Z Axis caluculation help

[jimbones]

Soon I will be attempting to use PCD and/or Dayton OmniMic to determine time aligning drivers of my 3 way (Lambda) speaker system.

I've read some about it. I understand parts of the process but not end-to-end. If there is anyone that has done this and can share their experiences and wouldn't mind guiding me through this via PM/emails I would be grateful.

Thanks,
Jim

 

[Josephjcole]

I imagine it would educational for people, to have the discussion open on the forum. I've worked a little with PCD, but would enjoy reading about others experiences.

Do you have any specific questions to start off with?

 

[robotbunny]

Hey Jim,

I'm only doing this with a 2-way in mind, I have not done a 3-way so, ymmv. I'll guess that you'd need a summed response for all three drivers anyway, so the process should be the same. The easiest method for finding acoustic offset through PCD is to use the overlay feature using the summed response of the drivers. Take frequency response measurements of the individual drivers, then wire the drivers in parallel for the summed response. Make sure phase data is in the frd otherwise it won't import into pcd. Use the import overlay feature and choose the saved frd of the summed response.

Next you want to physically measure the distance from center of tweeter to the center of woofer and convert to meters. This value should be entered into PCD as the woofer's vertical offset, example: 2" from center of tweeter to woofer = 0.051, enter (negative) -0.051

Once you begin acoustic measurements, don't move the mic and measure the distance from mic to tweeter. You'll need to enter this distance in meters into PCD.

You'll see the wavelengths and begin adjusting z offset until the two images converge.

Edit: J. Bagby wrote it up here and it works perfectly: https://www.box.com/shared/ouxjjsx0m8bs00cil5iq

 

[jimbones]

WHAT?! You mean I have to expose my ignorance about speaker building in an open forum?! Ha ha!!!! Thats fine.

My system is RS270, Peerless 830833 mid, SB29 tweeter. The Mid os OB. XO at 275 and 2500. OK if I understand this correctly I start with all my physical measurements (driver location on baffle) entered into PCD. i then take individual Fr measurements of each driver alone. Question: with or without XO in line?

Next I am supposed to play the mid/tweeter together (again with or without XO). We can discuss other steps later.

Thanks

 

[robotbunny]

No crossover. You're going to simulate one in PCD eventually, unless it's just the z-axis and offset you're after. You may also need to have some z-axis correction if the tweeter of woofer is further out (recessed or not). Just make sure it's all in meters/mm.

If you follow Bagby's sheet exactly, it's a breeze.

 

[dumptruck]

If you're using PCD, what are you doing with a crossover already existing before measurements? No crossovers. If you're trying to determine your phase alignment after an existing crossover, take a set of vertical polar measurements instead. There's no need for PCD.

Anyway, I've been doing it like Bagby's writeup there. Results match the sims every time. Only thing to watch out for is that if you're doing a 3-way (or just a large 2-way), you need to measure from pretty far to keep from being too far off-axis from any driver, OR follow up the first three measurements with on-driver-axis measurements to use, after you determine the z-offset.

For three drivers, after measuring each from a single mic location, you can try measuring all three at once wide open (you'd need to use 2-3 amp channels of course), or you can do the TM and MW pairs separately. The latter will be worth the extra few minutes, I'd say.

 

[jimbones]

Z axis and PCD

No crossover. You're going to simulate one in PCD eventually, unless it's just the z-axis and offset you're after. You may also need to have some z-axis correction if the tweeter of woofer is further out (recessed or not). Just make sure it's all in meters/mm.

If you follow Bagby's sheet exactly, it's a breeze.

I'm glad you warned me about mm. I designed and fabricated the speaker using PCD. Now I want to do the z axis. I'll bypass the XO. I dont care if I need to make XO adjustments later. BTW, I found PCD to be very good for the XO design.

 

[jimbones]

Z axis

If you're using PCD, what are you doing with a crossover already existing before measurements? No crossovers. If you're trying to determine your phase alignment after an existing crossover, take a set of vertical polar measurements instead. There's no need for PCD.

Anyway, I've been doing it like Bagby's writeup there. Results match the sims every time. Only thing to watch out for is that if you're doing a 3-way (or just a large 2-way), you need to measure from pretty far to keep from being too far off-axis from any driver, OR follow up the first three measurements with on-driver-axis measurements to use, after you determine the z-offset.

For three drivers, after measuring each from a single mic location, you can try measuring all three at once wide open (you'd need to use 2-3 amp channels of course), or you can do the TM and MW pairs separately. The latter will be worth the extra few minutes, I'd say.

I used the Fr plots form the manufacturer and traced using SPL Trace. I followed you for everything except the vertical polar measurement. Could you clarify? Since mine is a 3 way it seems pertinent.

 

[dumptruck]

Oh, if you're not measuring the response yourself, you can't determine the z-offset, period.

 

[wolf_teeth]

Oh, if you're not measuring the response yourself, you can't determine the z-offset, period.

Not to an exact point, no, but you can estimate it.... See first blog post below, and note that fudging forward 5mm or so on longer throw woofers can be beneficial.

Some optimization criteria... and some tips... - Blogs - Techtalk Speaker Building, Audio, Video Discussion Forum
Tips, Volume 2- The setup!! - Blogs - Techtalk Speaker Building, Audio, Video Discussion Forum!!
"Tips! Volume 3"... - Blogs - Techtalk Speaker Building, Audio, Video Discussion Forum
Tips! Volume 3, Section B.. Series Xovers.... - Blogs - Techtalk Speaker Building, Audio, Video Discussion Forum
'Tips! Volume 4', minimizing, evaluating, and rechecking a simulation... - Blogs - Techtalk Speaker Building, Audio, Video Discussion Forum


Later,
Wolf

 

[jimbones]

OK, let me clarify. For the original build without any z axis calculations, I used the stock Fr graphs as provided from the manufacturer. After I used various programs (Edge, SPL TRace etc) I built up the XO and measured and readjusted the XO (Omnimic was used for this.

Now that I want to do z axis i understand I need to measure individual drivers sans XO. (I read that it is advisable to put a cap on the tweeter though, just not sure what value). I will start by measuring each on thier own on the stock baffle. Is this correct?

 

[Josephjcole]

Now that I want to do z axis i understand I need to measure individual drivers sans XO. (I read that it is advisable to put a cap on the tweeter though, just not sure what value). I will start by measuring each on thier own on the stock baffle. Is this correct?

Yep, set your speaker up, set your mic up, measure each driver seperately, then measure all the drivers together without the crossover. Don't move the speaker or the mic when measuring the different drivers.

I don't know about a cap on your tweeter. I would think you'ld be alright as long as the volume isn't to high. Certainly when you measure it on it's own you can start the measurements around 500Hz or so. I'll defer to an expert on this one though.

 

[jimbones]

Yep, set your speaker up, set your mic up, measure each driver seperately, then measure all the drivers together without the crossover. Don't move the speaker or the mic when measuring the different drivers.

I don't know about a cap on your tweeter. I would think you'ld be alright as long as the volume isn't to high. Certainly when you measure it on it's own you can start the measurements around 500Hz or so. I'll defer to an expert on this one though.

Hmmm all drivers together..without XO. Can't put them all on one channel in parallel so what is the connectivity options? One on each amp channel only gives me 2. measure 2 at a time? woofer/mid and mid tweeter?

 

[dumptruck]

2 at a time will give you more clear results when you take things into PCD anyhow, so go with that. I don't know what your layout looks like, but I'd probably start with the mic positioned on the midpoint between the high and mid centers. Is your mid driver a sealed-back type?

 

[jimbones]

2 at a time will give you more clear results when you take things into PCD anyhow, so go with that. I don't know what your layout looks like, but I'd probably start with the mic positioned on the midpoint between the high and mid centers. Is your mid driver a sealed-back type?

Open back mid (Hybrid OB design, I also have a rear facing tweeter as well)

The mid-tweet are positioned close together (1 inch apart) so if I position the mic exactly between them it should be good I guess.

 

[jimbones]

Oh crap, I just realized I have to open the box and connect to the woofer directly. Can't go through LP filter on woofer I guess

 

[jimbones]

OK so I measured the mid and tweeter and saved the curves with the phase. Tomorrow I will do both together, then what?

 

[dumptruck]

Hopefully you didn't move anything at all in between measurements. You load them up in PCD, type in all the physical distances (except Z), load the response of both drivers at once as an overlay, and adjust Z until it matches. Also post your files and distances here if you'd like help.

 

[robotbunny]

You're all over it dumptruck. Great username, btw. Ya, as long as the constants remain unchanged the data is useful. Changing mic distances in any axis between sessions can correlate to inaccurate data and if the wavelengths of each driver are not timed together from the summed response, you've got distortion.

Just follow that paper Jeff wrote, and you'll be golden. Once the data is in PCD, the heavy lifting is over.

Yes, please post some graphs.

 

[jimbones]

Found Z Paper

You're all over it dumptruck. Great username, btw. Ya, as long as the constants remain unchanged the data is useful. Changing mic distances in any axis between sessions can correlate to inaccurate data and if the wavelengths of each driver are not timed together from the summed response, you've got distortion.

Just follow that paper Jeff wrote, and you'll be golden. Once the data is in PCD, the heavy lifting is over.

Yes, please post some graphs.

I found the Z paper you are referring to (pun intended ) I will do the measurements over. question; do I work the midrange setback fist then the tweeter or is it a all at once thing?