Jump to content

3 way crossover design in Xsim, feedback sought.


Recommended Posts

The snapshot that hopefully appears with this post is a work in progress for a WMTM tower loudspeaker. Any constructive feedback on the frequency response would be appreciated. I am a little concerned that the curve for the mids runs higher than the total response perhaps indicating cancellation with the woofer. Is this a big problem?

 

Thanks 

 

Ken

IMG_1788.JPG

Link to comment
Share on other sites

@Brian, the miking was 1 m on axis, 2.83 volt signal. @davewantsmoore, this confirms my interpretation thanks. How would this come across aurally compared to in phase...?

Reversing the woofer contacts and then smoothing out the summed response should put the two drivers back in phase?

 

Thanks,

 

Ken

Link to comment
Share on other sites

15 minutes ago, Ken Arrowsmith said:

1 m on axis

 

By that I assume it was on the tweeter axis vertically?

 

If so, I expect the MTM mids will be beaming/lobing, and the woofer is being measured way off axis, which may be part of it.

 

It's a very complex question with little to go on at the moment.

 

 

Link to comment
Share on other sites

Brian, thanks for taking an interest..

As I recall you are correct the measurements were taken 1m on axis to the tweeter.. are you implying the measurements should be on axis for each driver? That would make sense to me. How then do we measure the two paralleled mids....? On the tweeter axis would be equidistant from the centre of the two cones...?

 

Any advice appreciated.

 

Ken

Link to comment
Share on other sites
On 10/09/2021 at 1:42 PM, Ken Arrowsmith said:

The snapshot that hopefully appears with this post is a work in progress for a WMTM tower loudspeaker. Any constructive feedback on the frequency response would be appreciated. I am a little concerned that the curve for the mids runs higher than the total response perhaps indicating cancellation with the woofer. Is this a big problem?

 

Thanks 

 

Ken

IMG_1788.JPG

 

Link to comment
Share on other sites
On 10/09/2021 at 1:42 PM, Ken Arrowsmith said:

The snapshot that hopefully appears with this post is a work in progress for a WMTM tower loudspeaker. Any constructive feedback on the frequency response would be appreciated. I am a little concerned that the curve for the mids runs higher than the total response perhaps indicating cancellation with the woofer. Is this a big problem?

 

Thanks 

 

Ken

IMG_1788.JPG

Initial thoughts are ... woofer crossing way too low in terms of dB

Assuming the woofer natively is around the 82-89 dB range I would be looking to cross this with the mids around the 200 -700 Hz range and somewhere around the 85 to 90 dB range

I would then be looking to cross the mids  depending upon the mid woofers frequency response around the 1500 to 4500 Hz range and then deal with the tweeter as required. L-Pads and or notch filters can be added at that point.

In most cases reverse the polarity of the mids or you will get them playing too low and out of phase.

 

Generally I'm more concerned with the system resistance and prefer that the speakers dont dip down under 4 ohm, especially with 8 ohm speakers. Time alignment and phase issues can be dealt with a little later, firstly I like to get the speakers crossing smoothly at my required crossover points with a 6 to 12 dB slope.

BTW phase looks okay in this image as I dont see any huge dips at the crossover points.

 

There one thing I constantly have to remember when using Xsim and thats to bring the source power up to the range the speakers are going to be typically playing at as this changes how the resistance slopes work. 

 

Three way speakers can be a ***** to get to play nice together, especially in the type of configuration you are designing. I'm current having to redesign my crossovers for my Strads as I nuked the tweeters  and am in the process of updating them.

Hope this is some help.

image_38766.jpg.f834c4f9c6d7901ef6c907c9648a1d5c.jpg

The Strads are now 8 years old and the tweeters have expired.  Currently using the Arias pictured below alongside their bigger cousins The Strads.

 

 

image_38767.jpg.7e434b2403de5a26770e2b00db8d6a3d.jpg

Link to comment
Share on other sites

On 13/09/2021 at 10:43 AM, Ken Arrowsmith said:

How would this come across aurally compared to in phase...?

Difficult to generalise.    Lack of clarity, might sum it up (pun alert).

 

The summed response should be at least a couple dB above the mid through that region, but it's getting towards 5dB below.   This means over 50% of the sound pressure in the room is being cancelled, by the other driver.

 

On 13/09/2021 at 10:43 AM, Ken Arrowsmith said:

Reversing the woofer contacts and then smoothing out the summed response should put the two drivers back in phase?

Not necessarily no.   It may improve it... it may not.

 

The solution lies in designing the rolloffs of the drivers (via their crossover filters) so that they are inphase with each other where they overlap.

 

Using a high pass filter (or increasing the strength of the high pass filter) on the mid driver looks like a good start...... but there's 17,000 ways to skin this cat.

 

 

Depending on what you do with the filters.... the driver levels look wrong too.

 

The mid is too loud (or the tweeter to soft) ... and the woofer is too low.    It's difficult to tell if that's just because of the too strong (or "wrong") lowpass filter on the woofer ......  or, OTOH, the rise in level at 100Hz, could be not real (and just a measurement effect.... I assume these measurements are inside a small room).   It's super difficult to troubleshoot.

 

Something is up with the HF phase..... this might be to do with your driver responses themselves (ie. the data you're feeding Xsim) ... or how the filters interract.

 

Do you have the raw driver responses (measured from a common point, 1 or 2m)? .... if so, I would recommend you go back to these, and think about the target responses for the drivers, and the filters to use.

Link to comment
Share on other sites
1 hour ago, kiwilistener said:

BTW phase looks okay in this image as I dont see any huge dips at the crossover points.

It isn't just dips at XO points that are a tell about the phase.

 

Between ~400 and 2000Hz we can see that the combined response is actually less than the mid driver itself.   This means that another driver (the woofer) is cancelling some of it's output..... because they are not in phase.

 

If we could see the individial driver phases, we would see the issue clearly (mid / woofer output of phase)

 

 

1 hour ago, kiwilistener said:

 I would be looking to cross this with the mids around the 200 -700 Hz range

 

Yes.... a stronger high pass filter (or actually just adding one, if it doresn't have one) on the mid driver.... to get it's -6dB point somewhere around 400Hz looks about right.

 

.... and then change whatever filter is on the woofer to cross it over to the mid.

Link to comment
Share on other sites
23 hours ago, Ken Arrowsmith said:

As I recall you are correct the measurements were taken 1m on axis to the tweeter.. are you implying the measurements should be on axis for each driver? That would make sense to me. How then do we measure the two paralleled mids....? On the tweeter axis would be equidistant from the centre of the two cones...?

 

I'm really just saying that what the measurements might appear to show aren't necessarily what we could take as conclusive.

 

The mids will be beaming - giving a very strong directional output at that mike position, which may be quite different a few degrees to one side. And the woofer output, being so off-axis vertically, will be showing lower output, which would be stronger at 1m, on axis. But if you measure all with the mike down there, you'll be getting comb-filtering from the 2 mids, which are now at differing distances from the mike, not to mention the tweeter. And a different apparent phase error. It's why I've never been interested in MTM. Your head is forever stuck on tweeter axis (vertically). Funny that it's called the D'Appolito arrangement, and he's the guy who wrote the highly regarded bible on speaker measuring.

 

Hoisting them into the sky and measuring the whole caboodle from the designed listening distance with long gating would give a more real picture.

 

With the sensitivity thing, you might be able to solve both problems by doubling the woofers, rather than mids.

 

The real help from this community will start after you show pics (yeah!) or say if the drivers are already boxed, give model numbers for all drivers, and some info about the crossover.

 

Could be informative and fun for all. (It's a bit frustrating having to guess).

Link to comment
Share on other sites

Great to have all your feedback, I originally posted this as I am trying to help out a friend on this project. I had better let him know and if he is amenable I will provide a lot more detailed information, I appreciate it is difficult to help out without all the info..

 

Cheers,

 

Ken

Link to comment
Share on other sites

 Share

  • Recently Browsing   0 members

    No registered users viewing this page.





×
×
  • Create New...
To Top