Double Stack ESS AMT-1 with Wings--Possible Kit for Heritage

[Pete H]

Hate it for you Rudy.  We've all had some oh sh!t moments and painful is pretty accurate.  

[Thaddeus Smith]

25 minutes ago, Rudy81 said:

Word to the wise....be VERY, VERY careful when working with the AMT wiring!

 

I just learned an expensive lesson, all self inflicted.  I am working a new prototype and was using one of my AMTs to take measurements and setup the new prototype.  After my initial measurements, I re-assembled the stack and set it aside.  Later, I needed the AMT again for setting up my lumber pieces and I FORGOT that I had re-assembled the stack and attached the wires to the power strip.  I pulled the AMT and one of the wires ripped out of the diaphragm.

 

I called ESS and I now realize it is impossible to solder or FIX what I did.  What a bummer.

 

I just put in another order for a backup pair since I just trashed one.  Painful!

 

oh dear. mine are staying packaged up in the box until I've got everything built and purchased, ready to go.

 

edit: do you need one of mine in the short term?

[Rudy81]

1 hour ago, Thaddeus Smith said:

 

oh dear. mine are staying packaged up in the box until I've got everything built and purchased, ready to go.

 

edit: do you need one of mine in the short term?

 

Thank you sir, but no.  I now have the perfect 'model" while I wait for a new diaphragm to arrive and two extra drivers in case of future issues.

 

Apparently, the wires are soldered with  some very precise and exotic process.  Soldering irons are totally out.  I'm sure if I even got close to the diaphragm with a hot iron, the plastic itself would melt instantly. 

[Rudy81]

Besides trashing one of my AMTs, I did manage to get some work done.  Here is the latest contraption.  I am including measurements and drawings in case anyone is interested. Should I get promising SPL and xover point results, I will finish the rounded corners and go from there.

 

 

 

 

[Thaddeus Smith]

Man, that looks awesome. Looking forward to the measurements and listening impressions.

[Rudy81]

1 hour ago, Thaddeus Smith said:

Man, that looks awesome. Looking forward to the measurements and listening impressions.

 

I hope to just use what you see above to take my measurements.  I don't think rounded edges and a top would make much, if any, difference in the response. 

 

Overall, the base of the structure will be roughly 24"x24"......which is a reasonable size and can be used on top of most speakers we are dealing with.  I will post sweeps ASAP.

[Thaddeus Smith]

1 minute ago, Rudy81 said:

Overall, the base of the structure will be roughly 24"x24"..

 

That's basically the dimensions of my LS bins, so that would be awesomely stout.

[Rudy81]

Ok, here are the results comparing the AMT stack in my current 7" asymmetrical wing setup vs. the prototype 12" symmetrical wing setup. These were taken with no xover for just the stack, no LF was on at the time.  Sweep was run from 400hz to 20khz so as not to push the driver too much....I'm already down one diaphragm.

 

I will leave things hooked up tonight in case anyone wants some other measurement. Please let me know what you think.  The mic was never moved between reading and I did the best I could to keep the driver position consistent. The brown line is the 12" symmetrical prototype results. If @Chris A could interpret, that would be a bonus.

 

 

 

[Rudy81]

Based on my previous tests with varying wing sizes and these results, I believe that for me, the smart play is to keep what I have.  If anything, maybe add symmetrical wings to the 7" setup. The results above show just a couple of db improvement in the lowest range, which in my case is not necessary.  One of the down sides, is the new prototype size. It is rather large and poses some placement challenges for my center channel. The other down side is the slight loss of performance between 1.2khz and 2.2khz.  All in all, it is a game of trade-offs and the added work, cost, size and weight don't make the very marginal gain worthwhile for me.  That is, unless I missed something.

 

[Thaddeus Smith]

5 minutes ago, Rudy81 said:

Based on my previous tests with varying wing sizes and these results, I believe that for me, the smart play is to keep what I have.  If anything, maybe add symmetrical wings to the 7" setup. The results above show just a couple of db improvement in the lowest range, which in my case is not necessary.  One of the down sides, is the new prototype size. It is rather large and poses some placement challenges for my center channel. The other down side is the slight loss of performance between 1.2khz and 2.2khz.  All in all, it is a game of trade-offs and the added work, cost, size and weight don't make the very marginal gain worthwhile for me.  That is, unless I missed something.

 

 

Unless Chris sees something that makes this not suitable, keep this one and I'll compensate you for it and provide materials to create the twin and finish them for production use.

[Chris A]

This is interesting.  I see some things between the 7" wings and 12" wings that may be difficult to trace (first order effects, that is) to wing size but nevertheless are there:

1. I see no real change in the spectrograms above 6 kHz--and that's expected.
2. Around 3 and 5 kHz, there are a few changes in the duration of the spectrogram pulse length with the longer wings, somewhat beneficial with the longer wings (i.e., shorter impulse durations) at 5 kHz and a little shorter duration with the shorter wings at 3 kHz.  This is a bit surprising in that there should be no differences due to any wing lengths above 1500 Hz (the 1/4 wavelength point for the basic AMT-1 body without wings). I gather that there are some three dimensional effects occurring in the horizontal+vertical loading of the ribbon that are not entirely related to 1/4 wavelength support otherwise directly attributable to the wings.  Perhaps the higher output of the 12" wing version at 3 kHz explains this effect.  Perhaps some of this is due to the window lengths used by REW to calculate the SPL response curves, and the duration of the relative impulse responses at those frequencies.  I'd have to play with the time windowing (if that's even possible within REW in the SPL response view) to see what those effects might be. 
3. The changes in duration of the impulse response at 2 kHz look like an improvement of almost 50% reduction in the impulse response duration with 12" wings over 7" wings.  This is spectacular but not easy to explain.  Perhaps the lower SPL response of the 12" wing version at 2 kHz explains it all.
4. By the 1 kHz point, there seems to be no differences in the duration of the spectrogram between the two different wing sizes, other than the fact that the 12" wing version has 2 dB more output down to 600 Hz.  This is actually easier to explain: there really shouldn't be any real changes due to the wing lengths until the frequency gets down to 500 Hz or below.  So some of the differences occurring above 1 kHz frequency must be related to 1/2 wavelength or perhaps even full wavelength effects (instead of 1/4 wavelength effects within the horns).
5. The extension of bottom end frequency response isn't really strongly seen below 550 Hz or so using the longer wings.  This means to me that the ribbon is getting about all the lower frequency loading that it's going to get (the resistive part of the horn loading to the ribbon itself).  But it's clear that the SPL response is 2+ dB higher with the longer wings at 700 Hz. This indicates that the wings are primarily controlling the polar coverage better at lower frequencies rather than providing improved acoustic loading to the ribbon itself (i.e., the acoustic transformer effect on the ribbon).  So how important is better polar control at 700 Hz? It depends on whether or not you're crossing at 700 Hz or below.

Chris

[Chris A]

At this point, I realize that some of my observations above are getting into the weeds a bit, but I think it's important to note what the experiment was meant to show.  In my view, the real issues are:

a) supporting ribbon acoustic loading better at lower frequencies using longer wings, and

b) improved polar control.

 

Since all we're seeing here are on-axis measurements, it's not clear how much the longer wings are improving polar coverage at lower frequencies (and that takes many more measurements at off-axis angles to show that), but the evidence of the SPL response curve differences seen at lower frequencies with longer wings indicates to me that the polar support (point "b" above) is the major effect that's occurring.  And that's good for those folks considering using longer wings for supporting lower crossover frequencies than the ~800 Hz that Rudy has chosen for his crossover to his bass bins. 

 

In the case of the Khorn--the most stressing case for Klipsch Heritage applications--the desired crossover frequency is 400 Hz or lower.  The reality is that the current dipole radiating AMT configuration really supports crossing over at 500-550 Hz.  That's not a show stopper for Khorn owners, but it is a constraint that wasn't described in the original Hawthorn Audio Rainier AMT (the monopole-radiating horn loaded AMT). 

 

But the stacked/winged AMT assemblies still are a better deal than the typical Khorn midrange and tweeter stock K-400/K-55 and K-77 assemblies in terms of listenability (subjective performance), apparent depth of soundstage (dipole radiating midrange/tweeter), inherent time alignment, and polar control consistency.  Vertical polar coverage in the stock Khorn midrange and tweeter is better than the AMT assembly, but there is no lobing in the AMT assembly in the vertical or horizontal directions, unlike the separate K-400 and K-77 horn mouths. 

 

This means that crossing at 550 Hz with Khorn bass bins using DSP crossovers is still a better deal (in my opinion) than the stock K-400-series midrange horn and K-77 tweeter.  [I've not discussed power handling above, but note that the AMT-1 assemblies can handle 80 watts (stacked AMT-1s), and the K-77 tweeter can handle 2-4 watts.  That's a pretty big difference. You may have higher total SPL capability using the K-400/K-55 driver assemblies, but you've got a lot more SPL output power margin using the stacked/winged AMT assemblies vs. the K-77 tweeters.  Since few Klipsch Heritage owners experience midrange driver failures, but many experience tweeter failures (due to overloads), the AMT-1 assemblies currently look to be a more survivable alternative for "SPL heads".  Two inch compression drivers on K-510 or K-402 horns are even much more survivable than the stock midrange/tweeter and the AMT assemblies, and have even more SPL headroom than either former alternatives, but at much higher monetary cost and at a loss of dipole operation that provides that light, airy feeling. ]

 

I suspect that the Hawthorn Audio reports of being able to cross over at 350-450 Hz might be related to having a full four-sided horn and an acoustic backstop (closed chamber) for the rear side of the AMT ribbon, i.e., theirs wasn't really a good dipole radiator.  Either that, or the reports of being able to cross at that low in frequency were exaggerated a bit, were not considering polar control, and were assuming a bit of EQ boosting was included in those quoted lower crossover frequencies.

 

Chris

[Chris A]

Rudy, it would be nice to see the relative phase responses (from the "overlays" facility) and to zoom in on the group delay plot so that the minor vertical division is 0.5 ms instead of 50 ms. Alternatively, it would be nice to have the .mdat REW measurement files associated with your above plots, via email, and I'd be happy to post the plots and describe those differences.

 

Chris

[Thaddeus Smith]

4 hours ago, Chris A said:

In the case of the Khorn--the most stressing case for Klipsch Heritage applications--the desired crossover frequency is 400 Hz or lower.  The reality is that the current dipole radiating AMT configuration really supports crossing over at 500-550 Hz.  That's not a show stopper for Khorn owners, but it is a constraint that wasn't described in the original Hawthorn Audio Rainier AMT (the monopole-radiating horn loaded AMT).

 

What are you seeing as the recommended crossover point for other heritage speakers in this kit idea?

[Chris A]

It depends on which bass bin you're using.  In the case of La Scalas, there is really no reason to cross as low as 400 Hz, other than you can cross that low with the K-55 drivers and K-400-series horns.  The bass bin is good to over 1 kHz, so you have plenty of choices of crossover frequency with La Scalas. I'd pay attention to the phase curve in the La Scala, and where the phase starts to take off, that's a good frequency to cross over in order to avoid phase mismatches within the crossover interference band from the two crossing horns/drivers.

 

The same is true for Belles, except that as you get above 500 Hz, the truncated mouth of the Belle bass bin begins to present lobing in the output vs. horizontal off-angle (unlike the La Scala bass bin that has a full mouth closure evident by the sharp nose of the dog house). 

 

Cornwall bass bins all need to be crossed near the full wavelength point in order to match polar coverage.  That full wavelength point is up around 900 Hz, so you have something like 700-800 Hz crossover frequencies used in the various Cornwall versions.  This is a function of the active cone portion diameter of the woofer, only.  The same is true of Heresies: their 12 inch woofers would indicate that it actually will begin to match polars with the midrange horn only at 1100 Hz (although the Heresy is crossed at 700 Hz [Heresy II] or 850 Hz [Heresy III].  Matching polar coverage in at least the horizontal direction at the crossover point is one of those requirements that was apparently not recognized until perhaps the 1980s.

 

Chris

[Chris A]

I should point out that the Cornwall and Heresy direct radiating bass bins (like all single woofer bass bins) will begin to lose all directivity below their full wavelength point.  This is one of the big differences between horn-loaded bass bins and direct radiating bass bins: the direct radiators all lost directivity control below their full wavelength point across the woofer, while the horn-loaded ones are typically designed to hold their polars down to ~100 Hz. 

 

The only way to delay that loss of directivity in direct radiating bass bins is by using more woofers arrayed in a flat baffle (like the KPT-942/4 bass bin with four 15" woofers).  Even with four 15" woofers, the delay of loss of directivity control only exists down to ~450 Hz, which in my experience is still too high to avoid the less than desirable effects of directivity loss at  midrange frequencies.

 

Chris

[Rudy81]

Chris, file en route to you.

 

Micahel, I will keep the new mock-up for you.

[dkalsi]

2 hours ago, Chris A said:

The same is true of Heresies: their 12 inch woofers would indicate that it actually will begin to match polars with the midrange horn only at 1100 Hz (although the Heresy is crossed at 700 Hz [Heresy II] or 850 Hz [Heresy III].  

Chris

 

I plan on experimenting with the Eminence Delta Pro 12A + AMT1s in the coming days. I'm nearly done completing a test cabinet and hope to start taking measurements by the weekend. 

 

I know the ESS used the AMTs down to 800hz, but given my distortion measurements, I don't plan on using them less than 1200hz. I'll see how the horizontal polar response if the Eminence mattes the polar response of the AMTs at that 1200hz x-over point. 

 

The Delta PRO should not start beaming until about 1300Hz (where the wavelength = 264mm). The Delta PRO's effective piston diameter is 260mm. So I suspect that above 1300hz, the wavelength will drop below 260mm and the cone geometry will start beaming. 

 

[Chris A]

Thanks, Rudy.

So for a little more data (to help those contemplating going this direction with their Klipsch Heritage models), below you will see a single, non-wing AMT-1 (yellow trace), a double-stacked winged AMT-1 assembly with front and back 7" wing extensions (red trace) and with front and back 12" wing extensions (green trace).  There is no EQ used on any configuration, below):

 

 

As you can see, the on-axis low frequency SPL cutoff improvement is about 100-150 Hz lower with the stacked, winged versions (not a great improvement, but an important one for being able to cross over a little lower to a Klipsch Heritage bass bin, such as a Khorn (most importantly) and La Scala/Belle.  I'd put the natural crossover frequency at 610 Hz for the single non-wing AMT-1, and 450 Hz (with a little 3 dB PEQ boost on the bottom end to flatten response a bit) for either the 7" or 12" stacked/winged versions. 

 

Next up is the group delay plot, which shows the rate of change of phase response, and which the human hearing system can hear values above about 1 ms:

 

 

This is probably the best indicator of perceived sound quality of each alternative. (That local rise at 4 kHz for the green trace is a measurement anomaly, and should be ignored.)

 

In general, what you get with the 12" symmetrical wings is lower group delay, followed by 7" symmetrical wings, and then a single AMT-1 without wings.  Correspondingly, lower group delay means better control of polars of each alternative configuration.

 

Phase response is about the same, with the higher changes seen in phase at lower frequencies (the apparent 360 jump in phase for the green trace is the same measurement anomaly discussed above for the 12" symmetric wing case...and should be ignored):

 

 

Next are the spectrograms.  I've deliberately chosen 1/6 octave smoothing (the same smoothing used for psychoacoustic smoothing at these frequencies) to highlight the differences in the "self-damping" or the narrowness of the spectrogram spikes, which indicates the effects of lower moving mass and correspondingly "faster" sounding response.  Additionally, I've normalized the spectrograms to the peak at each frequency (the vertical scale)

 

First the single, non-wing AMT-1:

 

 

Note the shift to the right of the peak energy time curve below 1 kHz.  Next the double stacked, 7 inch wing extension (symmetric wings) AMT-1 assembly:

 

Note again the peak energy time curve below 1 kHz: there is a strong improvement in the verticality of this curve for this configuration.  Finally, the double stacked and symmetrially winged (12 inch wings) configuration:

 

The last spectrogram show the most vertical peak energy time curve, whose delay to the right in limited to less than 0.6 ms, and whose duration of a delayed tail of response below 1 kHz is minimized among the three configurations shown.  This is important from the standpoint of achieving a flatter phase response when the bass bin is integrated with the AMT-1 assemblies.  This phase flatness is a big factor in the subjective auditory response of the loudspeaker(s).

 

Additionally, you see some broadening of the spectrogram at 3 and 5 kHz, which is really reflecting the local rise in relative SPL at those two frequencies as compared to the 7" stacked/winged configuration and the single non-winged AMT-1.  After EQ is applied to flatten the response (remember that this thread is about using DSP crossovers that have inherent EQ compensation capabilities...and not about trying to use passive crossovers), the spectrogram of the 12" symmetric wing extension configuration will look very much like the 7" wing case.

 

Chris

[Chris A]

1 hour ago, dkalsi said:

I know the ESS used the AMTs down to 800hz, but given my distortion measurements, I don't plan on using them less than 1200hz.

Remember that this thread is really about a double-stack AMT-1 assembly with wing extensions, which significantly decreases the harmonic distortion levels at higher SPL.  That's the subject of this thread.  The single AMT-1 comparisons above are to show the degree of improvement in each category of each measurement types vs. the baseline of one AMT-1 without wings.  If you're worried about harmonic distortion below 2 kHz, then the subject of this thread is stacked and winged AMT-1s to directly address and decrease harmonic distortion at higher SPL (i.e., we're talking about non-linear distortion here), also discussed the first post of this thread. 

 

1 hour ago, dkalsi said:

The Delta PRO should not start beaming until about 1300Hz (wavelength = 264mm)

"Beaming" is in my terminology "directivity" and it is a good thing to have, especially if it can keep the nearfield acoustic energy off the walls, floor and ceiling.  This is perhaps the strongest reason for using horn-loaded bass bins and perhaps the second best reason for using AMT-1s, in my experience. 

 

Also, flatter directivity vs. frequency curves are actually significant measures of merit for the entire loudspeaker configuration, not just having some level of directivity at some particular frequency irrespective of precipitous changes in directivity above and below the frequency of interest. 

 

Chris