Jump to content

I finally got around to triamping my K-horns


Greg Oshiro

Recommended Posts

Having owned, modified and tested Klipschorns quite a bit over the years, I’ve come to believe that Klipsch should be offering a prepackaged electronic crossover system and associated equipment to correct for the time delay, ever-present room issues and unique folded bass horn behaviors - and deliver the ultimate Klipschorn experience to EVERY user. Just an opinion, of course. 

  • Like 1
Link to comment
Share on other sites

1 hour ago, BeFuddledinMn said:

Having owned, modified and tested Klipschorns quite a bit over the years, I’ve come to believe that Klipsch should be offering a prepackaged electronic crossover system and associated equipment to correct for the time delay, ever-present room issues and unique folded bass horn behaviors - and deliver the ultimate Klipschorn experience to EVERY user. Just an opinion, of course. 

 

There are a number of reasonably priced DSP units that could fit the bill (you can always pay more of course ....). These would not be "perfect" but they would give a good 80% solution. (and would be affordable for a greater number of users).

 

Unless a person is willing to do some measurements, then they would be pretty much be stuck with settings that others have made using commonly used components. Since the Klipsch (OEM) woofer, mid-horn / driver and tweeter have not changed much over the years, then once those settings were obtained they would be usable by the majority of users.  If folks start using non-OEM components, then the new (appropriate) settings would need to be obtained and used. Contrary to what some others might think, setting this up "by ear" would be a major mistake.

 

BTW, another benefit of doing the measures yourself (and they are not that difficult once you invest some time and do your homework), then you can also toss in some amount of room correction (done sparingly is best)

 

Good luck,

-Tom

 

 

Link to comment
Share on other sites

Well, not satisfied with the response that was achieved with the Shinall (Khorn Clone) performance using the Yamaha SP2060 a few days ago, I started over from scratch with PEQs and delays.  The selected crossover point was revised from 250 Hz to 280 Hz and the channel delays were revamped relative to the bass bin:

 

300249405_ShinallKKS103dBon-AxisSPLandPhaseResponse.thumb.jpg.afa4a21924cf0b3d71b252b38c646d4e.jpg

1518733688_ShinallKKS103dBon-AxisSpectrogram.thumb.jpg.6bf362b201384a25841174dcc2b5b6da.jpg

 

And a few comparison plots using the Danley SH-50 as the comparison (Shinall in yellow trace, Danley SH-50 in green):

 

1980112939_ShinallKKSvs.DanleySH-50PhaseReponse.thumb.jpg.0f7ec9129f3bca95b7bf9f7384e7ae75.jpg1981075898_ShinallKKSvs.DanleySH-50GroupDelay.thumb.jpg.2d88b59c6b9197052e76ea8bbe201545.jpg

1617972403_ShinallKKSvs.DanleySH-50StepReponse.thumb.jpg.f19501fc073da52c2805939a17b2f0d0.jpg

 

These are much more satisfactory performance plots, and the subjective listening performance improvement also coincides with the improvements shown in the plots.

 

Chris

Link to comment
Share on other sites

On 1/14/2020 at 11:10 AM, Chris A said:

Welcome (assuming you're new to the forum).

I just realized who I was talking to (Mark S). 

 

Sorry about that.

 

Chris

Link to comment
Share on other sites

By the way, in the past I was somewhat amazed at the narrow band that the midrange can effectively cross over to the tweeter in Klipsch Heritage models, and that the bass bin overlap with the midrange is actually larger (on axis, without considering phase response) than has been reported.  Below you will see the overall SPL response of the Shinall after extensive work, and below that you will see the individual driver/horn overlaps for the bass bin, midrange, and tweeter with flattened SPL for each driver/horn.  No crossovers used for each driver/horn in the lower plot:

 

489616061_ShinalloverallSPLandDriverResponses(NoXovers).thumb.jpg.b1f9d3ca371a4da4c507afb2c2587e3a.jpg

 

Notice the fairly narrow overlap band of the tweeter and midrange - and the actual crossover point is at 5 kHz, not 4 kHz as can be seen as a more likely crossover frequency.  This is due to the phase response of the tweeter, which takes off in the wrong direction (lagging phase) below 5 kHz. 

 

Additionally, I lowered the crossover point of the bass bin to midrange from the nominal 400 Hz to 280 Hz.  The reason for this is not only the complex polar coverage of the bass bin above 300 Hz, but also the phase response of the bass bin inexplicably takes off, again in the wrong direction--as leading phase above 300 Hz. 

 

1537231330_ShinalloverallPhaseandDriverResponses(NoXovers).thumb.jpg.e3561075223abf608ae80014fb467bc1.jpg

 

Both of the phase response issues serve to set the crossover frequencies in the Khorn (although a clone is being used here, it's a fairly accurate one).

 

Just my observations on the results of this exercise and the lessons learned along the way.

 

Chris

  • Like 2
Link to comment
Share on other sites

2 hours ago, Chris A said:

... the individual driver/horn overlaps for the bass bin, midrange, and tweeter with flattened SPL for each driver/horn.  No crossovers used for each driver/horn in the lower plot:

 

To clarify:  you drove each full-spectrum and flattened each's response within their operating range only?  Or (also) effected crossovers by diminishing out-of-band response of each driver - in/for that plot?

Link to comment
Share on other sites

7 hours ago, Chris A said:

Your first choice, not the second.

 

Chris didn't you employ the Yamaha 2060 as an active XO to manipulate and maximize each transducer? I mean each transducer is not driven full range in the end but per your settings within the 2060? precise HP and LP filters employed, equalized (DSP corrections), and corrected for group delay? 

 

Also, your all SPL graph above, is this with any smoothing? Specifically the orange trace.

Really nice work with the 2060, the overall response is quite flat. 

Interesting the graph is taking a nosedive @ 40hz. Down ~ 9 db @  33 hz.

 

Audio life within the digital domain.......there is no going back now.

Link to comment
Share on other sites

Part of the process that I use (and many others) to dial in the response using a DSP crossover is to:

  1. first look at the raw response of each driver/horn individually, then
  2. flatten the response of each driver within each driver's usable passband (what you see above) using REW to help optimize the PEQs, then
  3. look at leveling the channel gains for each driver, then
  4. decide where to put the crossover frequency for each pair of drivers/horns being crossed by plotting all the individual driver responses on the same plot, like you see above.
  5. At this point, the decision can be made to use a simple lower order or higher order crossover filter for each channel (low pass on the lower frequency drivers, high pass on the higher). 

In my main rig (Jubilees, K-402-MEH, etc.), I'm using PEQs to do the crossover filter duties by using the natural points of crossover responses of the paired drivers/horns to tell me where to put the crossover frequencies (i.e., "fractional order crossovers" because the resulting slopes at the crossovers points are usually between second and fourth order slopes of 12 to 24 dB/octave).  Using PEQs instead of regular crossover filters eliminates the phase shifts associated with those electrical filters and yields a much flatter phase response overall. 

 

For the example above (Shinall KKS Khorn clone), I used first order filters at both the crossover frequencies due to the nature of the natural response of the drivers/horns used, which have narrow frequency bands of opportunity to use without SPL or phase issues interfering with the desired combined response, while still retaining very low phase growth across the entire operating range, as can be seen above.

 

The point that @glens was trying to clear up was whether the plot above reflected the end of step 4 or the end of step 5.  So the plot above reflects the step 4 state of the process--i.e., natural response of each driver, but its response is initially flattened using PEQs within the DSP crossover. 

 

There are more steps to completing the process started above, namely setting the relative channel delays in order to minimize phase and group delay growth issues and to get smooth frequency response through the crossover interference bands.  This is usually an iterative process and sometimes this step can take a little while while looking at SPL and phase response while changing the relative channel delays. 

 

After a satisfactory crossing of drivers using the right channel delays for the loudspeaker under test, the overall smoothness of the SPL response of the loudspeaker can be dialed-in using the DSP crossover input and output PEQ filters to fine tune the flatness of response. The result of that exercise is shown in this post above.

 

By using the steps that I describe above, very good results can be expected using DSP crossovers for even the most challenging loudspeakers, such as Khorns.

 

Chris

 

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

14 minutes ago, Chris A said:

By using the steps that I describe above, very good results can be expected using DSP crossovers for even the most challenging loudspeakers, such as Khorns.

 

Chris

 

 

Nice work. Now did you employ any smoothing in those last two graphs you posted above?

Link to comment
Share on other sites

The smoothing used with any plot within REW can be seen in the legend, usually below the plot area. I use psychoacoustic smoothing (Psy) for my plots, generally, because this most closely follows how the human hearing system perceives the result.  This is the best way to determine the right balance of the number of PEQs used--neither too few nor too many.  Other methods of determining how many PEQs to use typically leave you with either a loudspeaker that's still tonally unbalanced or over compensated, trying to correct for room modes.

 

This process above is for using IIR filters used in most DSP crossovers.  For the ultimate in SPL and phase response flatness, after the above steps are successfully accomplished, FIR filters can be applied to further flatten the phase response of the loudspeaker.  This usually takes much higher computing resources than the typical IIR filters found in DSP crossovers, but yields loudspeaker performance that's visually and sonically "perfect" (from an SPL and phase response viewpoint). 

 

While my activities with dialing in loudspeaker using DSP crossovers is focused on IIR filtering, I find that the performance attained using only IIR filters to achieve "minimum phase" performance (i.e., essentially zero all-pass phase growth of the loudspeaker is achieved--i.e., the excess phase curve is almost coincident with the overall phase curve) is really good--and that's usually "good enough". 

 

At some point, I plan to cross over into using FIR filters to correct phase further (perhaps even the bass response, which requires very many FIR filter "taps" per loudspeaker, implying greater delays overall) but for now, IIR filtering gets me far enough.

 

Chris

  • Like 1
Link to comment
Share on other sites

Got it, thx. I have been experimenting with REW on my two systems, learning the software and it's potential. There is a bit of a learning curve for sure and I have not tried the Psy smoothing yet. But I shall in time. My base cornwall 3 performance graphs from day one testing below.

 

In stereo mode, mic @ my listening position:

 

cw full range n stereo at my head.jpg

 

 

waterfall display is right speaker alone @ 1 meter:

 

30-400hz waterfall.jpg

  • Like 1
Link to comment
Share on other sites

17 hours ago, Chris A said:

...Additionally, I lowered the crossover point of the bass bin to midrange from the nominal 400 Hz to 280 Hz.  The reason for this is not only the complex polar coverage of the bass bin above 300 Hz, but also the phase response of the bass bin inexplicably takes off, again in the wrong direction--as leading phase above 300 Hz. 

 

Chris

Was this taking off of the phase response above 300 Hz not seen last year when we married the K402 to the Khorn bass bin and crossed around 450 Hz?

 

Link to comment
Share on other sites

Yes.  But since your bass bin phase was taking off at 210 Hz (a bit too low for the K-402 with BMS 4592ND), it was better just to deal with the phase growth at that point and cross at 450 Hz, about at the same phase growth as the stock Khorn, but without the extra 360 degrees of phase growth due to time misalignment in the passive crossover Khorn bass bin (about 6' long) to midrange horn (about 28" long) :

 

77853840_MarkHKhornBassBinSPLandPhase.thumb.jpg.3f1abf7d21900334fdb4937b5222c751.jpg

 

Chris

  • Like 1
Link to comment
Share on other sites

Actually, the above graph was of the Shinall earlier on when I just started the tri-amping process.  Below you will find your K-402-Khorn bass bin response, which shows the change in phase starting at ~130 Hz, so crossing at 450 or 250 Hz made little difference:

 

1368074911_MarkHJub-KhornSPLandPhaseResponse.thumb.jpg.e43235e9cbadab06d4c335a1ec233fb3.jpg

 

Chris

Link to comment
Share on other sites

  • 6 months later...

Ages since my last post on time-alignment on this forum. Since than I happily played with MiniDSP DIRAC to do some of it rather nice.

Now the tri-amp bug got me. Overhere in the Netherlands we make HYPEX FA123 plateamps with DAC/DSP/NCore amps: https://www.hypex.nl/product/fusionamp-fa123/154. Sounds like a great way to do this at a very affordable price around $400 a piece. Remember NCore amps are on par with Benchmark and Purify.

Haven't seen it covered in this forum but it would be interesting to hear your opinions.

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...