Thanks to you PJs! It's a pleasure to read your advice!
REW - Room EQ Wizard Room Acoustics Software
I tried this free software, before that my soundcard decided to explode . It's easy to use and so good.
This is the graph of my horns + sub (without tweeter ), that's my last measurement, the day after boooom
I buyed from ebay auction this old rca audio generator, all tube circuit, new rca tube. Maybe I should clean with some dry spray cleaner potentiometer and rotary switches, more i use commands, better the sine and square waves are, maybe some oxide inside.
You touch the problem with flame shape
I tried measure amplifier with oscillator on, it disturbs too much signals, after all it's a strong rf emitter, then I decided to unplug kv and regulator power supply, I only kept the 300v anode for the first part of circuit.
With amazement I discovered that I got troubles with this stage. I havn't found the reason why things happen, but this stage has an incredibile gain, with very few turn of imput potentiometer distorsion raise quickly.I got some doubt about potentiometer, I've never seen a pot running so strange. Circuit should be "easy", I will try triode connection instead of pentode, maybe the gain is really too much.
I'm thinking about how to find the right way to determinate which load can be good to drive oscillator, now, even with 3w output pentode, raising the voltage gain seems not touch the volume of sound too much (over the threshold). Maybe I need more current from amplifier than voltage, current that I can find with el34 tube, or maybe trouble is elsewhere
In this 3 picture little curve is imput, 0,5v\square div, than 2v imput. Nearly full closed pot, 5v/square 2ch at output from pentode. Even with signals under volt distorsion raise quickly and volume doesn't change. sorry I don't remember the frequency, something near 500hz maybe.
In first pic that's the maximum "clean" gain, after this gain raise with dist. Different shape for square wave