Single ended EL34 amp design

[Stephen1966]

also there is detailed vintage telefunken datasheet of EL34 with many output stages defined, just pick up one which you like, also schematics are provided, so no need to think too much for that simple thing

Could you share that datasheet?

I've been painting by numbers for too long

[Stephen1966]

Some more thoughts about the screen grid stopper...

Merln again:

The stopper provides current limiting and compression... Since in reality the screen current will not quite attain this value we should use the next highest standard resistor, which is 470 ohms for a wire-wound, although 1k is more common and will cause the characteristics to compress more.

Greater resistance = more compression. One of the more desirable traits of guitar amps is their ability to articulate the thing we call "touch sensitivity". Dumble's amps are very good in this regard in my experience. Touch sensitivity also manifests itself as a greater dynamic range and - for want of a better word - transparency. Transparency of detail, such that there is a sufficiently audible difference between the uncompressed non-linear tube amp and the compressed linear solid-state amp. Compression can be added later with a good pedal if needed, but the articulation, touch sensitivity, transparency whatever we call it, should be there... to begin with.

A 1k will work to limit the current effectively, no question. But I'm asking myself, isn't it better to optimise this down to the smallest possible value where the tube can "breathe" more easily? We might later play around with different values and for no significant audible difference - the screen grid only plays a part in "filtering" the signal and tubes have a lot of other things going for them after all - but designing this amp from the ground up, from a single tube in fact, gives me the chance to build this dynamic character into its DNA. Current levels are going to dance around all over the place. A smaller value screen grid stopper is just going to allow more breathing space and solve its part in the whole equation.

Edit:

Again, this is mostly an intuitive understanding... What are your thoughts about the effect of compression with the screen grid stopper?

[Stephen1966]

Screen-grid stopper
Rethinking this and bearing in mind that the voltage drop derived from the transfer characteristics chart ranges between 100V and 120V, the reality is that the tube (hopefully) never reaches saturation at 130mA and so, questions of compression aside, it's better to select a value on the conservative side.

Because while:

• R = 100/0.130 = 769 Ohms

and

• R = 120/0.130 = 923 Ohms

if the Ia tops out at 100mA, we get:

• R = 100/0.100 = 1k

and

• R = 120/0.100 = 1k2

The voltage drop can only be estimated from the transfer characteristics chart and may be off by as much as +/-20% but running the amp below saturation is very likely and so a higher value resistor is a safer bet.

Okay 1k

[bepone]

a lot of nice informations can be read from this 2 datasheets
https://frank.pocnet.net/sheets/129/e/EL34.pdf
https://frank.pocnet.net/sheets/128/e/EL34.pdf

[Stephen1966]

Biasing

from operating point above, 350V, 0.065A,

Rk needed for bias =470R 5W-10W, including g2 current in calc.

Ck-from 47uF till 470uF / 63V depending of desired bass content

...

I went over my biasing figures again and taking:

• Vgk = -22V

• Ik = Ia + Ig2 (queiescent readings)

I get:

• Rk = 22/(0.065 + 0.0097) = 294.5 Ohms

Anything above this in the readily available near values would probably be okay though and 470 Ohms is close enough.

The power dissipation required is quite high as you would expect though:

• (Ik^2) x Rk

• (0.065 + 0.0097)^2 x 294.5 = 1.64W

and

• (0.065 + 0.0097)^2 x 470 = 2.6W

So a conservative doubling of those figures, results in a minimum approximation of 5W and 10W is probably not a bad choice.

I got a more agreeable reading when I came to calculate the value of the cathode cap though.

A bit of fudging here but to get a low 10Hz roll off with the approximate calculated value of Rk (295 Ohms) we get:

• Ck = 1/(2 x pi x 10 x 295) = 53.95 uF

Using a near value for Rk at 330 Ohms, we get

• Ck = 1/(2 x pi x 10 x 330) = 48.22 uF

Which is the approximate lower-end value (47uF) you arrived at.

Then, when we substitute in a 470 Ohm resistor, we get:

• Ck = 1/(2 x pi x 10 x 470) = 33.8 uF

Very few instruments demand clarity of the fundamental at such a low frequency however, so we might increase the roll off frequency to 40 Hz, giving us

• Ck = 1/(2 x pi x 40 x 470) = 8.47 uF

We can take it in the other direction though, with a roll off frequency of 5 Hz

• Ck = 1/(2 x pi x 5 x 470) = 67.72 uF

I mean, it's buyer's choice really and no one seems to have a problem using 100uF cap in that position - 'a generic value for single ended amps,' as Merlin describes it.

The voltage rating has to be three times as high as the normal cathode voltage though, so 3 x 22 = 66V. A 63V rated cap would probably be okay, but 100V is a safer choice.

That's as far as I've got and the control grid stopper resistor and grid leak (Rg1) resistor are for another day when I come to integrating this with a preamp design. Your values look reliable though and I will bear them in mind.

The fun really starts now though as I redo the whole thing by playing around with the primary impedance taps on the transformers and crunching numbers based on plugging different impedance speakers into different secondary impedance taps of the output transformer.

This is a trick, I believe that Dumble used:

It's easy to see that a transformer that exhibits a voltage and current ratio of 20 will have an impedance ratio of 400. This means that connecting a speaker of 8 ohms to the secondary will reflect an impedance of 3.2k to the primary. Doubling the speaker impedance to 16 ohms would also double the primary to 6.4k. (This is a useful trick that may allow you to obtain your ideal primary impedance by connecting the transformer to a speaker impedance other than the rated value.)

[Stephen1966]

a lot of nice informations can be read from this 2 datasheets
https://frank.pocnet.net/sheets/129/e/EL34.pdf
https://frank.pocnet.net/sheets/128/e/EL34.pdf

Thank you for those. The Mullard datasheet is the one I've been using. I thought the "vintage" Telefunken datasheet you referred to might have had some extra pages of sample schematic pages but its basically the same as the Mullard datasheet, except in German! I have trouble enough with English so I will stick with the Mullard datasheet but it is now easy enough to compare it with the Telefunken data.

Also, worth linking perhaps, is the Phillips datasheet that I believe Merlin used. Perhaps not as inclusive as the Mullard but Mullard was a subsidiary of Phillips at this time as I understand it, so... straight from the horse's mouth

el34-philips1969.pdf

[Stephen1966]

PS. I should point out that there is an easier way to work out the pentode stage in Merlin's book "Designing Tube Preamps..."

It's only because I am anally retentive that I went the "tedious" way of working out everything from m.

[bepone]

Biasing

I went over my biasing figures again and taking:

• Vgk = -22V

i made mistake bifore , it is not 470R , something between 290-350 ohms will be..

from where did you get -22V?

[bepone]

That aside, drawing up the load lines for this circuit has been a PIA. Is drawing loadlines becoming a lost art?

not really, doing it all days.. 30 min job..
the real PIA job is to recreate curves which are not existing in any datasheet, and you need them.. and ofcourse it is possible

[Stephen1966]

Biasing

I went over my biasing figures again and taking:

• Vgk = -22V

i made mistake bifore , it is not 470R , something between 290-350 ohms will be..

from where did you get -22V?

Vgk = 22V, (-22) it comes from the mutual characteristics chart.

[Stephen1966]

That aside, drawing up the load lines for this circuit has been a PIA. Is drawing loadlines becoming a lost art?

not really, doing it all days.. 30 min job..
the real PIA job is to recreate curves which are not existing in any datasheet, and you need them.. and ofcourse it is possible

I was thinking of using interpolation (blend) tools in Illustrator. Are you going to spill the beans and tell us how you do it?

[bepone]

from where did you get -22V?

Vgk = 22V, (-22) it comes from the mutual characteristics chart.

if this is your graph to follow , then Ug=-26V

Mullard el34.jpg

[Stephen1966]

from where did you get -22V?

Vgk = 22V, (-22) it comes from the mutual characteristics chart.

if this is your graph to follow , then Ug=-26V
Mullard el34.jpg

No the other graph. Blue dot. Va after the screen-grid dropper and screen grid stopper voltages are subtracted.

[bepone]

No the other graph. Blue dot. Va after the screen-grid dropper and screen grid stopper voltages are subtracted.

your graph to follow is PINK graph.. and blue dot on it is ok.. trick in single ended amps is that they "recreate" voltage supply which doesnt exist..because of inductance behaviour during the change. so your 350VDC will have peak at 680VDC.. how this is possible? because inductance of transformer makes the other "not existing" part.

so actually,pink graph is ok.. you are biasing 65mA at 350V for EL34, at -26V, and RK for that operating point is 26V/0.075mA roughly=350 ohms 7-10W..

[bepone]

I was thinking of using interpolation (blend) tools in Illustrator. Are you going to spill the beans and tell us how you do it?

ok i need to find some time