Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Multi-Way Conventional loudspeakers with crossovers

Small active 3-way floor stander with a punch
Small active 3-way floor stander with a punch
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 10th December 2018, 05:05 PM   #81
jReave is offline jReave  Canada
diyAudio Member
 
Join Date: Oct 2012
Ok, just to be sure we are on the same page, these are the articles I was referring to:

Accurate in-Room Quasi-Anechoic Frequency Response Measurements
How to Find the Relative Acoustic Offsets

Notice that with Jeff's pseudo-anechoic method, you will be designing for a more or less flat FR, whereas with a more ungated in-room measurement you should design for a descending response (low to highs).

Download PCD if you have Excel here: Passive Crossover Designer. Functionally can vary depending on which version of Office you have - earlier is better.

But finding the acoustic offsets can also be done in XSim if you simply make a couple of adjustments. First skip all the driver offset parameters. So just input the correct mic distance. Then setup 3 drivers in the xo section. Wire up drivers 1 and 2 as you would normally but without any components in the circuits, but leave driver 3 unattached on its own. Drivers 1 and 2 will be the individual driver measurements and driver 3 will be the summed measurement. Make sure you include the phase info - I've made that mistake a time or two. I normally just extract minimum phase. In the FR graph, choose the 'S3 (driver only)' curve and the System curve. Now simply increase the physical offset ('mod delay') of the larger of the 2 drivers in the 'Tune' section of that driver until your 2 curves more or less coincide. That will give you the correct difference between the 2 path lengths between your mic and the 2 drivers. Repeat for your next driver pair.

That's a quick overview. Let me know if you need more detail than that.
  Reply With Quote
Old 11th December 2018, 08:10 AM   #82
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Quote:
Originally Posted by jReave View Post
Ok, just to be sure we are on the same page, these are the articles I was referring to:

Accurate in-Room Quasi-Anechoic Frequency Response Measurements
How to Find the Relative Acoustic Offsets

Notice that with Jeff's pseudo-anechoic method, you will be designing for a more or less flat FR, whereas with a more ungated in-room measurement you should design for a descending response (low to highs).

Download PCD if you have Excel here: Passive Crossover Designer. Functionally can vary depending on which version of Office you have - earlier is better.

But finding the acoustic offsets can also be done in XSim if you simply make a couple of adjustments. First skip all the driver offset parameters. So just input the correct mic distance. Then setup 3 drivers in the xo section. Wire up drivers 1 and 2 as you would normally but without any components in the circuits, but leave driver 3 unattached on its own. Drivers 1 and 2 will be the individual driver measurements and driver 3 will be the summed measurement. Make sure you include the phase info - I've made that mistake a time or two. I normally just extract minimum phase. In the FR graph, choose the 'S3 (driver only)' curve and the System curve. Now simply increase the physical offset ('mod delay') of the larger of the 2 drivers in the 'Tune' section of that driver until your 2 curves more or less coincide. That will give you the correct difference between the 2 path lengths between your mic and the 2 drivers. Repeat for your next driver pair.

That's a quick overview. Let me know if you need more detail than that.
Yeps that's what i read as well. Didn't read the "..Quasi-Anechoic.." though. I will study those two and test it out tomorrow -hopefully.
Thanks for the tips! Hope i can figure out the XSim software.. still haven't used it..


Until then here is what it looks like at the moment:
  Reply With Quote
Old 11th December 2018, 08:35 AM   #83
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Default MW19P-04 - Mid/Midbas driver

Time for some measurements!
This post will focus on my mid/midbas driver only.
Let’s just jump in!

The MW19P driver is mounted in a 13,4-ish L cabinet with a ScanSpeak flow resistor in the back. As for damping, I still didn't receive my felt and wool so i just added MDM-4 on all sides and two layers on the angled back plate. Also a roll behind the tweeter chamber.
This is the raw response measured at 60cm at a height between the mid and tweeter drivers. Gated at 4ms - First reflection arrive at 4,4ms.
Click the image to open in full size.




I then added a -5dB baffle step compensation with a high shelf Q=0.7 at 400Hz. I Used REW to generate generic EQ parameters to equalize the response to my target. This resulted in 5 notches with Q from 2 to 5 and -2 to -6dB from 3000-9000Hz.
From 60cm gated at 4ms this is what it looks like
Click the image to open in full size.




A near field measurement 10 cm from the dustcap of the compensated woofer without the bafflestep compensation looks like this:
Click the image to open in full size.




And with bafflestep compensation it looks like this:
Click the image to open in full size.




And the the far field merged at 300Hz with the near field without bafflestep compensation looks to the best of my abilities like this:
Click the image to open in full size.




I tried two different crossover topologies. Both crossing at 2500Hz between the tweeter and midrange.
The first is LR24 slopes on everything. On the HF-end of the midrange this is what the driver measures like along with the target curve (60 cm 4ms gating):
Click the image to open in full size.




And I tried a Harsch topology. This means a 4th order Butterworth on the LF driver. It looks like this (60 cm 4ms gating):
Click the image to open in full size.



For both cases above the phase looks good.





All in all it now looks pretty good in my eyes. Almost textbook
Or did I do something wrong? Is there something I am forgetting to look at?

With regards to low end cutoff. The woofer appears to have a pretty clean roll-off with a F3 of 70-ish Hz. How would you move this rolloff to f.ex. 150 Hz? Any kind of high pass filter will generate a wired roll off never hitting a straight line due to this natural roll-of being so close.

Should i use a LR-Transform or something instead?

Last edited by DannerD3H; 11th December 2018 at 08:42 AM.
  Reply With Quote
Old 11th December 2018, 08:58 AM   #84
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Default R3004/662000 – High driver

This post will focus on my High frequency driver only.

The ScanSpeak R3004/662000 may not be the best match to the MW19P. It was a driver I had from my old speakers and I really love the look of it. So I will try to see if it can work.

This is the drivers raw (with a LR24 at 500 Hz for protection) on axis response measured at 60cm at a height between the mid and tweeter drivers. Gated at 4ms – First reflection arrive at 4,4ms.
Click the image to open in full size.


I am a bit surprised of how uneven it measures. I guess the bumps at 2K and 3,8K along with the dips are diffraction issues?
I did however use REW once again to dampen the bumps and achieve my target curve. Dampening the bumps was probably a mistake? Or?
Rew gave 3 nothecs woth Q2-4. And it looks like this:
Click the image to open in full size.


With the LR24 slope:
Click the image to open in full size.


And the 2nd order Bessel needed on the High frequency driver in the Harsch topology:
Click the image to open in full size.

The LR24 is textbook I guess?
The Bessel could do with a little fine tuning.
Should I remove the dampening of the bumps? Or are they not defractions?
  Reply With Quote
Old 11th December 2018, 09:14 AM   #85
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Default Tweeter and Midrange integration

This post will focus only on the integration between the tweeter and the midrange. All graphs here are gated at 2,2ms instead of 4ms because i moved the speakers and the first reflection now arrives at 2,4ms.

The MW19P would prefer a low crossover point (around 1500 I guess) and the tweeter would prefer a higher (2500 or above – 2000 sounded harsch). But anyway.. This is what my LR24 looks like crossed at 2500Hz. The tweeter is delayed 64us. This should be 22mm. I would have thought it to need less than that. Maybe I will get a better image of that when I try your links JReave#.
This is the FR-response and phase:
Click the image to open in full size.

And the Step response:
Click the image to open in full size.

The FR response is pretty, the step is as expected, but the Phase? What is happening here? If I look at the mid alone the phase is fine, but this phase shift happens right at crossover point on the tweeter alone. It follows the LR24 as you can see on the tweeter measurements. Is that normal with a LR24?

The same phase shift happens with the Harsch topology if the tweeter is not delayed. The Harsch topology says that the high frequency driver should be delayed half the period of the crossover frequency. This would be 264us (200+64us from time-alignment as in LR24). However this does not look good. 300us however, looks pretty decent both the FR-response and the Phase:
Click the image to open in full size.

And has a pretty good step response:
Click the image to open in full size.



Here are some comparisons between LR24 = Purple, Harsch with 264us delay = Green and Harsch with 300us Delay = Red:
Click the image to open in full size.


And here are the LR24 and Harsch on axis Blue and Red respectively and 10 degrees off axis Cyan and Orange respectively.
Click the image to open in full size.

Here the LR definitely outperforms the Harsch!


What are your thoughts about my results so far?

Last edited by DannerD3H; 11th December 2018 at 09:17 AM.
  Reply With Quote
Old 11th December 2018, 09:48 AM   #86
Juhazi is offline Juhazi  Finland
diyAudio Member
 
Juhazi's Avatar
 
Join Date: Mar 2007
Location: Jyväskylä, Finland
Midrange looks like having too much BSC, but distance 60cm is actually a bit too close to estimate it. 1m is better, but more prone to reflections. Anyway, it is easy to change IR gating in REW and you can look at the response with different settings, then you can catch and eliminate in your mind at least floor bounce dip.

Mid's low end obviously needs some eq lift or just use lower Q with Shelf 2 in HypexFilterDesign. Shelf 1 has fixed Q value. Remember to check distortion with xo set!

Tweeter's response got strange with autocal, abandon it and set BQ's manually. Tweeter's narrows rapidly above 10kHz, so on-axis should have rising response, some drivers have that naturally.

I have zero experience with S.Harsch xo, no comments. I love LR2 despite the MT off-axis response dip. It can be compensated with eq, on-axis should be a little hot around xo, reflecting off-axis dip. Usually straight response at 10¤-15¤ sounds best. This is what I aim for in my MR18 project. Then listening to music for several days tells more about it.
__________________
Radikal aktivist AINOgradient speaker project
  Reply With Quote
Old 11th December 2018, 11:02 AM   #87
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Juhazi#

Maybe I should also post a smoothed far field measurement to show a better picture of the overall tonal balance. I think the BSC is close to right. But will try 1m instead.

Why would you lift the mid's low end? Remember I will be crossing to my side mounted 10" subs at 100-200Hz (so far 150Hz is my favorite). So I have to "lower" the low end to match this crossover point. But I find it hard to not roll it off to fast! My target is a 2nd order Bessel with FC at 150Hz.

Yeah I will revert the EQ on the tweeter and try manually. But you wouldn't do anything about the bumps on the tweeter?.


I like a dropping response from 20-20k in listening position. But maybe its too much if it's already dropping at 60cm?
  Reply With Quote
Old 11th December 2018, 02:16 PM   #88
Juhazi is offline Juhazi  Finland
diyAudio Member
 
Juhazi's Avatar
 
Join Date: Mar 2007
Location: Jyväskylä, Finland
Low end of mid - to get xo curvature right you should first have straight response at least one octave beyond the xo point, preferably 1,5-2! If your xo is 150Hz mid should be eq'd straigth from 75Hz up! Same for lowpass.

REW autocal is designed for correcting room response (power response in farfield), and that is why autocal wants to see a downhill up high. I haven't used it at all, perhaps one can set target curve manually or from options?

Tweeter response in (semi-)nearfield and short gating always suffers from bumps and dips that are caused by diffractions of from baffle edges and even the driver's front plate. At least you should check how they behave off-axis. Farfield/listening spot measurements with long gating or RTA/pink noise tell you the power response above 500Hz, whic tells how it sounds to you (sound perception, neurophysiology).Check at least this before eq'ing treble wiggles! 0-500Hz wiggles are boundary reflections and room modes, there use only ½ octave (low Q) corrections, except perhaps the stronges mode bump around 40-50Hz.

Fusion
Click the image to open in full size.
__________________
Radikal aktivist AINOgradient speaker project

Last edited by Juhazi; 11th December 2018 at 02:21 PM.
  Reply With Quote
Old 11th December 2018, 02:58 PM   #89
DannerD3H is offline DannerD3H  Denmark
diyAudio Member
 
Join Date: Jun 2013
Location: Denmark
Regarding low end of mid - Aah okay, I was thinking of doing that but thought that there may have been a better/more correct way of doing it. Will try that when I have mid-high in place then

You can adjust the target curve to i.e. flat if that is what one wants.

But when do you want flat and when do you want the downhill?
Is it just a question of how far away you measure?
Or is there also a "how" you measure? This "Power response" is new to me. I will google that..

Aaha.. thanks for that info! The hardest part about this for me is my lack of experience and knowing what to correct and what not to correct, so this is gold. Thank you
Even though i thought i have read so much it still isn't enough xD
  Reply With Quote

Reply


Small active 3-way floor stander with a punchHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
3 way floor stander speaker placement in small room ... help dannyjmf Multi-Way 9 10th September 2018 02:34 PM
Floor Stander MarKen for the A7P? mikejennens Full Range 10 26th February 2018 02:17 PM
Help with floor stander design under $1K hahbs Multi-Way 9 12th December 2016 12:39 PM
Initial thoughts - 3-Way active floor stander as first project... bushmeister Multi-Way 29 14th February 2013 02:41 AM
First DIY Build help. 3-way floor-stander. Elias318 Multi-Way 20 9th June 2012 01:18 AM


New To Site? Need Help?

All times are GMT. The time now is 11:08 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki