> the circuit is similar to ...
It does not seem right for the two cathode resistors in the output stage to be so different: 100Ω and 1,500Ω. In general, the mid-point should sit about half-way between B+ and ground. I estimate this plan's midpoint sits around 40V.
> ...but the output tube is now ecc 83, .... which output stages are capable of driving a 2:1 output transformer except the SRPP and the WCF?
Reality check. 2:1 into 600Ω is 2,400Ω at the tube. ECC81/12AT7 has internal resistance of ~10K; ECC83/12AX7 has internal resistance of ~60K. Neither is a good choice for a 2K4 load.
Topology can help make the most of what you have, but no matter how you harness two mice they won't pull your wagon.
Take the 'AT7. The very best the two tubes can do is for one side to turn full off and the other side to turn full on. Fully on, 12AT7 acts a little like a 10K resistor. Assuming B+ is 250V and the tubes idle at half of that, or 125V. Then 125V feeding a 10K tube and a 2K4 load in series can only pull-up or pull-down 125V*(2K4/(10K+2K4))= 125V*0.19= 24V peak. After the 2:1 transformer, only 12V peak, 8.5V RMS. This is +20dBm, but also gross clipping; usable output may be a lot less. I think THD will be large at the customary +18dBm rating of simple mike preamps. Voltage gain is about 15 to the primary, 7.5 to the load. This is low for a tube with Mu=60. It will be straining.
With the 12AX7 it is even worse. And you usually use the 'AX7 for gain, but with such a low load impedance, gain will be poor. Maybe as low as unity.
As a wild-hit guess, find a tube with Rp about 2K or 3K. That will be a not-bad match for ~2K4 load. In the 12A?7 series, 12AU7 is 6K, so you could use two bottles. NYDave has such a plan somewhere except using one of the 12AU7's bigger brothers (12BH7?). 6DJ8 is actually a capable driver at this level, except it wasn't really built to run high current and heat in normal operation, and life at ~125V ~10mA may be short.
I think SRPP sucks. I am out of step with modern fashion, I know. However when it does work well, I think equal results could be had with one triode, and superior results with the two triodes in GC-GP configuration. If you need some gain and must drive LOW impedance loads well, a volt-amp plus a WCF will shame any SRPP, at the cost of one more triode. (And you may find a pentode-triode to replace that '86.)
