Single ended EL34 amp design

[bepone]

working a little bit more on this working point..distortion:

Mullard el34.jpg

[bepone]

Pout from the graph will be theoretical 9.2W
Then substracted loss 12-15% in output transformer = cca 8W on the speaker

[Stephen1966]

Just want to show you quickly how we might use illustrator to draw interpolated graph curves. Here are a couple of screenshots using the blend tool.

This is what it looks like before:

Screenshot 2023-05-22 143329.jpg

And this is what it looks like after I went into the blend tool options and specified a distance of 1mm between segments. Setting it for 3mm in this, produces something like the curves corresponding to 25V divisions. You would have to play around with it and you need to start with vector curves but it's easy enough to be useful. These are "ideal" pentode curves after all, and we live in the "real" world

Screenshot 2023-05-22 143603.jpg

The green colour is my choice to highlight the affected curves.

[Stephen1966]

25V divisions:

Screenshot 2023-05-22 145900.jpg

[martin manning]

here are tools for this, such as https://www.vtadiy.com/loadline-calcula ... alculator/

[Stephen1966]

here are tools for this, such as https://www.vtadiy.com/loadline-calcula ... alculator/

Yes, that's great... thank you. Also, this one:

http://www.trioda.com/tools/triode.html

[Stephen1966]

Pout from the graph will be theoretical 9.2W
Then substracted loss 12-15% in output transformer = cca 8W on the speaker

I'm just wondering... that loss... is it a rule of thumb for most output transformers or is it different for single-ended and push-pull transformers?

[Stephen1966]

working a little bit more on this working point..distortion:
Mullard el34.jpg

This is great! Thank you, I hadn't got that far. Second harmonic distortion at 3% would be practically inaudible and so the values calculated would give a relatively clean output section. Not sure I will go with that yet though because I was aiming for about 10W output power. The difference between 8W and 10W is marginal - whole careers have been formed with less

I want to play a bit more with the impedance. The THD of the Class A, SE amps tops out at 10% in the Mullard values but I don't know if those are second or odd order harmonics - I guess I could try a new load line chart with the Mullard values and it might indicate the harmonic content. The Hammond 125 series transformers are interesting because they offer three values of primary impedance. I should run some numbers but I am curious to see what happens (mathematically) when I pair up (for example) the secondary impedance of the OT with a mismatched speaker impedance - running an 8 Ohm speaker off the 16 Ohm tap, for instance.

It's this that might allow me to play with the gradient of the loadline. Either a higher Va or a steeper gradient more towards Class B.

[bepone]

25V divisions:

Screenshot 2023-05-22 145900.jpg

very nice, because in pentodes all curves are possible to extract from this transfer curve (main graph for all)

[bepone]

Pout from the graph will be theoretical 9.2W
Then substracted loss 12-15% in output transformer = cca 8W on the speaker

I'm just wondering... that loss... is it a rule of thumb for most output transformers or is it different for single-ended and push-pull transformers?

the best trafos (big core, thick wires) is possible to make with 7% loss, excellent result for hi-fi
in guitar world, there are thin wires, small cores, so 12-20% of loss is some standard..

[Stephen1966]

I decided I needed a fresh set of curves from the Muller datasheet.

First, with the new curves drawn as vectors over the top:

Mullard el34.jpg

Then, just the new curves:

Mullard el34-copy.jpg

[bepone]

with more thinking, you cannot use this set of curves because your Vg2 will be less than 360V.

So if you want cathode bias amp, class A which draws always the same current, you need at least 400VDC of raw UB+ after rectifier...
then after CRC or CLC first block of rectification, you will have 390V on transformer, 385 on anode, 26 on cathode means Ugk=340V, Ug2k 340V or less

[bepone]

i think is better and more relaxed rectifier wise, to interpolate curves for Ug2= cca 300V

[Stephen1966]

with more thinking, you cannot use this set of curves because your Vg2 will be less than 360V.

So if you want cathode bias amp, class A which draws always the same current, you need at least 400VDC of raw UB+ after rectifier...
then after CRC or CLC first block of rectification, you will have 390V on transformer, 385 on anode, 26 on cathode means Ugk=340V, Ug2k 340V or less

So we need this one?

Mullard el34-copy_Page_07.jpg

I took another look at Merlin's chart and it is the same (Vg2 = 360V). I'm confused, his Vg1 curves follow the Mullard I've been using.

[Stephen1966]

Here's something I can do with vector curves.

Mullard el34-interpolated.jpg

This shows the effect mentioned earlier of moving Vg1 = 0V down to the -8V position and the new -8V curve intersecting the Pd (Wa) = 25W curve at about the same position as the old -16V curve. Not saying the load lines I plotted earlier are those I would use, in fact, I can probably get in the region of 400V off the power transformer and around 400V for the plate or 390V would be easier to work with, further down the line with the preamp.

Could you walk me through what you mean by "more relaxed rectifier wise" with Vg2 around 300V? I'm not sure I understand what you are talking about.