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

First project - Will this crossover design work?
First project - Will this crossover design work?
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 16th March 2020, 03:14 PM   #21
Douglas Blake is offline Douglas Blake  Canada
diyAudio Member
 
Douglas Blake's Avatar
 
Join Date: Nov 2019
Location: Niagara Region Ontario
Quote:
Originally Posted by Mygyth View Post
@system7 Thanks, just subbed in that woofer. Looks like i can get a decent crossover quite a bit cheaper (Image 1). Also should have a minimum of ~8.4Ohms impedence. Which, since i gather higher impedence is better makes this a significant upgrade on my last setup?
A couple of tips ... when viewing curves in X-Sim, look at the curves marked "S1", "S2" etc. The "driver only" curve is showing you what the driver itself would do... not what your crossover is doing with the driver.

From your last thumbnail, I would move R4 behind C4, this is to prevent it from carrying current all the time. I would try to adjust C2 without R2 in the circuit... If you right click on it you can temporarily short it and remove it later, if it's not helpful.

But it looks like you're getting close to a "first blush" design. Do keep in mind that computer simulation in XSim and it's cousins does not take your box or your room into consideration, so it is just a rough estimate at best.

Not bad for a beginner, though.
  Reply With Quote
Old 16th March 2020, 06:44 PM   #22
jReave is offline jReave  Canada
diyAudio Member
 
Join Date: Oct 2012
Yes, I echo the above - not bad for a beginner.

But you need to read this now - So you want to design your own speaker from scratch! - because your sims currently have 3 fatal flaws:

1) you must include baffle diffraction effects in your frd files
2) you have to account for the different acoustic centers of the drivers and their distance to your ears
3) you also have to extract minimum phase from both your frd and zma files and align each driver's phase in the xo region

I could also say that you haven't included the effects of your box on the woofer's FR and impedance files which are necessary for your design but with a 2-way with a high enough xo point, don't actually make much of a difference to the actual xo sims. Personally, I always include them anyways.

If you give us the dimensions of the old radio that you found, we might be able to make some further recommendations or even do up or double check your sims ourselves.

A couple more tips for XSim:

- in the FR graph, get rid of the phase for the system response too. It tells you nothing. Instead, check the "show with phase" box for S1 and S2. Now what you want to do is try to get those lines to match up as closely as possible in the region of the xo or in other words in the frequency range where the 2 drivers are both contributing to the summed response. Do this while at the same time keeping the summed response as flat as possible.
- you can get rid of the "driver only" curves in the Impedance graph as well
- to extract minimum phase from your files, go to the "Tune" window for each driver and for the FRD and ZMA phase source, select "derived" instead of "measured"
- in XSim, acoustic center differences are input in the "Tune" window under "mod delay" but it's a little complex.

Essentially, the acoustic center is the physical point of the driver from which the sound wave starts. It's always best to actually measure this with the drivers on the actual baffle, but for the purpose of simulations you can generally take that point to be where the cone of the driver meets the voice coil former or in the case of a tweeter dome where the dome meets the surround. The easiest way to do this is to blow up the drivers' pdf's to actual size on your computer screen and then take a physical measurement with a ruler.

One might think that the mod delay should then just be the difference between the 2 measurements but it's a little more complex than that because 1 driver is probably above the other in relation to your ears. So a little math is required.

In the pic below:
A = your ears
B = the tweeter acoustic center
C = the woofer acoustic center
D = the distance between the center of the tweeter and the center of the woofer
E = the front of the baffle

So what you are trying to calculate is the difference in lengths from A to B and from A to C. To make things simple, assume the distance AE is 1 meter and lets say the acoustic center for the tweeter is 5mm back, the acoustic center for the woofer is 15mm back and that the center to center spacing between drivers is 100mm. So from our good old friend Pythagoras, the distance AE is going to be the square root of 1000^2 + 100^2 which equals 1005mm. So now the difference from A to C and from A to B is AC - AB = (1005 + 15) - (1000 + 5) = 15mm or .6".

Back in Xsim now, you leave the tweeter delay at 0 and input .6" for the delay of the woofer. Hope that makes sense.
Attached Images
File Type: jpg Acoustic centers.jpg (52.7 KB, 56 views)
  Reply With Quote
Old 16th March 2020, 10:54 PM   #23
Pharos is offline Pharos  United Kingdom
diyAudio Member
 
Join Date: Feb 2020
Location: Seaford
Quote:
Originally Posted by Douglas Blake View Post
Sorry to but in again ... but higher doesn't mean better. It is a little easier on the amplifier because it draws less current, but that's about the only advantage.

"Better" generally comes from carefully matching drivers and box designs. In theory, the right drivers in the right box don't need a complex crossover. Sometimes it can be as simple as a single capacitor to keep bass energy out of the tweeter.

The crossover does play a role, of course, but mostly it's there to correct for our compromises.

What I meant by my previous post is that if you are doing a design and you get a real low dip in impedance someplace on the response curve, it could overheat and damage the amplifier... That is if you are designing for 8 ohms and at some point the impedance dips to 1 ohm, even while the rest is still at 8 or 10 ohms, your amp is at risk.

The opposite... that characteristic rise to 20 ohms or so at woofer resonance is not harmful. It's just hard to look at.
Couldn't agree more, as stated re. Z, in my last post.
  Reply With Quote
Old 17th March 2020, 03:30 PM   #24
Douglas Blake is offline Douglas Blake  Canada
diyAudio Member
 
Douglas Blake's Avatar
 
Join Date: Nov 2019
Location: Niagara Region Ontario
@phsros

Seems we are in agreement then...
  Reply With Quote
Old 19th March 2020, 07:44 PM   #25
jReave is offline jReave  Canada
diyAudio Member
 
Join Date: Oct 2012
Perhaps some further help at this point would be beneficial.

In my mind, no need to change the woofer yet. The 4ohm version gives you more sensitivity than the 8ohm and with small drivers that try to reach low, all the help in terms of SPL that you can get, the better. I would consider running it vented though as that's its preferred alignment and the F3 goes from about 120Hz sealed to about 64Hz in a vented box. About 2L tuned to 70Hz with a 3cm x 20cm port gives a little boost in the low end that matches the natural rising response of the woofer and helps a tiny little bit to compensate for baffle step loss. Max SPL though is going to unavoidably be a little low with small drivers like these when they are trying to run more or less full range. About 90/92dB at 1m. See pic 1.

Just to show how diffraction can work and how important it is, pics 2 and 3 show baffle diffraction effects if you put the drivers on a 6.5" x 9.5" front baffle. Note that these are just examples because when you put the boxes into the old radio frame that you posted you are going to get something different again. That was why I asked if you could post the dimensions of that unit.

Pic 4 shows what the drivers' FR looks like when you add in the baffle diffraction and the vented box response to the woofer but before any xo work.

Pic 5 is a quick attempt in XSim given all the generated data. Minimum phase was derived for all frd and zma files when entering them into the program. Using the Dayton spec sheets, I arrived at acoustic centers for the tweeter of 7mm back and 25mm back for the woofer. Unfortunately, this was a bit of a guess as the woofer is not actually pictured scaled properly to the measurements given. So I took my best guess. After calculating path differences and converting to inches, I input a woofer delay of .87". Impedance for the woofer has also been changed to match its 2L vented box alignment. Note the double impedance peaks now in the LF instead of only the single one you get with a closed box.

I tried to use the minimum number of parts because with inexpensive drivers it doesn't make any sense to do otherwise. The summed response is decently flat, the drivers' phase align very well in the xo region, impedance is totally respectable for a 4ohm amp and the xo ends up at about 2600Hz. I might actually like steeper slopes in the xo for these 2 drivers, but as I say not really worth it to start adding more and more components.

Pic 6 is what the response looks like with my frd and zma files and the 2nd xo you attached in post #16. This is just to demonstrate how different the response can be if the original files going into the simulation are not quite accurate.

Hope that helps. Please let me know if you have any questions.
Attached Images
File Type: jpg TCP115-4 VB .jpg (219.3 KB, 46 views)
File Type: jpg Twtr BD2.JPG (150.6 KB, 42 views)
File Type: jpg wfr BD.JPG (151.8 KB, 42 views)
File Type: jpg FR's w BD.jpg (78.8 KB, 42 views)
File Type: jpg XSim w BD.JPG (376.9 KB, 16 views)
File Type: jpg XSim w old xo old files.JPG (372.2 KB, 16 views)
  Reply With Quote
Old 20th March 2020, 05:46 PM   #26
waxx is offline waxx  Belgium
diyAudio Member
 
Join Date: Apr 2014
Location: Belgium
I see a lot of good thoughts, but some strange circuits for this crossover. But BSC and time alignment are indeed rather essential for a good crossover. You can do it acousticly (make sure in the box that all drivers are aligned) or in the crossover.

I did follow the example of the poster above, and used the published files the way he did (derived and so) in Xsim and used my logical way of building a cr and it's a bit different, but better i think. I don't get why you use resitors on the parallel inductors and caps in a crossover.

Anyway, this is my take... The Zorbel before the filters is optinional, but makes the speaker easier to drive by amps. Sometimes amps can't handle high impendance spikes and it influences the sound.
Attached Images
File Type: png 2WP2O.PNG (95.8 KB, 16 views)

Last edited by waxx; 20th March 2020 at 05:48 PM.
  Reply With Quote
Old 20th March 2020, 06:20 PM   #27
jReave is offline jReave  Canada
diyAudio Member
 
Join Date: Oct 2012
Hey waxx

I've used the extra resistors in the parallel legs for several reasons:

1 - they help to shape the response around the knee of each filter
2 - they help to decrease the conductivity to ground in the parallel legs (not that that is much of a problem here with the values I'm using)
3 - they cost next to nothing to do so
4 - and they fine tune and help align each driver's phase response in the xo region, helping to overcome the problem of the physically unaligned acoustic centers of the drivers on a vertical baffle

You'll need to display each drivers' phase in XSim to see how your design meets that last criteria. That's also a fairly large dip in the response between about 250Hz and 1500Hz. More than I'd be happy with. And I also have to wonder if you have included baffle step loss accurately as your LF response is showing no SPL loss from the spec sheet's infinite baffle measurements. It should be down somewhere around 6dB from the spec sheet's 85/86dB in that area.
  Reply With Quote
Old 20th March 2020, 06:39 PM   #28
waxx is offline waxx  Belgium
diyAudio Member
 
Join Date: Apr 2014
Location: Belgium
Quote:
Originally Posted by jReave View Post
Hey waxx

I've used the extra resistors in the parallel legs for several reasons:

1 - they help to shape the response around the knee of each filter
2 - they help to decrease the conductivity to ground in the parallel legs (not that that is much of a problem here with the values I'm using)
3 - they cost next to nothing to do so
4 - and they fine tune and help align each driver's phase response in the xo region, helping to overcome the problem of the physically unaligned acoustic centers of the drivers on a vertical baffle
Thanks for that, i'll experiment with it.

Quote:
Originally Posted by jReave View Post
You'll need to display each drivers' phase in XSim to see how your design meets that last criteria.
How do yo do that, i don't see that option

Quote:
Originally Posted by jReave View Post
And I also have to wonder if you have included baffle step loss accurately as your LF response is showing no SPL loss from the spec sheet's infinite baffle measurements. It should be down somewhere around 6dB from the spec sheet's 85/86dB in that area.
I used the files from Dayton Audio their website. If i would build this for real, i would work on measurings from the drivers in the box off course. But here it was more a study of the case like presented by the OP... I don't have these drivers or am building the box...
  Reply With Quote
Old 20th March 2020, 06:54 PM   #29
waxx is offline waxx  Belgium
diyAudio Member
 
Join Date: Apr 2014
Location: Belgium
Actually i just found how to do that, and used your explenation in my next version. And it's indeed better. And i could drop the zorbel, as even the most demanding amp that runs 4R can load this..

And Again i learned something on this wonderfull forum... Thanks!
Attached Images
File Type: png 2WP2O-V2.PNG (93.9 KB, 18 views)

Last edited by waxx; 20th March 2020 at 07:08 PM.
  Reply With Quote
Old 20th March 2020, 08:02 PM   #30
jReave is offline jReave  Canada
diyAudio Member
 
Join Date: Oct 2012
waxx, I'm liking your latest version better now. On the plus side, you've got a flatter FR and the phase is decently aligned around the xo.

On the negative side, the tweeter resonance frequency of about 1100Hz is only about 11 or 12dB down from the summed response. That's worrisome. Throwing a resonance compensation filter on it would help or else going to a steeper xo slope by adding a 2nd series capacitor would do it as well at lower cost.

Also, no argument from me that measurements are the way to go but when that's not possible or if just trying to improve the accuracy of the simulations, then it's imperative to include the baffle diffraction effects. These too can be simulated and then added in to frd files as I've done above. Frequency Response Modeler will do both of those things as well as more (like splicing in the modeled LF sealed or vented response) but it's an Excel program which not everybody has access to. In that case switching to an all-in-one program like VituixCAD might be the way to go.

And yup, learning is definitely cool.
  Reply With Quote

Reply


First project - Will this crossover design work?Hide 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
Pulseaudio Crossover Rack - multi-way crossover design & implementation with linux Tfive PC Based 622 Yesterday 12:13 PM
crossover design sb17nrx ,sb29rdnc. Will this work? mallikreddyk Multi-Way 40 27th October 2018 06:28 AM
3 Way Active Crossover Design project powagrip Multi-Way 13 23rd February 2005 08:06 PM
have a speaker design- need crossover- will this work? trespasser_guy Multi-Way 1 26th August 2002 05:23 AM


New To Site? Need Help?

All times are GMT. The time now is 11:48 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Copyright ©1999-2020 diyAudio
Wiki