Heil Air motion transformer for Cornwalls

[Chris A]

I haven't experimented with wall boundary standoff placements yet.  I'm not sure that I really can since the geometry of placement in-room as surrounds is pretty much set. 

 

The left AMT-1 is set about 2 feet from wooden plantation shutters behind it in an extension of the room via bay window, and it sounds very nice. 

 

The right side Cornwall is touching the wall with an Auralex Sonofiber pad on top of the cabinet-wall boundary covering the wall just behind to absorb backwave reflections-- and it's about 14" away.  This sounds less spacious but still very nice. 

 

I used REW measurements to balance left and right AMTs and CW bass bins.  That's the rationale for adding the Sonofiber pad--from looking at the left and right spectrograms and largely choosing backwave attenuation on the right side.  It is amazing to me that the AMTs do so well that close to the rear reflection boundaries.  If I had an extra 2-4 feet, I'd probably take it...but I don't.

 

Chris

[Khornukopia]

I have a pair of Cornwall IIs that I could set these on and then copy much of your settings, but I have also been wondering about placing the AMT's on top of my tri-amped Klipschorns and increasing the gain level on the tweeter amp. The corner placement seems to be a complication. Do I put a flat or convex wood panel one foot behind the AMT, or make that part of the corner behind the AMT into an anechoic space?

[Chris A]

Actually, you should get an increased backwave reflection from the corners since they form a (squashed) trihedral horn vs. the gain from a flat back wall. And since Khorns are typically placed in corners symmetrically, i.e., 45 degrees from each wall or false corner, you will get a fairly nice backwave reflection from the AMT and perhaps a little more diffusion from the tetrahedral surfaces that should increase the sense of depth of the resulting images. 

 

In any case, it's certainly worth a try on the Khorns.  Except for the three or so PEQs that I showed above (and reproduced again below) if you cross below 2 kHz, all you need to do is match SPL of the AMT with the midrange using a hand-held SPL meter. Note that I haven't had a lot of run-in time on the AMTs, so I reserve the right to update the PEQs below based on simply using them for a little while.

 

<snip>

 

Chris

[Khornukopia]

I will see how that works out, sitting on top center, flush with the front edge. When they get here, I will want to hear them on the Cornwalls first, so I can correlate to all the good info in this thread.

[jwc]

I'm late to this.  In the images earlier, what Passive XO is being used? 

[Chris A]

Okay, so based on my latest measurements today:

 

1) The delay value for the AMT-1 channels is about zero (0), so a passive crossover can be used, but I'd recommend a first order.  I'm using 6 dB/octave highpass and lowpass crossover filters centered at 800 Hz break-point frequencies.  This is good for the passive crossover guys--a simple capacitor and inductor is all you need--no more crossover circuitry is needed, according to what I just measured.

 

 

This of course includes quite a few PEQs to smooth the response to ±2.5 dB (using psychoacoustic smoothing, as shown above).

 

2) I was fairly surprised by the difference between the left channel and right channel measurements, so I'd say that you would probably need to EQ in your own room via your own measurements.  I can certainly post mine here, but I have to say that the variability will probably nullify much of the gain that you might hope for. The good thing is that the frequency response without EQ correction is probably in the ±5-6 dB range, with about ±3 dB of ultra-smoothed central tendency wander of the response.  This is okay, but I would prefer a flatter response.  YMMV.

 

3) If you're measuring the AMT-1s within about 2 feet of the reflective wall behind it, be prepared to see some interesting measurements.  I found that I couldn't use the same EQ on the left channel as the right--and that's really a first for me due to the fact that my listening room is pretty symmetric, except for the bay window on the left.  Apparently, this is enough to significantly affect the output of the AMT-1 as measured by the microphone at 1 m...significantly...

 

4) I'm at a loss to explain the shift in measurements from last week to today.  Part of this is that I took a bit more time today to lay down absorption blankets, etc. so I could get much better phase measurements, but the delay values that I measured last week are not the same today.  I'd put a few hours of listening on the AMT-1s before trying to dial them in.

 

Good luck. 

 

Chris

[Chris A]

By the way, here's a new spectrogram plot for those that want to see the performance with first order filters and zero delay:

 

[jwc]

debate on the passive values.....

 

K33 woofer and Heil.

 

~0.8-1.0 mHy

~25uF

 

?

jc

[Chris A]

I've got XX µF and XX mH (in series for each case) using XSim and the imported impedance measurement of the AMT-1 and K-33 woofer from DATS.

 

An XSim prediction of the output (just now updated with K-33 input impedance):

 

 

 

<snip>

 

and input impedance:

 

 

<snip>

 

Chris

[jwc]

53.6 µF....reason for that?  Educate me here please.

[Chris A]

Probably related to the measuring and putting in the input impedance data for both the AMT-1 and K-33. 

 

See here for the tool used: http://libinst.com/Xsim/XSimSetup.exe

 

That's Bill Waslo's site and freeware tool.  He hangs out at diyAudio a lot.  He was also the author of Liberty Instruments Praxis, SynergyCalc, and Omnimic, etc.  He's a great guy.

 

Chris

 <snip>

 

[jwc]

Is the Heil rated on the site as 4 ohm?  I see where the website says this (2.9) ohm.  Your measurements though don't look like that.  If close to 4 ohm.....I understand your 53.6uF

[Chris A]

I guess I could provide the input impedance import files (.zma) for the AMT-1 and woofer that I used (they're pretty small files):

 

AMT-1.zma

Cornwall Woofer Only direct connect.zma

 

Chris

[Chris A]

Perhaps you didn't catch the input impedance plot on page 3.  I'll help you out:

 

 

I believe that when you input the zma file, it overrides the value that you put in manually in the filter tuning dialogue.

 

Chris

[jwc]

OK.  So both the Crites and Heil are around 4 ohms.

 

 Not sure where I got the 8 ohms from.  Maybe just an assumption.

 

 

[mboxler]

Chris

 

Am I reading the Crites .zma file correctly?  Looks like the driver, at 800 hz, is 8.5 ohms and 38 degree phase (inductive).  

Seems as though a 1.7 mh inductor would be needed, or am I missing something.

 

Mike 

[Chris A]

Well, that's interesting.  Apparently XSim doesn't adjust the impedance to the imported zma file when determining filter values.

 

I reran the Cornwall woofer (K-33) in the Cornwall cabinet (with ports), and here is the impedance curve for it:

 

 

So the woofer's impedance at 800 Hz is 5.9 ohms, and the AMT-1 is 3.5 ohms.  Plugging those values in yields 56.2 µF and 1.15 mH. 

 

 

 

Fortunately since these are first order filters, having values that are not quite exact is probably within acceptable tolerances due to the shallow slopes (6 dB/octave).

 

Chris

[Chris A]

I'm going to delete the figures and text above that show incorrect values so that if anyone reads this, it will cut down on errors being propagated...

 

Chris

[mboxler]

Thanks, Chris!  Still learning, and I could never figure out what .zma files were.  Opened it up in notepad and it became clearer.

 

At first I thought the phase column represented the electrical phase shift across the driver.  But it now appears that it is acoustical phase shift.  Correct???

 

Again, thanks for the info.

 

Mike

 

[Chris A]

The .zma files result from only what DATS is measuring, and that's electrical impedance only.  Moving mass effects (including the coupled air and ports in the Cornwall box for the woofer) and inductance/capacitance of the drivers also are reflected in the measurements, but there are no acoustic measurements in those .zma files.