Jump to content

Recommended Posts

I am new to active crossovers, new to using REW to test speakers and Klipsch speakers.  So I am hoping someone with more knowledge can "look over my homework" to make sure I am on the right path.  @Chris A or anybody else with active speakers or familiar with active crossovers, I would be happy for y'all to look at driver graph and then see if crossover points are correct.  This is a tri-amp setup via 2 MiniDPS 2x4HDs (one doing left channel and the other doing the right) with 3 AIYIMA A07 amps.

 

While I am not overly familiar with REW, I have watched a few videos on how to use it to measure your speakers.  Scott Hinson has video on how to measure each driver and I tried to follow his steps to get my response graph for each driver (1v @ 1m because I have neighbors).  While it may not be perfect since I could not take my speakers outside to measure each driver, I think they are close.  Then again, I maybe fooling myself, which is why I am hoping someone can look this stuff over and tell me I am all wrong or on the right track.

 

Lastly, I attached 4 MiniDPS settings "screens".  The low pass, for the woofer, the band pass for the squawker, and the high pass for the tweeter.  The 4th is the time delay that I derived using the tech paper that MiniDPS posted on measuring impulse response to time align your drivers.  There is no delay for the woofer, a 2.8ms delay for the squawker and a 4.45ms delay for the tweeter.  I have tried to verify these settings and there is still a few uSeconds of variance between all the drivers.  If I change the delay on the tweeter, the impulse either fall just in front or just behind the squawker--I can't seem to get them to align perfectly.  This could be because: I am not familiar enough to using REW to measure the impulse correctly OR I think I am hitting the delay resolution of the MiniDSP OR I could just have totally messed up the original impulse response measurement and these number of nowhere remotely correct.  So fi these delay settings seem incorrect, I would be happy to revisit them.

 

With all this said, they still sound a LOT better than it did with the original AA crossover.  So I think I am on the right track.

 

Thanks!

 

 

Graph1.png

LowPass1.png

midband1.png

hipass1.png

timedelay1.png

Link to comment
Share on other sites

20 hours ago, 1bigpig said:

While it may not be perfect since I could not take my speakers outside to measure each driver, I think they are close.

I've not used outdoor measurements to correct the SPL response, only to do polar coverage measurements--like this one:

 

K-402-MEH horizonal normalized sonogram.jpg

 

The notion that you have to do outdoor measurements to correct for in-room response really doesn't work very well, I've found (i.e., in my experience).  It's better to use 1 m measurements in front of each loudspeaker in-room with plenty of absorption material on the floor between the loudspeaker and microphone.  I've also found a lot of opinions from folks on other audio forums that have some very strange notions of what they can and (more importantly) can't do--that are actually not very well calibrated opinions.  The best thing to do is to make different measurements in-room for yourself and compare results.  That seems to be a unique point of view to those that hold those strange notions of how "not to do it".

 

Even the low frequencies--which begin to come into nearfield measurement at frequencies below 200 Hz--are better measured in--room at 1 m, and if you get below ~50 Hz, then you can shift to listening position measurements--due entirely to room mode effects all over the room. 

 

20 hours ago, 1bigpig said:

There is no delay for the woofer...

This is usually the case for everything but perhaps a Klipsch Cornwall with its long midrange horn and direct radiating 15" woofer.  Even then, the amount of bass channel delay is very small relative to the wavelength periods at crossover.

 

20 hours ago, 1bigpig said:

If I change the delay on the tweeter, the impulse either fall just in front or just behind the squawker--I can't seem to get them to align perfectly.  This could be because: I am not familiar enough to using REW to measure the impulse correctly OR I think I am hitting the delay resolution of the MiniDSP OR I could just have totally messed up the original impulse response measurement and these number of nowhere remotely correct.

As you might know, the order of the crossover filters used also affect the required channel delays.  If, for instance, you use first order filters instead of eighth order filters, the needed delays will decrease by 90 degrees of phase per decrease in order of filter employed.  If you look at the phase response (assuming you've got absorption material on the floor to absorb early reflections between the microphone and loudspeaker), you can read the needed higher channel delay to add or subtract directly from the excess group delay plot.

 

In terms of hitting the delays right--it is possible, especially knowing that you can make microsecond adjustments in the channels.  If you look at the excess group delay lag or lead as you decrease in frequency (after punching the "generate minimum phase" or "generate minimum group delay" button), then you can read the exact amount of channel delay to add to the higher frequency channel.  You might need to zoom in pretty far on the vertical axis to read it correctly for the higher frequency channels since the wave periods are very short.  Here's a thread on that topic:

 

20 hours ago, 1bigpig said:

With all this said, they still sound a LOT better than it did with the original AA crossover.  So I think I am on the right track.

Yes--and that should be made clear to all that are thinking about going down this path, too.  Better sound quality is better sound quality.

 

Chris

  • Like 2
Link to comment
Share on other sites

Here's an example of using excess group delay to find the needed channel delay adjustments:

 

Example Group Delay Time Misalignment (Bass Bin 4 ms).jpg

 

Sometimes as you get very close to the correct delays, you'll find that inverting the polarity of either the higher frequency or lower frequency driver channel will bring the two channels closer together in terms of excess phase or group delay (i.e., the first derivative of the phase curve).

 

Chris

  • Like 2
Link to comment
Share on other sites

@Chris A Thanks!  I am gonna look at the "Using REW to Determine Time Delay".  I am also reading Scott Hinson's "Determines crossover points via driver distortion".

 

I am trying to learning all this stuff.  There is a LOT of features and stuff to learn in REW.  I have not explored the "group delay" or "minimum phase" features of REW.  That is something that I am gonna investigate.

 

When I took my measurements, I did not put anything on the floor between the speaker and the mic.  I will definitely do that and take another set of measurements.  Do you recommend anything in particular?  Cargo blanket, rubber mat, bedding/comforter or something else between the doghouse and the mic?  FYI, I am using a UMIK-1 for my measurements mounted on a camera tripod with a meter yard stick to measure distance.

 

Lastly, I understand that 1st, 2nd, 3rd and 4th order in passives changes the phase by 90' for each order.  I was under the assumption that in the digital domain, the crossover order did not change the phase.  Am I wrong with that assumption?

 

I got some homework to do tonight.

Bruce

 

PS the MiniDSP has 4 settings.  I am thinking of copying the cross over points and slopes of different Klipsch crossovers to see how "different" they each sound with a press of a button--but that is for another day.

 

Link to comment
Share on other sites

I use an old foam rubber mattress topper, but anything that's relatively thick will do--such as several thick blankets laid out horizontally between the microphone and loudspeaker under measurement.

 

4 minutes ago, 1bigpig said:

Am I wrong with that assumption?

 

Yes.  It works just the same as passive in terms of phase growth--unless you're using FIR filtering, in which case you can control both amplitude and phase vs. frequency, but typically at a greater required level of computing capability (especially at lower frequencies). 

 

24 minutes ago, 1bigpig said:

PS the MiniDSP has 4 settings.  I am thinking of copying the cross over points and slopes of different Klipsch crossovers to see how "different" they each sound with a press of a button

 

That's a great idea.  I think it will be a pretty big revelation/eye-opener.  I know that I was ready to sell a pair of Belles, one of which I was using between Jubilees, and until I tri-amped and flattened the SPL response, I didn't think it was going to work at all.  Short story: it did work...

 

 

Chris

  • Like 2
Link to comment
Share on other sites

ibigpig,

Is the LaScala still running K-55 / K401 mids and K77 tweeters ?

 

Recently I used 2 miniDSP 2X4 HDs to triamp my K-Horns. However, I used an Eliptrac horn and B&C DCX464 co-axial mid/hi drivers. ChrisA was a great help and very patient with my beginner questions on using REW. One of the hardest things I found while using REW was getting the x and y axis settings right, sometimes people say to use a particular REW display and when you do it doesn't make sense. ChrisA solved my queries by showing the control panel for the display so I could set the right parameters.

Some more reading for you -
https://community.klipsch.com/index.php?/topic/195572-bc-dcx464-as-the-mid-hi-drivers-for-a-k-horn/&ct=1620012789&_fromLogin=1

 

 

 

 

  • Like 1
Link to comment
Share on other sites

@Wirrunna yes I am running a stock '79 LaScala and '80 with Atlas 55V drivers.  One has a CTS K-77 and the other is an EV K-77.  I have not opened the doghouse yet, as no need.  I did purchase some Crites K77 (DE120) replacements tweeter, but have held off doing the swap so far.

 

I did rebuilt the AA crossovers to ALK's "what wrong with the AA crossover" specs.  It sounded way better than the stock AA--probably dues to the age of the old caps.

The original AA had a lot of "brightness" and a fair bit of sibilance.  The "new" AA is much smoother but still has some honk/harshness when turned up.  Turned up is relative, as I am running a 300b SET to drive these.

 

This is what drove me to look at the active crossovers. To learn something new and the ability to adjust the honky-ness of the mids via crossover.  The MiniDSP allows you to EQ the system, but so far, my experiences with room EQ (REW and Audissy) has left my systems sounding dull and uninteresting.  I do understand a lot of music is just that--dull and flat, but I find that if I remove most of the sibilance and harshness but run the speakers with just basic correction (drivers, not whole system), it sounds a lot more dynamic and exciting.  Maybe that will get fatiguing after long listening sessions, but right now, I don't have enough time to sit down and listen to more than a single album at a time--if that much.

 

A lot more learning is what I have to do.  My understanding of digital crossover order and phase was incorrect.  I also need to learn how to better use REW--offset with the desire to listen to actual music instead of sweep tones.

 

Just for the record, the ALK "Whats wrong with the AA crossover" crossover points are:

ALK AA
Woofer
1.3mH 975 hz Low pass 6dB per octave

Squawker
47mF 420 hz High Pass 6dB per octave
.3mH & 2.2mF 6000 hz upper band pass 12dB per octave

tweeter high pass
~5000 to 5500 hz high Pass 18dB per octave

 

Once I look at the schematics of some other Klipsch crossovers, I will see if I can reverse the crossover points and slopes.  Then I will make a MiniDSP .xml file for these different crossovers so other MiniDSP users can load them to "simulate" the crossovers of yore.

 

Back to the grindstone...

Link to comment
Share on other sites

1bigpig, was the stock '79 LaScala an AA ? I thought they would have been AL3 . I am familiar with ALK crossovers ( see https://www.stereonet.com/forums/topic/118361-k-horn-crossovers-from-ak-3-to-universal-to-es/)

I had changed the K77 to a Crites DE120 and the crossover to 4500 between mid and hi to try and avoid the upper mid harshness at higher volumes, ALK told me to get rid of the K401 horn and use an Eliptrac. That certainly changed the sound and got rid of the harshness and getting phase fairly flat amongst other things brightened up the music.

 

With the two miniDSP 2X4 HD  you could easily change your mid to hi crossover frequency to 4500 if you put in the DE120s .

Link to comment
Share on other sites

@Wirrunnathese had AA crossovers as stock (if not stock, the crossovers are the age as the speakers).  I saw a pair of '85 K-Horns that had AL crossovers.  Unfortunately, the bottoms has a little water damage (probably had water damage from Hurricane Harvey in 2017 and while the speakers were only in a few inches of water, the veneer and bottom MDF material was all bubbled and warped.  I passed, but I keep kicking my self thinking this would have been a "better" project since I would have re-veneered the front and build some corner baffles so they would not have to been placed in a real corner.

 

Unfortunately, I have not had time to mess with my speakers this week.  With what little time I have had, I have been reading up on phase alignment. While I have questions, I think they will be answered as soon as I run some sweeps and look at the results--or at least I will have a better formulated question to ask here.

 

Speaking of tweeters, have you heard DaveA's SMAHL tweeter?  I have a mental block with tweeter (and midrange horn) mounted to the back side of the motorboard.  The cutout acts like a diffraction slit.  The SMAHL tweeter is flush mounted to the outside face.  The downside is that it is just a lens and not a true horn.  Either way, they both use the same DE120 driver.

 

Lastly, has anyone ever put a 45 degree chamfer on the outside of the mid range horn cutout?  Would it make a difference with sound dispersion/refraction.  Logic says it should be smoother but maybe someone has tried it and found it made no difference because of the long wavelengths of the midrange.  First world problems, right?!

 

Bruce

Link to comment
Share on other sites

I have not heard DaveA's SMAHL tweeters but have read many satisfied user posts.

My 1990 K-Horns mounted the K-77 flush with the front of the motorboard with z brackets that were screwed into the back of the motorboard. A google search for z brackets will bring up examples. 

This old thread

 shows them, Crites may have them. You would have to do some minor hacking to the motorboard to mount them.

I have not seen mention of chamfering the exit of the K-401 mid horn so I suspect that it would not make much difference to the sound.

 

As far as phase alignment goes, I suggest that you get SPL frequency flat and the drivers time aligned first, then go down the rabbit hole of phase alignment via FIR filters on the miniDSP 2X4 HDs. I found there is a lot of measure, tweak, load new config, measure, tweak, load new config ...... and go nuts and restore to the config I saved before starting applying the latest brain fart. Start again in a few days.

Link to comment
Share on other sites

  • 2 weeks later...

@Wirrunna just wanted to post a quick update since I have been absent.  I decided to try  DaveA's SMAHL tweeters.  I decided that I could not get past the tweeters being mounted behind the motor board.  I was able to install them (with a slight amount of manual sanding of the opening) and then promptly went out of town.  So I got back and have been listening to them.  My sound memory is not what it use to be after being gone for a week, so I don't know for sure if it is a significant upgrade or not.  Without any adjustment, it does seem to be brighter.  I have not had a chance to EQ the tweeters yet.  I believe it is either no EQ or EQed for the K77--which is why it might be too bright.

 

What IS an upgrade is my decision to upgrade to a MiniDSP 4x10.  Now I can control everything in one program AND I do notice less noise than the 2 2x4HDs.  I guess if I knew I was going to do this project from the start, I would purchase the 4x10 instead of two 2x4HDs.  It is about $100 more, but the unit is nicer and has less noise.  I already had a 2x4HD from another project, so buying a second unit was a cheaper way to go active for the La Scalas.

 

For kicks, I decided to plug in the ALK "Super" AA passive crossovers I built--I added banana plugs to make it a super fast swap the speaker between active and passive-- and I do think the analog crossover has very smooth presentation.  It is not an apples to apples comparison, as I am using a tube amp with the passive compared to solid state amps for the active.  I guess I need to wire in the tube amp into the mid-range and see if I get the best of both worlds...

 

So more testing to do and I want to measure the new SMAHL tweeters to see what their curve looks like.

Bruce

Edited by 1bigpig
Link to comment
Share on other sites

12 hours ago, 1bigpig said:

For kicks, I decided to plug in the ALK "Super" AA passive crossovers I built--I added banana plugs to make it a super fast swap the speaker between active and passive-- and I do think the analog crossover has very smooth presentation

 

I believe that this is entirely a function of the channel gains, delays, and PEQs that you've set in the 4x10 HD.  If you want to match the exact presentation of the passive crossovers, that's very doable.  Just measure the SPL and phase response of the loudspeakers with passives, then transfer that response to the DSP crossover.  Voila! An exact match.

 

Chris

  • Like 1
Link to comment
Share on other sites

@SchuI posted the individual drivers on my first post.  But this is it again.  I have not had a chance to run a sweep on the new SMAHL tweeter yet.  Maybe sometime this week.  All sweeps were full range 1v at 1m.  UMIK mic height was adjusted for each driver.

 

The blue line is a CTS K77 tweeter.

 

Bruce

 

drivers.thumb.png.c9e0908a2d404a16626676c9bc01809c.png

Edited by 1bigpig
Link to comment
Share on other sites

@Schu

 

This is a quick and dirty SMAHL tweeter sweep.  I say quick cuz, my wife said "Turn that sh!t down!"  So this is all I got.  The tweeter was at 1m but the mic was not exactly the same height as the tweeter (thus "dirty").  Funny thing, it matches the DE120 "factory" curve almost exactly.

 

B&C DE120 factory curve:

 

Bruce

SMAHL.png

Capture.PNG

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...