can I get away with a smaller (~5W) section like in one of the dual dissimilar triodes? Also, what sort of output iron would be suitable for the 509s in enhanced triode mode?
I use MOSFETs for mine- it's a lot to ask out of a tube to drive a load like that. It's more than just max current, it's also the nonlinearity of the load. You need super low source impedance.
I haven't driven 509s in that service, but 6LF6s that I've used want to see 20-30mA at full tilt.
(edit: 400V is too low unless you're happy with 30W- these tubes give better performance at 700-800V)
I haven't driven 509s in that service, but 6LF6s that I've used want to see 20-30mA at full tilt.
(edit: 400V is too low unless you're happy with 30W- these tubes give better performance at 700-800V)
Interesting - I wanted to stay vacuum state with this project for the most part. An alternative to using MOSFETs might be to use a crude and hairy monster like the 6080 in the driver stage with 100-150V plate voltage. This would require a separate supply for the driver stage, but what the heck. I'm already in for a separate transformer to run the 509 filaments (10A total!) anyway.
I plan to use an old R-390A receiver power transformer to run the power stage. It has a 280-0-280 @200ma winding. I could float the center tap and use a full wave bridge to get 800V for the 509 plate supply. I'll have to check the isolation ratings on the plate winding to see if I can pull it off. The alternative would be to use a biphase doubler, which is a lot more parts. The transformer also has a 6.1V/8A winding I can use for the 6080 filaments.
I plan to use an old R-390A receiver power transformer to run the power stage. It has a 280-0-280 @200ma winding. I could float the center tap and use a full wave bridge to get 800V for the 509 plate supply. I'll have to check the isolation ratings on the plate winding to see if I can pull it off. The alternative would be to use a biphase doubler, which is a lot more parts. The transformer also has a 6.1V/8A winding I can use for the 6080 filaments.
I'll be nice and not ask what happened to that R-390A. What a great piece of gear....
You may find that 150V of B+ on the driver isn't enough- you've got to swing very, very positive to tickle those grids. A 12B4 might be a good candidate. Note that you can actually run the CF closer to class B; you don't really care about the negative swings, the output tube is cut off.
Take a look through some of the schematics on David Berning's site- he uses both tubes and solid state in various amps for that follower.
You may find that 150V of B+ on the driver isn't enough- you've got to swing very, very positive to tickle those grids. A 12B4 might be a good candidate. Note that you can actually run the CF closer to class B; you don't really care about the negative swings, the output tube is cut off.
Take a look through some of the schematics on David Berning's site- he uses both tubes and solid state in various amps for that follower.
The 6080 will be used as a straight voltage follower, so it will pass on whatever voltage swing was developed in the previous stage. The tube has a high transconductance and is capable of high plate current at a very modest plate voltage, so it will make an excellent cathode follower. The 6080 also has a max dissipation of 13W/side, so I can use a high bias current without frying anything. The buffer makes life easy for the voltage gain stage in front of it, since it won't have to supply significant current. I'm toying with the idea of using a replacable standard size plate that fits in a hole in the chassis so I can easily sub in various front end circuits. I'm thinking of trying a standard voltage amplifier/concertina splitter (6CY7) vs a cascoded differential front end ( 5687 or 6SN7/6CG7).
12B4 IS a good candidate for the job!
I used them as FC driver stages in my first PL519 screen drive prototype and it worked great.
In the mkII version I used mosfets instead, which worked just as fine though they didn´t look quite as good
🙂
I used them as FC driver stages in my first PL519 screen drive prototype and it worked great.
In the mkII version I used mosfets instead, which worked just as fine though they didn´t look quite as good
🙂
If a 12B4A will make the grade as a driver for an EL509, then triode #2 from a 6CY7 should work as well. Check the specs - it's a hot little number, with the same dissipation as a 12B4A, and similar transconductance (not quite as much, but still pretty good). As a bonus, you get an extra pair of triodes for gain and splitter duties.
have you looked at 6CM7? this looks good although for a 150volt screen drive, looks to me gain may not be enough if global nfb is to be used.
I have a nice little pile of 6CM7s (another project in mind), but the 2nd triode on the 6CY7 is a bit beefier (higher transconductance), and the 1st triode has more mu. The two triodes in the 6CM7 are more nearly similar. There's another subset of these vertical deflection tubes with two low mu monsters. I haven't tried picking up any of those yet. Another obvious choice might be the 12BH7, but that tube is just a bit too trendy lately, and prices are starting to go through the roof.
BTW - SY, as far as I know, no R-390A was harmed to get the transformer I'm using for this amp. Fair Radio sells them as power supply spares.
BTW - SY, as far as I know, no R-390A was harmed to get the transformer I'm using for this amp. Fair Radio sells them as power supply spares.
6cm7 here costs around $2 apiece, looking at them, the sections appear identical yet they have dissimilar curves from the tube manual.
novar 6FM7 can be purchased here at the same price and one section has even more plate dissipation.
very interested how this thread develops as i too have 2 6HF5's waiting to see action, i already have the power supply and output traffo and an 8H choke on hand which i diy'ed.
i am using a 600volt B+ and a 4.2k output traffo with 4,8,16ohm taps.
novar 6FM7 can be purchased here at the same price and one section has even more plate dissipation.
very interested how this thread develops as i too have 2 6HF5's waiting to see action, i already have the power supply and output traffo and an 8H choke on hand which i diy'ed.
i am using a 600volt B+ and a 4.2k output traffo with 4,8,16ohm taps.
i will be using IRF820 mosfet followers as SY suggested as it makes a lot of sense to me, tubes are so precious to waste as cathode follower imho. loss of gate voltage does not mean the end out the ouput tubes, screen grid goes negative and tube is cut-off, very neat indeed.
i made a dedicated shunt regulated +/-200volts supply for the mosfets, i used bjt shung regs and the psu is good up to 30mA as SY hinted.😀
hope i get more hints from him and others too😀
i made a dedicated shunt regulated +/-200volts supply for the mosfets, i used bjt shung regs and the psu is good up to 30mA as SY hinted.😀
hope i get more hints from him and others too😀
The power supply for the MOSFETs just isn't very critical. The rejection at the drain is very high, and you can feed the source through a very high value resistor, which reduces the influence of the negative rail. The '820 is a reasonable choice (that's what I use), but the DN2540 might even be better... The nice thing about MOSFETs (besides excellent linearity in source follower configuration) is that you can swing right up to within a couple of volts of the rail.
Do make sure to have gate protection diodes, a gate-stopper resistor as close to the case as possible, and heatsink them well.
Do make sure to have gate protection diodes, a gate-stopper resistor as close to the case as possible, and heatsink them well.
The MOSFETs ae a neat solution (I have scads of IRF820s, and I also have some HV depletion mode MOSFETs), but I want to try vacuum first out of sheer perversity. I just scored a whole mess of 6GF7 vertical deflection tubes on Ebay. They were really, really cheap. These are a high-mu (64) triode and a second heftier low-mu unit (10-11W) in a compactron-style bulb. As such, the envelope style is consistent with the EL509s, so they'll look nice together, and I'll have enough capability for some serious screen drive.
That's a nice tube- I've used them in small push-pull amps as combo driver/phase-splitter/output. Max plate voltage rating is kinda low for the swing you need in this application.
I was looking at possibly using the Hammond 1650R for output iron, or would the 1650N or R be more appropriate?
indeed, that is why i am trying to source them in hongkong with the help of a friend, since they are all made in china nowadays, i thought it might be better if it were pruchased from there.
i have 8pcs of 6LQ6/6JE6 in need of novar 9 pin sockects, at U$2 apiece i will be spending more for the sockets which probably would cost me U$5 each.
The 1650N looks better. A lower impedance is a good thing. I'm using some Northlake units with 2K plate to plate.
Interesting. I was doing some rough back of the envelope calculations for reflected voltage and current, and sort of halfway concluded that the 1.9k Hammond transformer could work. This means that the 509s would be capable of some pretty hefty peak power. I'm going to go back and stare at the tube curves a bit to see if this will really work.
- Status
- Not open for further replies.