Hello!
I have never built any speakers, but I have read a lot about the subject and have practiced a lot in VituixCad and WinISD.
Now it has come to the point that I want to build my own 3-way speakers. I have a Umik-1 and DATS v3.
I know that I should have bought an XLR measurement microphone with a small amplifier for measurements with loopback, but now I have what I have and thought to run with it.
I was thinking of doing the measurements with 'acoustic timing reference' in REW. This using a small dome tweeter in a fixed position from the mic, and only move the speakers during the measurement process.
I thought to do according to the following steps (I will edit the steps gradually)
1. Design a ported speaker cabinet with WinISD.
2. Build the box with 19 or 22 mm MDF and 3D-print the port with flares.
3. Measure the tweeter on-axis with a distance of 1 m (set Impulse/Actions/Estimate IR Delay and "shift and update timing offset"). Continue with the rest of the measurements with the same settings. Repeat for off-axis in steps of 15° hor and vert. Use the same volume/settings throughout all measurements.
4. Measure the woofer's near field on-axis (5 mm away from the cone), measure nearfield of the port, measure woofer on-axis at 1 m distance (gated measurement to minimize room reflections), repeat measurements for off-axis in steps of 15° hor and vert.
5. Measure the midrange on-axis with a distance of 1 m and repeat for off-axis in steps of 15° hor and vert. Measure tweeter and woofer in paralell, so I can confirm the delay later on.
6. Merge the woofer's nearfield+port with farfield using VituixCad.
7. Measure the impedance of each mounted driver with DATS v3.
8. Import the measurements into VituixCad.
9. Set the listening position on Y-axis for mid and woofer (distance from the tweeter). The drivers are centered on the baffle on X-axis. Import and compare the tweeter+woofer parallel response and double check if drivers are set properly.
10. Design the crossover, order the components, build the crossover outside of the speaker cabinet.
11. Make test measurements. Make adjustments if needed. Finalize the build.
Questions I have:
Have I got the steps right or am I missing anything?
What type of stuffing material and how much to get rid of standing waves? I know it's a subject with many different opinions. How do you do it?
I have never built any speakers, but I have read a lot about the subject and have practiced a lot in VituixCad and WinISD.
Now it has come to the point that I want to build my own 3-way speakers. I have a Umik-1 and DATS v3.
I know that I should have bought an XLR measurement microphone with a small amplifier for measurements with loopback, but now I have what I have and thought to run with it.
I was thinking of doing the measurements with 'acoustic timing reference' in REW. This using a small dome tweeter in a fixed position from the mic, and only move the speakers during the measurement process.
I thought to do according to the following steps (I will edit the steps gradually)
1. Design a ported speaker cabinet with WinISD.
2. Build the box with 19 or 22 mm MDF and 3D-print the port with flares.
3. Measure the tweeter on-axis with a distance of 1 m (set Impulse/Actions/Estimate IR Delay and "shift and update timing offset"). Continue with the rest of the measurements with the same settings. Repeat for off-axis in steps of 15° hor and vert. Use the same volume/settings throughout all measurements.
4. Measure the woofer's near field on-axis (5 mm away from the cone), measure nearfield of the port, measure woofer on-axis at 1 m distance (gated measurement to minimize room reflections), repeat measurements for off-axis in steps of 15° hor and vert.
5. Measure the midrange on-axis with a distance of 1 m and repeat for off-axis in steps of 15° hor and vert. Measure tweeter and woofer in paralell, so I can confirm the delay later on.
6. Merge the woofer's nearfield+port with farfield using VituixCad.
7. Measure the impedance of each mounted driver with DATS v3.
8. Import the measurements into VituixCad.
9. Set the listening position on Y-axis for mid and woofer (distance from the tweeter). The drivers are centered on the baffle on X-axis. Import and compare the tweeter+woofer parallel response and double check if drivers are set properly.
10. Design the crossover, order the components, build the crossover outside of the speaker cabinet.
11. Make test measurements. Make adjustments if needed. Finalize the build.
Questions I have:
Have I got the steps right or am I missing anything?
What type of stuffing material and how much to get rid of standing waves? I know it's a subject with many different opinions. How do you do it?
Last edited:
@AllenB - You know a lot more than I do, but are you sure that is correct? I thought (1) with timing reference you would keep z offset at zero for all three (2) If all three drivers are centered on the baffle you would keep x offset zero for all three (but not if any of the drivers were offset horizontally) (3) but Y-offset will NOT be zero for the midrange and woofer because each was measured on the driver axis but the design axis is the tweeter?
I think you have this right, but just to be clear, the port is also measured near field
yes, and just to be clear, midrange and woofer impedance are in the actual box, not free-air.
While taking test measurements and listening, it is common to leave the crossover outside of the cabinet and run wires to the terminal jacks and out the port as needed. Of course you CAN build it and put it inside the cabinet, but if you are just tweaking one component here or there, it is a pain to have to get inside the cabinet each time.
I think you have this right, but just to be clear, the port is also measured near field
I personally would not measure in steps of 5 degrees. You won't gain much. In fact you don't gain much from doing 10 degrees instead of 15 or 20. And being your first time measuring, you will likely make a mistake and have to redo measurements. (Not to assume you will mess up, but I've designed several speakers using VituixCAD and often mess something up and have to re-do my measurements.)
Yes, but really just one step. VituixCAD will do the on-axis and off-axis measurements all at the same time.
yes, and just to be clear, midrange and woofer impedance are in the actual box, not free-air.
While taking test measurements and listening, it is common to leave the crossover outside of the cabinet and run wires to the terminal jacks and out the port as needed. Of course you CAN build it and put it inside the cabinet, but if you are just tweaking one component here or there, it is a pain to have to get inside the cabinet each time.
VituixCAD. Use the Merge tool, it is awesome.
Indeed, and this does need a better explanation on my part. Firstly I agree that z offset does stay at zero if you time them properly.
If you have a horizontal offset then it gets complicated because to do it right, you need to measure at the listening distance or compensate accordingly.. otherwise you have a conflict over rotating the speaker around the tweeter centreline or the horizontal offset.
As for vertical shift, that depends on whether you move the mic in height for each driver. This naturally is going to produce the slightest error between each, but it's usually accepted as insignificant.
Of course, Vituixcad can be applied to take advantage of using these offsets for lobing purposes, you just need to ensure you don't double account for these factors.
(I don't have time to double check what I've written just now)
A simple way to check if you have timing problems is to measure the tweeter say at 50cm in the middle between woofer and tweeter.
Then measure the woofer.
And the last one measure the woofer and tweeter together.
All these measurements you never move the microphone.
Then load the woofer and tweeter measurements into Vituixcad and load the measurement of the combined woofer and tweeter as a reference.
Now you can adjust the delay to match the reference.
Then your ready to design a filter in Vituixcad.
Rob
Then measure the woofer.
And the last one measure the woofer and tweeter together.
All these measurements you never move the microphone.
Then load the woofer and tweeter measurements into Vituixcad and load the measurement of the combined woofer and tweeter as a reference.
Now you can adjust the delay to match the reference.
Then your ready to design a filter in Vituixcad.
Rob
I'd rather use 20 mm MDF 😊
And make sure to use generous flares when 3d printing the port - also on the inside!
Okay, even if the measurements of the drivers is not taken from listening position?
I thought z offset didn't need to be adjusted only x and y, if the measurements were done on-axis of each driver.
Thanks for the reply, I understand what you explained regarding z, x and y. That with timing reference, I only specify x and y but not z, as I have taken each measurement on-axis on each driver. This is to create the listening position from the tweeter. I explained it a little poorly in my steps.
Yes the port will be measured nearfield, just before the beginning of the flaring of the port (where the port is even circumference).
I thing I will do steps of 15° instead then, thanks 🙂
Thats great that the merging can be done in one step!
The impedance will be measured with the woofer mounted. The midrange and tweeter is sealed, so it does not matter.
Regarding the crossover, that sounds smart 🙂 thanks for all the tips!
Thanks, I will have this in mind. The drivers will be centered horisontionally, so I will only enter the vertical distance from the tweeter in VituixCad. The speaker/drivers will be positioned at the same height as the mic during the measurements. 🙂
Stv's post regarding 20mm was just a joke...you have a typo in the original post that says 2mm... Pretty thin fit a speaker 😁
Member
Joined 2003
When using an acoustic timing reference, care must be taken to ensure distance from acoustic reference to mic remains constant at all times, in addition to the distance from mic to speaker baffle surface. When measuring multiple drivers, its probably easier to adjust the speaker height to bring the driver in line with the mic, as any movement in the mic location will require careful adjustment of acoustic reference location which will be a pain without some additional jig.
Start with the tweeter to define reference time location, high frequencies are beneficial to define the impulse peak for timing determination. VituixCAD measurement guide for REW will have you enter 1 meter distance to the measurement timing offset, since electrical loopback will represent actual time of flight from speaker to driver. With acoustic reference, timing offset is actual delay from acoustic reference to DUT, not the DUT to the mic. With your first measurement, go to the Impulse tab, select Actions -> Estimate IR Delay. Confirm delay value makes sense, and select "shift and update timing offset". This will offset all subsequent measurements to align t=0 based on delay of the first measurement. So don't move anything and continue with the rest of your measurements.
Apart from this, all other steps for measurement and merge process can be followed per the VituixCAD measurement guide for REW.
https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html#How_to_start_with_VituixCAD
Start with the tweeter to define reference time location, high frequencies are beneficial to define the impulse peak for timing determination. VituixCAD measurement guide for REW will have you enter 1 meter distance to the measurement timing offset, since electrical loopback will represent actual time of flight from speaker to driver. With acoustic reference, timing offset is actual delay from acoustic reference to DUT, not the DUT to the mic. With your first measurement, go to the Impulse tab, select Actions -> Estimate IR Delay. Confirm delay value makes sense, and select "shift and update timing offset". This will offset all subsequent measurements to align t=0 based on delay of the first measurement. So don't move anything and continue with the rest of your measurements.
Apart from this, all other steps for measurement and merge process can be followed per the VituixCAD measurement guide for REW.
https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_20.html#How_to_start_with_VituixCAD
Last edited:
Ok. This will allow you to take advantage of Vituixcad calculating lobing. It does still raise questions but realistically it is suitable.