EL34 PP acoustic guitar amp

[Ian444]

I tried your ideas Davorin, I ended up totally removing the 180R's in the CRC filter, changed the 5K4 to 1K1 (2 x 2K2 in parallel) to supply the screens, and fitted 1 ohm cathode resistors. Power is now 35W at the onset of clipping, and 45W with hard clipping. Thank you
new schematic here

There is still significant voltage drop across the 1K screen stoppers, around 80V peak on the scope, I can't remember if that was at 35W or 45W. The screen supply voltage at the 47uF cap after the 1K1 resistor at quiescent is 465, at 35W its 410V and at 45W its 373V. B+ at quiescent is 468, at 35W it drops to 447 and at 45W its 436V. The tubes are biased at 42mA around 19W at quiescent.

I'm wondering how to work out the average screen current, if I used the voltage drop across the 1K1 resistor feeding the 47uF cap that feeds the screens, I get for example at 35W, 447 - 410 = 37V drop across the 1K1, so power dissipated would be 37^2/1100 = 1.25W? Is that right? The 80V peak drop across the 1K screen stopper shows a 6.4W peak but that is for a very short time only. The specs say 8W max.

Does it look like I could do with a stiffer screen supply? I do have one of those C354 chokes used on ST-70's, rated at 1.5H, 55 ohms and 200mADC. Is it worth trying that instead of the 1K1 resistor? I guess I could also call it done and leave it as it is, 35W is probably enough.

This has been an interesting exercise, now I have an insight into why guitar amps are designed the way they are, it just seems odd at first if you come from a hifi background. Two different worlds

 

[davorin]

First replace resistors 2x270K with 2x220K! If You have choke 5-20H/50mA, than replace 1K1 resistor with them.

 

[Ian444]

After a couple days here and there of tweaking, I changed a few things, the unbypassed cathode resistor sounds better than the LED, I swapped the 12AT7 for 12AU7 for less gain, adjusted the anode/cathode resistors for the 12AU7, put the 470K back in from 1st stage 12AU7 to 2nd stage, and reduced the EL34 grid leak resistors to 180K. There is a small amount of crossover distortion at clipping but it sounds quite OK. I tried to alleviate that with zeners limiting the max negative signal excursion on the EL34 grids, this increased the clipping power to 50W but added nothing good sonically so I removed them. The freq response is within 0.5dB from 200Hz to 14kHz. At 100Hz -1.2dB, 80Hz -2.2dB, 50Hz -4.1db, 20kHz -1.2dB. When cranked up it sounds great for blues with electric guitar, even the neighbours like it. "Keep playin that guitar" they say It seems that for a really clean guitar amp you really don't need much gain at all, maybe x 4 gain and straight into the PI would be sufficient (a good job for an IC?). For a bit more drive for blues/rock the 12AT7 is the ticket! But the only distortion stage in this amp is in the cathodyne when it starts squaring off the sine wave due to grid current in the EL34's, and its a very nice tone when cranked up. Thanks for the help! I'm really happy with how it turned out, there were a couple of periods of despair but not too bad and all worth it, I learnt a lot from this one. I am surprised how much screen current is pulled in pulses (with pure sinewave input) with the EL34, which I have not worked with before. I can understand why for hifi one might use a regulated screen supply, especially with true pentodes. Just need to build a box for it now. Next year I'll build another one for myself, with tone controls and a bit more drive.

For completeness, the final schematic is here.

 

[Ian444]

I fitted a different OPT with 3K6 primary from the USA and still get 35W clean, oh well, no worries, I guess 35W clean is what you get from 2 x EL34, and 50W with who knows how much distortion. The USA transformer does sound nice (Magnetic Components). I fixed a minor hum problem and updated the schematic in the previous post. The hum is 25mV with the volume on max and the input shorted. 25mV is acceptable, but with the volume turned to zero, it is around 2mV, so the rest of the amp is quiet, but the input stage not quite so quiet. It would be nice to work out what is the cause of this. Would make a magic amp with only 2mV with max vol.