I tried those Autoformers in my tweeter highpass and I am more than pleased with the result. They are on sale now so might be a good opportunity - I have ordered 2 more pairs already.
They eat up 2-3db more top end than expected in the 10-20k range, which must be taken into account when adapting the filter. I could not measure any distortion at 2 amps and the sound is indeed very mellow - the perfect match for my titanium tweeters.
They eat up 2-3db more top end than expected in the 10-20k range, which must be taken into account when adapting the filter. I could not measure any distortion at 2 amps and the sound is indeed very mellow - the perfect match for my titanium tweeters.
Actually the HF roll off might not have been caused by the autoformers but by my speaker cables´ inductance combined with the lower impedance of my redesigned high pass. I will measure the frequency response of the autoformers into a resistor when my new pairs arrive or curiosity gets the better of me and I rip one out. I might have finally found a reason to upgrade my speaker cables and or rearrange my setup to make them shorter 
Of course I have made a measurement. The load is a 6 Ohms power resistor that has some inductance so I substracted its influence in the graph. I also built a filter with higher impedance and simulated equal response to see how much loss comes from the cables etc.
At 20 kHz the autoformer has an additional loss of about -2.3 dB and my cables add another -1dB. Compared to the resistive filter the on axis response has a big bump now yet it sounds smoother, which baffles me a bit. Glad I did not see those graphs beforehand and just tried them out.
At 20 kHz the autoformer has an additional loss of about -2.3 dB and my cables add another -1dB. Compared to the resistive filter the on axis response has a big bump now yet it sounds smoother, which baffles me a bit. Glad I did not see those graphs beforehand and just tried them out.
Attachments
I tried another tap to inject the voltage (inductance is still above 30mH) and measured again with a different amp, which improved the corrected response quite a bit to -0,8dB at 20kHz. With a notch filter I managed to flatten the response up to 19kHz, which sounded flat and lifeless. The former definitely makes it harder to compensate for my Twetters´natural HF roll off and I decided to live with -3db @20 kHz. So do they sound so sweet because or despite of the HF roll off issues? I don´t know....yet 
