TRIMAX OUTPUT TRANSFORMER - what tube?

I just got a huge Trimax output transformer - 5k primary 600 secondary rated at +37dbm and 45ma
I wanted to build a single ended triode class A output stage because these kind of transformers don't come up very often.
What tube do you guys recommend 2A3, 45, 6F6, 6K6, 6V6????

Triode would be cool but I cannot find a triode that will get close to the +37dbm for the rated current of the transformer.

Pentode or beam tetrode in class A seems a better option.

6v6 and 6f6 will get to about the right level for the transformer without sucking up too much current.
6l6 would do it easy but I would have to bias the cathode colder and I think it would move into its less efficient operating conditions.

Having a conversation with myself - thats cool ;D

> 5k primary 600 secondary rated at +37dbm and 45ma ... I cannot find a triode that will get close to the +37dbm for the rated current of the transformer.

+37dBm is 5 Watts.

5K at 45mA is 225V. 225V at 45mA is 10.1 Watts. This would be the "optimum" device condition. You can however apply more voltage and heat.

On the face of it, they expect a Class A power stage with 10.1W input and 5W output, 49% efficiency.

A good pentode will do 40% efficient.

A good triode with efficient loading will do 20%-33%.

Therefore the specs do not directly give an optimum condition.

I'm sure the original application was pentode 6V6, working at 45mA and perhaps 250V-275V supply, 15V-20V bias, 11W-12W dissipation, 36dBm output.

You can however apply more voltage and heat.

Still, +37dBm in 5K is 44.77mA peak AC. The idle DC is 45mA. The spec assumes you can swing 99.5% of the way to zero current. In hollow-state, that's not happening. However I would not obsess about a few mA and a dBm or two.

EL34 triode will make ~~2.3W, 33.6dBm, at 5K load, 45mA, 330V plate-cathode, -27V bias. 6550/KT88 will not be very different. The total 360V supply is in the zone of common "Champ" power transformers with neat filtering. But note that these big bottles need three times the heater power of 6V6. A 6L6 is not radically different, though a somewhat poorer triode, at only twice the heat-power of 6V6.

Without plotting, I bet 2A3 will do the same, except with twice as much bias, drive, and appropriately higher total B+. 300B will also work, handsomely, if you are rich.

6080 is also possible, but will need 140V bias and drive. The strain in the preceding stage will smother the poor linearity of 6080, making a very tough design or a bad amplifier. (And 5 times the heater of a 6V6.)

I don't think '45 is readily available. Likewise a couple '10 at 400V will do +33 or +34dBm in 5K load, but healthy '10s are very-very-very rare.

Note that a triode into a transformer needs much cleaner B+ than a pentode. The PSRR of a transformer coupled power triode is about 3, of a pentode closer to 10. A Champ power supply is not adequate without a C-L-C filter or C-R-C with kilo-uFd caps.

> I cannot find a triode that will get close

WE 300, page 4, 350V, -76 bias, 50mA, 5000 load, 6.2 Watts, -30dB second, -45dB third. Knocking current down 10% would give 5 Watts or +37dBm on the nose, 3% 2nd, <1% 3rd. There is also a 400V 40mA 5K condition with more power at more distortion. If the iron is VERY old, it is possible this is the design target. But 6V6 (or broadcast-marked 6V6) has been popular for a very long time.

Why the heck would you want +37dBm in 600 ohms in this 21st century?

You find many Gates and RCA amps using the same transformer for both 6V6 and 6L6 purposes.  Some with current ratings identical to your description.    6V6 is the only practical choice there.  The others you list will be hard to source in matching pairs; you willl have to buy a pile and match them yourself.  Practical may not matter. 

Thanks PRR thats a great rundown of choices.

Why the heck would you want +37dBm in 600 ohms in this 21st century?

Click to expand...

Well because I have the iron - why else? Hugely inefficient I know but I love the sound. I built a preamp with a 6V6 SE output and for certain applications I personally think it is my best sounding preamp - and I have collected a lot of awesome preamps.
I'd like to learn more about the old big bottles.

6V6 is the only practical choice there.  The others you list will be hard to source in matching pairs

Click to expand...

Its not Push Pull pairs - just SE.

Well, gee, I'm not reading with comprehension!

So, what is your design goal?  To max out the rating?  What will you be doing with this output stage?  That is the real question.  To revisit PPR's statement:

Why the heck would you want +37dBm in 600 ohms in this 21st century?

Click to expand...

Focus on the +37 dBm into 600 ohms part; do you need that level for your purpose? 

> Focus on the +37 dBm into 600 ohms part

54.9V rms. 77.6V peak.

It was designed to run 6V6 pentode

I meant beam tetrode

I've just been playing with  807s 400V or more on the plate 250V the screen 45mA wouldn't be a problem just adjust the Rk.
I'm actually running PSE with an AudioNote 2k5 output,so 5k SE as your spec.

But the real reason I post is that PRR didn't mention partial AC feedback from the plate to g1 in a pentode SE stage.
I'm impressed at this early stage;it's something everybody was doing on SE pentode in the 40s but got lost with ultra-linear.

In my experimental case a 100k resistor between the plate/bottom of the transformer and the input side of the stage decoupling cap.

Worth trying if you go 6V6 or 6L6 SE.

Robert

Today at 05:05:37 am
> It was designed to run 6V6 pentode
Today at 05:06:56 am
> I meant beam tetrode

1) When you write something wrong, you can use the Modify button above your post to fix it. Certainly when you see your slip right away. Going "back in time" to change your words after everybody commented might be rude; it depends.

2) Me, I would argue that the difference between "power pentode" and "beam power tetrode" is as much about patent rights as the actual tubes.

For "small audio Power" a true tetrode is not as good as a triode.

Good audio power needs a negative "supressor zone" in front of the plate. You can get this with distance and geometry, with an actual grid, or with a pair of side-rods. These look very different to the Patent Examiner, but to the electrons, there's no real difference. Some trade-offs are easier with one structure or another, but designers were paid to be clever.... it was cheaper to dink around with metal than with Patent Profits.

So RCA was careful to market the 6L6/6V6 tubes as "beam power tetrodes", to avoid someone else's patent on the "Power Pentode". But the electrons don't read marketing literature.
___________________________________________________________
> PRR didn't mention partial AC feedback from the plate to g1 in a pentode SE stage.

The original question was "triodes".

I got off into triode-connected pentodes because development of 10W-40W triodes mostly stopped when EL3 and 6L6 appeared; but the lower-Mu pentodes make useful power triodes.

Triode action does about the same as your plate-grid resistor, without the lower input impedance.

Yes, this was a popular technique, and I have used it.

Taken too far, the driver stage is unable to fight the NFB to bring the power tube to full roar. So when you want more than a little NFB, you must move the injection point to the driver input. And impedance relationships, and NFB benefits, suggest now taking NFB from secondary.

An additional point is that plate-grid NFB reduced pentode power supply rejection, making it similar to a triode. When we all ran field coil speakers, the FC network was effective power supply filtering; as PM speakers invaded, power supplies got nasty and only a pentode will give a tolerably clean output. This changes again with modern low-cost caps and DIY economics, of course: I got very-clean with plate-grid NFB and a couple 470ufd caps I got for a buck a piece.