sealed enclosure + SET damping (help!)

[keto]

Hey all,

here are measurements of a midbass of fz=34.7, qts=0.33, vas=312L in a 2 ft3 box, fullrange (purple) and then with the coupling and parafeed caps in a SEX amp tuned to around 60hz (green).

It looks like with the addition of the two low-passes, the -6dB point is around 55-60hz (success!), but the magical transformation of the q=1 2nd-order to q=0.5 4th-order obviously needs work. Too lumpy.

The midbass (ae acoustics TD15M apollo) is reported to have a very flat impedance, but I am also willing to add a cap at the input of the SEX amp and not rely on the parafeed cap for the tuning of the circuit. But so far, that aspect of the filter doesn't seem to be the culprit, but that...

The enclosure is too small. To accommodate other possible x-factors (like driver specs and real-life damping-factors being different), I think a 4 ft3 box would do the trick. I could also go to, say, 5.5 ft3. But that would seem like overkill. Any thoughts?

thanks!

--keto

[Paul Joppa]

The parafeed cap will interact with the bass resonance pretty strongly, that's probably why the apparent Q is higher than expected. When speaker makers say "very flat impedance" they are almost never referring to the bass resonance, but rather to controlling the voice coil inductance effects at high frequencies.

Fortunately, placing that second pole before the input keeps more bass out of the driver and output stages.

[keto]

Thanks a lot Paul.

I used .022uF on the 100K input and .01uF on the 215K (new) output grid leak, both of which tune to about 72hz, and reinstalled the 1.0uF on the output.

Wow, major step in the right direction. Looks pretty much like a 4th-order roll-off and a -6dB @ about 60hz.

Here's a measured curve for the driver. As you said, that resonance was very present.

http://www.aespeakers.com/phpbb2/viewtopic.php?f=15&t=2596&p=11448&hilit=impedance+curve#p11448