2W SE EL34 amp in simulation

[tmbg]

I spent the weekend polishing my ltspice chops, as well as my triode stage design skills. I only sorta know what I'm doing here, so criticisms and suggestions are welcome :)

I have this goofball 360VCT/10.9V transformer kicking around, which gives me 260V when fullwave rectified, and I decided to see what I could manage to make with stuff I have on hand.

I first plotted curves with spice, assumed a 5k james OPT, and plotted a loadline:



The bias is pretty low compared to Pdmax, but with the limited B+ voltage I have available, hotter bias doesn't buy me much in terms of output swing. This loadline suggests 280Vpp swing for 40Vpp input, which is 100Vrms for 14Vrms drive. 100Vrms into 5k is 2W.


Next, I needed a driver stage. I figured I'd shoot for full output with 1Vrms of input, so I'd need a gain of 14 in the driver. 6SN7 with its µ of 20 would likely be a good choice, but I have 6SL7s on hand, which have a µ of 70. Gain is µ * ra / (RL + ra), and ra I estimated at the operating point by crudely drawing a tangent and figured around 40k, and datasheet says typically 44k. Using 44k and a gain of 15, I calculated RL = 9.42k. With a 10K plate load, I plotted 6SL7 curves and the (very steep) load line:



From there it was fairly simple to pick an operating point and bias appropriately, and then it was just a matter of cobbling together the circuit and doublechecking the transient analysis:



Gains are right where I expected, and the output power is about exactly 2Wrms.

2nd and 3rd harmonics are higher relative to the fundamental than I've seen in other circuits I've simulated, but there's almost nothing above 3rd, so I imagine it'll sound reasonably musical:



AC analysis seems almost suspiciously too flat; I'm wondering if these models don't model miller capacitance. Anyone know about that?



Anyway, I hope to actually breadboard this and see how it performs. That would be the first full amp design that I conceived with a purpose, clean-sheet designed, and executed to the point where it makes noise. I don't expect it to be particularly good, but I'm hoping it isn't wretched :)

[Shannon Parks]

Most of the models need real tweaking. Heck, even all the 2SK170BL models on the internet are garbage and need modifying. I would just add the parasitic cap values as discrete ones on your schematic from grid to plate if you don't see the corresponding value in the model. Also add some parasitic caps to your transformer model. If you get that dialed in along with your inductance/ratio, you'll be "good enough".

Shannon

[rock4016]

Since you have a low B+ have you considered using the 10.9V tap on your transformer to bias your output tube? You could use a doubler and get -25 or more volts. This would give you more power and possibly better bass and usually a little less distortion. You could then change your 6SL7 plate load resistor to 47K and get more gain and probably lower distortion from that as well, since it won't be loaded down so much.