EL84 + 6L6 in parallel in same amp! 3.3K primary?

[Hardtailed]

Yeah, I'm back with this old idea that's haunting me! Using a pair of EL84 and a pair of 6L6 in the same amp!

The biggest problem to me was matching the primary of the OT, but then it occured to me that:
- the Vox AC15 uses a 6.6K output impedance (vs the typical 8K for EL84s)
- you can run a pair of 6L6 with a lower B+ in a 6.6K impedance too, like people putting 6L6s in their 5E3!

So since we have a pair of each, we divide 6.6 by two which give 3.3Kohms.
Hey! The Hammond 1750N (Marshall 50W) OT that I love in my Marshall clone is 3.4Kohms! Close enough for rock 'n roll.

So let's accept that we will need to use lower voltage than we would typically use with 6L6s and run this thing pretty low. The Vox AC30 runs on 320V, let's use that.

And let's run the EL84 cathode-biased... again like the Vox, but use fixed-bias for the 6L6 (we split after the PI to the two power amp lines, caps will block the bias signal from reaching the EL84s). That'll help us keep as much voltage as we can for the 6L6s.

Judging from the data sheets, I am expecting between 40 and 45W of total output power. Care will have to be taken to evenly divide the signal accross the two pairs, but I'm basically hoping the EL84 will overdrive first, the idea being to keep a tight and firm low-end from the 6L6 while the EL84 are generating their nice harmonics! With the EL84 high gain, that should be easy.

Do I have the math right here?

2 cathode bias EL84
2 fixed bias 6L6
320V B+
3.3-3.4KOhms output primaries
Under 50W of output

[Hardtailed]

It occurs to me that I could also lower the reflected plate voltage of the EL84 by floating them above ground...

Basically, add another resistor after the cathode biasing one, and "ground" the grid in between.

Say we have a 390V B+, we need to drop 70V (or punish the EL84s like there's no tomorrow!). If all goes well, we are expecting about a 77mA of current at idle from the pair of EL84s, so a 900 ohm resistor would do, bypassed by a huge cap.
The elevated ground would also apply to the screens, so it all works out!
With the screen current on top of the anode current, let's say a 820ohm resistor (common value) in 10W rating would work.

Has this been ever done? Floating power tubes above ground to lower their cathode to plate voltage?

[HTH]

Has this been ever done? Floating power tubes above ground to lower their cathode to plate voltage?

more than likely - when it comes to valves its ALL been done before somewhere or other.

wierdly enough, there was a guy on another forum talking about this exact same thing today as a way to reduce the voltage across the output valve.

I'll be interested to know how this works out for you.

[Bruce / Mission Amps]

Yeah, I'm back with this old idea that's haunting me! Using a pair of EL84 and a pair of 6L6 in the same amp!

The biggest problem to me was matching the primary of the OT, but then it occured to me that:
- the Vox AC15 uses a 6.6K output impedance (vs the typical 8K for EL84s)
- you can run a pair of 6L6 with a lower B+ in a 6.6K impedance too, like people putting 6L6s in their 5E3!...

I'm curious where "the Vox AC15 uses a 6.6K output impedance" in a VOX AC15 came from.
Regardless, I'm sure this would work fine though.

[Hardtailed]

I'm curious where "the Vox AC15 uses a 6.6K output impedance" in a VOX AC15 came from.
Regardless, I'm sure this would work fine though.

That's what I remember reading (in discussions where people were comparing the Marshall 18W to the AC15)

The Hammond replacement transformer for the vintage AC15 is 6.2K (1650Y I believe). First example I could find by "Binging"

A little part in my brain thinks that the EL84s, by pulling somewhat less current, might "see" a higher impedance (less current, same voltage minus the drop accross the cathode resistor...).
As long as nothing blows, it's worth a shot. Just don't want to spend 80$ or more on an OT for no reason (though if I go with the 1650N that I so dearly love, I'll surely find something else to do with it if it doesn't work! I've had a friend asking many questions about my 50W Plexi clone!)

Thanks Bruce

[CSB]

Don't let it haunt you, just "build it, and they will come."

I've been running different types of output tubes in parallel for about 5 years now. Just use a 40 - 50 Watt, with an impedance somewhere around 4k. I also run the B+ up around 380V DC. No problems with the EL84 tubes, etc... Thats about where my Maz 18 runs it's B+.

I have EL84's in parallel with octals. I have tried it with 6V6's, 5881's, EL34's, and 6L6's. At the time, 5881'a and EL84's in parallel sounded the best to my ears. Heater current was typically a limiting factor for me when trying different tubes, so I stick with 6V6's most of the time.

Below is one of my first posts (at 18watt.com) on the amp. Its went through significant mods, but the basic principle remains the same.

18 Watt Community Center - Access Denied


CSB

[Hardtailed]

Don't let it haunt you, just "build it, and they will come."

I've been running different types of output tubes in parallel for about 5 years now. Just use a 40 - 50 Watt, with an impedance somewhere around 4k. I also run the B+ up around 380V DC. No problems with the EL84 tubes, etc... Thats about where my Maz 18 runs it's B+.

I have EL84's in parallel with octals. I have tried it with 6V6's, 5881's, EL34's, and 6L6's. At the time, 5881'a and EL84's in parallel sounded the best to my ears. Heater current was typically a limiting factor for me when trying different tubes, so I stick with 6V6's most of the time.

Below is one of my first posts (at 18watt.com) on the amp. Its went through significant mods, but the basic principle remains the same.

18 Watt Community Center - Access Denied


CSB

Most interesting read, thanks a lot!
It reassures me that someone has done it already... for two reasons: 1- it means it works, 2- it means I'm not crazy!

I see you started with a "Ted" 6M36 kit, so I'm guessing your OT is 4K?

Regarding the B+, I've read numerous times that even though EL84 can take a beating, there's a certain sweet spot around 300-310V which is the secret behind the Marshall 18W's sound compared to the SS rectified 2061.
The project is really about getting those EL84 to overdrive while the big bottles puts out a nice firm low-end, which is what you describe in your post. So it's really EL84 first, so if I can't float them 70V above ground, then I'm going with a 320V B+ (minus about 10V of negative bias).

The problem though is finding a suitable transformer to get the proper voltage yet enough current handling to feed the 2 6L6s...

If we look at the "good old" Hammond line (hey, I love the Hammonds in my Marshall clone, and they're cheap since I live in Canada , you need to go with at least 275-0-275 to get anything over 100mA, which gives you about 385V DC once rectified.
I guess I could tube rectify it down to about 350V which would be a good middle ground... but floating the EL84s seems like the more tempting solution.

Let's say a 270HX: 385V rectified, 200mA, 3A on the 5V line if I want to tube rectify, and 6A on the 6.3V line, so I can heat all those tubes easily (4 12AX7, 2 EL84, 2 6L6, for a total of 4.5A). Seems like a good match to me!

But can't find those specs with lower voltage...
Maybe their AC30 replacement would work but I can't find the exact specs

Regarding primary impedance, I'm trying to stick with the 1750 serie of OTs since they are designed for guitar amps, not HI-FI like the 1650 serie.
I really do want/need at least 8 and 16ohms output options.
Like I said, I really like the 1750N in my Marshall clone, I'm wondering if 3.2K is too low... if I want 4K, there's always the AC30 one, but I can't figure out why it's so much more expensive and I'm worried that it might run out of headroom with the 6L6s in there.
There's always the 50W 1750G... but again, why is it 50% more expensive than the 1750N, I have no clue.

I'll have to try and draw load lines to figure this one out. Just wondering how it works out with 2 different types of output tubes (like I said, something in me tells me they won't be "seeing" the same reflected impedance by way of drawing different current)

[Hardtailed]

Just occured to me that I could also elevate the ground connection of the EL84s using a Zener
As long as their grid is connected above the Zener.

I'm guessing that would yield a more stable and predictable voltage drop.

[black_labb]

Just occured to me that I could also elevate the ground connection of the EL84s using a Zener
As long as their grid is connected above the Zener.

I'm guessing that would yield a more stable and predictable voltage drop.

i was going to suggest that.

you could also create a voltage divider at the grids to the el84. where the usual 1m resistor to ground (or to zener in this case) you could put a 500k resistor, and a 500k pot. attach the signal in to one end of the pot, the grid of the el84 to the wiper and the other end of the pot to the 500k resistor. you would have to use a dual ganged pot if the amp is to be a push/pull for both sides of the signal.

i guess it would work as a master volume for the el84's. you can set it to alot of overdrive for the el84's, or less. i can imagine if you pushed it to a level where the 6l6's were overdriving a bit then the el84 would be pushed very hard and be closer to a square wave. by adding the mv in you could have a very versatile amp without changing what you designed it for.

[theunrulychef]

Most interesting read, thanks a lot!
It reassures me that someone has done it already... for two reasons: 1- it means it works, 2- it means I'm not crazy!

I see you started with a "Ted" 6M36 kit, so I'm guessing your OT is 4K?

Regarding the B+, I've read numerous times that even though EL84 can take a beating, there's a certain sweet spot around 300-310V which is the secret behind the Marshall 18W's sound compared to the SS rectified 2061.
The project is really about getting those EL84 to overdrive while the big bottles puts out a nice firm low-end, which is what you describe in your post. So it's really EL84 first, so if I can't float them 70V above ground, then I'm going with a 320V B+ (minus about 10V of negative bias).

The problem though is finding a suitable transformer to get the proper voltage yet enough current handling to feed the 2 6L6s...

If we look at the "good old" Hammond line (hey, I love the Hammonds in my Marshall clone, and they're cheap since I live in Canada , you need to go with at least 275-0-275 to get anything over 100mA, which gives you about 385V DC once rectified.
I guess I could tube rectify it down to about 350V which would be a good middle ground... but floating the EL84s seems like the more tempting solution.

Let's say a 270HX: 385V rectified, 200mA, 3A on the 5V line if I want to tube rectify, and 6A on the 6.3V line, so I can heat all those tubes easily (4 12AX7, 2 EL84, 2 6L6, for a total of 4.5A). Seems like a good match to me!

But can't find those specs with lower voltage...
Maybe their AC30 replacement would work but I can't find the exact specs

Regarding primary impedance, I'm trying to stick with the 1750 serie of OTs since they are designed for guitar amps, not HI-FI like the 1650 serie.
I really do want/need at least 8 and 16ohms output options.
Like I said, I really like the 1750N in my Marshall clone, I'm wondering if 3.2K is too low... if I want 4K, there's always the AC30 one, but I can't figure out why it's so much more expensive and I'm worried that it might run out of headroom with the 6L6s in there.
There's always the 50W 1750G... but again, why is it 50% more expensive than the 1750N, I have no clue.

I'll have to try and draw load lines to figure this one out. Just wondering how it works out with 2 different types of output tubes (like I said, something in me tells me they won't be "seeing" the same reflected impedance by way of drawing different current)

If you're not adverse to imported iron, Weber might have a good PT for this - for his AC30 DC30 kits. Specs here. Interesting looking project!

[Hardtailed]

i was going to suggest that.

you could also create a voltage divider at the grids to the el84. where the usual 1m resistor to ground (or to zener in this case) you could put a 500k resistor, and a 500k pot. attach the signal in to one end of the pot, the grid of the el84 to the wiper and the other end of the pot to the 500k resistor. you would have to use a dual ganged pot if the amp is to be a push/pull for both sides of the signal.

i guess it would work as a master volume for the el84's. you can set it to alot of overdrive for the el84's, or less. i can imagine if you pushed it to a level where the 6l6's were overdriving a bit then the el84 would be pushed very hard and be closer to a square wave. by adding the mv in you could have a very versatile amp without changing what you designed it for.

Yes, I was definitely going to put a "PPIMV" on the EL84, as well as a switch to turn off the 6L6 side. This would allow to use the amp like a classic Marshall 18W or similar. It is not the main point of the amplifier (power supply will probably be too stiff when used in this low-power mode), but it's just a few extra parts to have the option!

And, since this is for me, why spend time deciding on a hardwired EL84/6L6 balance when I can just add a pot for that and go with the mood!

[Hardtailed]

I've started working on the schematic, see in attachment.

I may have the Zener polarity reverse, I'm always confused with those!

Preamp will be a 3 channel affair (I'm basically building myself the most complete gigging amp ever!). With the clean channel based on a Fender blackface architecture, the crunch channel on a JCM800 and the lead channel is my own interpretation of a Mesa Mark II with much less filtering and less gain stages. Total 4 preamp tubes, with all channels sharing the same input stage.

Any input is welcome... seems to me that power amp configuration CAN work.