Home Theater Forum and Systems banner

1 - 20 of 27 Posts

·
Registered
Joined
·
17 Posts
Discussion Starter #1 (Edited)
Hello.

I'm working on a 3-way MTMWW build and have a crossover that looks good to me, but wanna make sure I'm not overlooking anything or not considering/understanding something important.



My primary concerns are:

1.) Is this impedance graph accurate? If I adjust things differently I get large spikes in impedance at the crossover frequency and while I think that's normal, I'm not sure how that transmits into the actual performance of the speaker.

2.) Is the midrange spreading too far over it's band? Should it have a sharper narrower BP response, or is this okay?

3.) Being that this is a 3-way, how do I make this kind of design bi-ampable? If separating woofer from mid-tweet, will the different system impedances play differently from each amp output?

I apologize if any of my statements aren't clear enough. Please feel free to ask specific questions so I can make things more clear.

Thank you
 

Attachments

·
Registered
Joined
·
5,772 Posts
1.) Is this impedance graph accurate? If I adjust things differently I get large spikes in impedance at the crossover frequency and while I think that's normal, I'm not sure how that transmits into the actual performance of the speaker.
It is not desirable to have spikes in the impedance. These would indicate that something is amiss with the design. For a sanity check I would compare the design against another online calculator like the one here: http://www.diyaudioandvideo.com/Calculator/APCXOver/

2.) Is the midrange spreading too far over it's band? Should it have a sharper narrower BP response, or is this okay?
That will depend on the driver you chose and what it's capabilities are. Do you have a set of drivers in mind?

3.) Being that this is a 3-way, how do I make this kind of design bi-ampable? If separating woofer from mid-tweet, will the different system impedances play differently from each amp output?
Do you mean that you want separate amplifiers, both being fed the full range signal, and then the crossover after that?
 

·
Registered
Joined
·
17 Posts
Discussion Starter #3
Hello and thank you for your response

To answer your questions more clearly:

The midrange driver I would be using would be two Audax HM170C0 from madisound.
Google search brings it up as first result - I can't post links just yet :p
It says usable range from 42hz - 6.5khz. 42hz I could care less about, as they aren't gonna go that low. As far as 6.5khz goes, well 6khz will already be about -10db from it's highest range. So perhaps it's already such low power by this point that it won't contribute too much sonically, leaving most of it for the tweeter. It looks like it adds 2db still.

For bi-amping, yes. I would be using two amps to drive two ranges of the speaker. Most likely one pair of binding posts would be for the woofers, and another pair for the mids/tweet. I wanna be able to run it both off of one amp if the posts are connected, as well as from two separate amps when the posts are disconnected from each other. Good flexibility that way.

And my concern with understanding impedance curves was mostly that I see many designs that have some good peaks in them. The B&W CM1 has a good peak of about 40ohms around about 2khz. Google search "b&w cm1 impedance curve" and it's the first image.
I've seen several like this and wasn't sure if my understanding that a flatter curve is better was being taken too far by myself :p
As you can see my crossover would keep all frequencies from 43hz all the way up between 4 and 8 ohms. So I was thinking that would be pretty good. But if higher peaks are not as much of an issue as I've believed, I can alter some of the slopes of the crossover points a bit to make them a bit sharper.
Maybe I shouldn't though?

Let me know what else I could clarify
 

·
Super Moderator
Joined
·
1,921 Posts
The impedance and impedance phase look fine to me. Spikes won't kill anything but performance could vary with how your amp is able to handle it, impedance dips are far more of a concern than high spikes.

I don't see any evidence in your transfer functions of any tweeter attenuation or BSC. Have you accounted for either?

For bi-amping I would separate the woofer from the mid and tweet. As for the midrange, three octaves is a great rule of thumb to shoot for on the passband for the mid. Anything besides that will allude to phase issues and bad polar response of the finished product.

Can you repost your plots with phase of each driver shown instead of total phase, as well as the x,y,z offsets that you used?
 

·
Registered
Joined
·
17 Posts
Discussion Starter #5
Driver phase with summed response is attached.



The distances I used for quick test (work-in-progress, but close estimated distances between centers of drivers) are as follows:

Mid 1 (above ear level) = +0.150m
Tweeter (ear level) = 0.000m
Mid 2 = -0.150m
Woofer 1 = -0.380m
Woofer 2 = -0.610m

All drivers kept on center horizontal point. I tried moving them to see what would happen and it made things look more messy, so I left them centered as I'd most want to build anyway lol.

Also I have not accounted for any of your two propositions.
I don't think I'd want any tweeter attenuation, but what would the benefit in BSC be?

Thanks
 

Attachments

·
Registered
Joined
·
17 Posts
Discussion Starter #6
To compare what I had at first to what I got from using the calculator provided by robbo's post, I attached a graph with driver phase, response, and total response.
 

Attachments

·
Registered
Joined
·
17 Posts
Discussion Starter #7
Tinkered around with the values a little more while watching phase lines.
Am I correct to assume that overlapping lines will indicate drivers playing at the same time with each other?
And if so, how much shift is considered "a lot"?

My current graph looks "good" (to my current understanding) with woofer-mid and with mid-tweet, but there is an area between 500hz and 800hz that seems a little odd.

Is this considerable or not so much seeing that the woofer is down 15db at 500hz and down 22.5db at 800hz?

 

Attachments

·
Registered
Joined
·
5,772 Posts
Russ asked earlier if you can graph the individual phase responses of the drivers. Does your software allow that?

Tinkered around with the values a little more while watching phase lines.
Am I correct to assume that overlapping lines will indicate drivers playing at the same time with each other?
And if so, how much shift is considered "a lot"?
The phase shift should, ideally, be as linear as possible. The fact that it jumps from the top of the graph to the bottom is simply so it can fit on the page. It is actually continuous ie 0° = 360° and -180° = +180°

My current graph looks "good" (to my current understanding) with woofer-mid and with mid-tweet, but there is an area between 500hz and 800hz that seems a little odd.
Is this considerable or not so much seeing that the woofer is down 15db at 500hz and down 22.5db at 800hz?
That is normal, and as above, is done so it can be displayed on the one graph.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #9
Hey thanks again.

By individual driver phase, wouldn't that be the fainter blue, pink and red lines I have added to my graphs since? Or are you looking for something else?

And it wasn't necessarily the bottom to top jump I was concerned about, just that the woofer phase and mid phase becomes much more spread apart starting at 500hz.

Is that ok considering the woofer won't be doing much at 500hz anyway?

Thanks
 

·
Registered
Joined
·
5,772 Posts
Ok, I didn't see them this morning (2 am local time)
I really don't know what would cause that blip in phase and frequency response. :huh:
As for the phase difference between the drivers, you could spend a lot of time changing the type and order of your crossover and time/phase align the drivers in the cabinets or simply try it and see how it sounds in real life.
Everything about speaker design involves compromises, including their room positioning.

Getting back to BSC most manufacturers make the mid/tweeters off centre to minimize the effects. there is an article that explains it here: http://sound.westhost.com/bafflestep.htm
 

·
Registered
Joined
·
17 Posts
Discussion Starter #11
That is interesting, but I think I'm gonna keep this a bit more simple than that for now, and see how it turns out.
That way the more builds I create and add different things, I can see how they audibly change things.

Also I think that little blip down there seems to be from a notch filter I had to apply to get a nice bump out of the response.

When I order all the drivers, I'll use driver testing software to get my own specs/parameters per driver. Then I can tweak my crossover to accommodate the updated measurements.

As far as driver distance from walls goes anyway, my current idea for the cabinet will be something like my crude picture below. I need to get my 3d modeling software installed on here :p

It's a top-down view and the filled section would be for the MTM, about 19"H 6.5"W flat surface before the curves, and about 6.5" deep.

The bottom white part is the dual woofer cabinet, and I can use the flat surface space on it to put a plant or lamp on it lol. Probably gonna make part of it a separate crossover container that I will have vented and be able to be opened easily in case I decide to make any modifications. It will give my the ability to put the MTM section more forward or back too while testing if I feel I need to do so.

I'm using Jeff Bagby's design software currently too.

Another quick question now would be that when I change my vertical listening position to say 2ft above reference, I get a graph that looks like the one below.

Is this right? I tried testing with my current speakers to listen to how the sound changes, and I'm thinking it's not too far off.
 

Attachments

·
Registered
Joined
·
5,772 Posts
Also I think that little blip down there seems to be from a notch filter I had to apply to get a nice bump out of the response.
That explains it, nearly all filters will add some phase distortion.

Another quick question now would be that when I change my vertical listening position to say 2ft above reference, I get a graph that looks like the one below.
All the drivers should be the same distance from the listener otherwise you will get interference between them (comb filtering). Again, a different arrangement of the drivers on the baffle can help minimize the effect although a better solution is to use speaker stands.
 

·
Registered
Joined
·
17 Posts
Discussion Starter #13
In this case all the drivers are equal distance to me - or as close as can be since you can't just overlap drivers on top of one another.

What I'm stating would be vertical off-axis response.
Say I'm sitting and my ears are at the same height as the tweeter, and I get my original FR. Then I stand up and my ears are 2ft higher than the tweeter and then I get the FR previously posted.

Also, these will be floor standing towers about 51" in height, so stands would be a no-no lol.
In this case the woofer cabinet is the stand for the MTM cabinet.

I'll try to work on making a better 3d model of what my current thought is
 

·
Registered
Joined
·
17 Posts
Discussion Starter #15
Lol that's probably the best response ever! Love it :D

Sometimes I do a lot of work and moving around while playing stuff and just don't want response to be SUPER poor when doing so.

If it behaves anything like all the speakers I've been listening to, then that's good for me, and when I sit and watch movies, play games, or just crank a cd, then I've got me sweet spot.

It's not a large multi-purpose entertainment room anyway. It's just my 12'x11'x8' room...BEDROOM... that I've managed to cram as much in as possible. I'm gonna have to post pics, you may get a kick out of what I've managed to pull off in such a sized room.

Tetris is never a waste of time!
 

·
Registered
Joined
·
17 Posts
Discussion Starter #17
Guess what?

Already implemented room treatments XD

The acoustic panels made such a difference!

But I still have to treat my rear wall and ceiling. I'm waiting until I get a fan installed on my ceiling and a custom shelf unit/bookcase on the rear wall. Then I'll treat so I know I'll be getting the right sized panels.

Also thinking about going from 4" thick to 6"

Clean bass is just hard to get in here. Although slowly getting better, it doesn't come close to having that theater slam or cleanliness of my local Best Buy's audio room.

Somehow I think adding a second subwoofer might help a bit to cancel standing waves and fill voids. But I'd either need to get another 5400 ($926) or try a pair of Dayton Ultimax drivers - they look like some good things can be acheievd.

Thinking about ordering the drivers and the DATS measuring software Sunday or Monday.

Now when it comes to testing the drivers, should I test each driver inside the cabinet in which it will be intended? And then design the crossover based on the cabinet?
 

·
Registered
Joined
·
17 Posts
Discussion Starter #19
Thanks.

I figured I'd probably be going with a sealed .707Q alignment for the mids anyway since I'm crossing them with the woofer around 250hz.

I'll probably start a new thread when I start working on the actual project then. Keep it all together for everyone to see and get answers to any questions of the same they might have.

Can't wait!
 

·
Registered
Joined
·
17 Posts
Discussion Starter #20
Ok, so after getting DATS I measured my actual drivers and now I'm not sure if they're gonna perform quite how I thought they might have before.

Main reason is that now I have the actual phase of each driver to work with.
Before when I was using manufacturer specs, there was no phase listed and it was basically always at 0degrees for the driver.

Let me just simply ask, how much phase shift is too much?
Like for my new graph, the woofers and mids are about 30degrees apart at crossover (313hz), and are about 60degrees apart under that.
The mids and tweeter are apart 45degrees at crossover (2.5khz) and more above that, and a little less under that.

What will this mean when listening to them?



Thanks
 

Attachments

1 - 20 of 27 Posts
Top