Hey Gable and thanks for the help 
Regarding rear mounting, the TD8Ms front panel is not flat and has a angled frame, the TD15Xs have a rubber thing going around the frame which is not designed to be mounted behing the baffle, so rear mount is not an option, we may only need the front or flush mount, which will reduce the needed inserts
Regarding rear mounting, the TD8Ms front panel is not flat and has a angled frame, the TD15Xs have a rubber thing going around the frame which is not designed to be mounted behing the baffle, so rear mount is not an option
, we may only need the front or flush mount, which will reduce the needed inserts
Last edited:
Gable, there is good measurement preparation paper by kimmosto, he recommends ARTA https://kimmosaunisto.net/Software/VituixCAD/VituixCAD Measurement Preparations.pdf
Hi pcgab, I'm planning to try RiTA software in near future, it's pretty close to Smaart and is freeware. It may worth a try.
Below there is a video that you can learn a little about it, the Youtuber also offer some videos comparing some softwares including RiTA and other famous.
The website is in Spanish but the software is in English.
YouTube
RiTA – EducaSound
Hi Gable,
Great that you will start with IEC baffle!
Here already some tips. Please ask for the info you are missing.
I have built a 225 Hz IEC baffle, made it in two parts and screwed together at the backside with some wooden boards. Easy to remove and replace. My baffle is placed a little above the ground in a way there is the same distance from ground, walls and ceiling for the longest FFT window without reflections.
For the inserts, in my case I only do flush mount. It is some work to make, but the most accurate without baffle diffractions and compatible with the in cabinet baffle mounting. In that way the drivers are all flush with the IEC baffle. You can leave the micro at same position on axis for all drivers. Interesting for AC measurement also.
As measurement, mls is recommended of course. 2 – 3 times repetitive is more than enough.
REW, I don’t know so good, but sometimes I use it to see measurements. Not possible to read absolute time of impulse start, I think. In REW I find it more difficult to set the start of the FFT window at the impulse start. In my Clio I have more resolution. It is important to set this time accurate to obtain the correct phase at the AC center. All by all I think REW can be used, it is also well known.
Maybe others can give you extra tips, also to measure AC.
No experience with Arta.
Measurement distance: 1m. 3m will be difficult in a room I suppose w.r.t. reflections.
Off axis: steps of 15 degrees in horizontal plane (one side) is ok. 15, 30, 45, 60 degrees. If possible to do 75 and 90 in your room, also welcome. Then I can calculate DI in a single hemisphere for drivers with the same horizontal and vertical response (midrange and woofer). But no really need for 75 and 90 degrees. Vertical off axis measurement in a room will be difficult probably. If possible at the upper side, nice. Then we can measure the chosen CD tweeter with horn also in the vertical direction. It has different horizontal and vertical off axis responses.
Voltage, minimum 1Vrms and normalizing to 2.83 Vrms is ok. For tweeter, start frequency 200 Hz, not lower.
Best to measure also the impedance on IEC. The results are a little different than free air, but more conform the infinite baffle response of the driver.
Nearfield measurements are interesting to check TSP. I do it with slow little sine wave sweep at 5 mm of the cone.
DC resistance of voice coil with accurate multimeter. Good to have a calibrated resistor around 10 – 20 Ohm to compare with.
Great that you will start with IEC baffle!
Here already some tips. Please ask for the info you are missing.
I have built a 225 Hz IEC baffle, made it in two parts and screwed together at the backside with some wooden boards. Easy to remove and replace. My baffle is placed a little above the ground in a way there is the same distance from ground, walls and ceiling for the longest FFT window without reflections.
For the inserts, in my case I only do flush mount. It is some work to make, but the most accurate without baffle diffractions and compatible with the in cabinet baffle mounting. In that way the drivers are all flush with the IEC baffle. You can leave the micro at same position on axis for all drivers. Interesting for AC measurement also.
As measurement, mls is recommended of course. 2 – 3 times repetitive is more than enough.
REW, I don’t know so good, but sometimes I use it to see measurements. Not possible to read absolute time of impulse start, I think. In REW I find it more difficult to set the start of the FFT window at the impulse start. In my Clio I have more resolution. It is important to set this time accurate to obtain the correct phase at the AC center. All by all I think REW can be used, it is also well known.
Maybe others can give you extra tips, also to measure AC.
No experience with Arta.
Measurement distance: 1m. 3m will be difficult in a room I suppose w.r.t. reflections.
Off axis: steps of 15 degrees in horizontal plane (one side) is ok. 15, 30, 45, 60 degrees. If possible to do 75 and 90 in your room, also welcome. Then I can calculate DI in a single hemisphere for drivers with the same horizontal and vertical response (midrange and woofer). But no really need for 75 and 90 degrees. Vertical off axis measurement in a room will be difficult probably. If possible at the upper side, nice. Then we can measure the chosen CD tweeter with horn also in the vertical direction. It has different horizontal and vertical off axis responses.
Voltage, minimum 1Vrms and normalizing to 2.83 Vrms is ok. For tweeter, start frequency 200 Hz, not lower.
Best to measure also the impedance on IEC. The results are a little different than free air, but more conform the infinite baffle response of the driver.
Nearfield measurements are interesting to check TSP. I do it with slow little sine wave sweep at 5 mm of the cone.
DC resistance of voice coil with accurate multimeter. Good to have a calibrated resistor around 10 – 20 Ohm to compare with.
Nothing wrong with MLS, but I changed to sine sweeps, because you can set the bandwidth limits of the test signal.
MATAA
(Of course this is the very best audio testing software that ever existed ;-), but I admit it's not easy to set up if you don't know what you're doing.)
How do you set start frequency with an MLS? Do a high pass filter to the MLS signal first?
Also, I never really how signal levels are set... I believe it's trivial, but some of my synapses seem to run in circles with this. Will the measurement results be the same in terms of SPL level if different test signal forms (with different spectra) are used as long as the signals all have the same RMS voltage (MLS, sine sweep, white/pink/red noise, etc.)?
Matthias,
How stupid can I be, of course with a mls measurement, frequency start cannot be set to 200 Hz. I meant for the tweeter impedance measurement with a swept sine, to limit the low frequency to 200 Hz. And also not more than 1 Vrms on a tweeter.
Concerning the relation, voltage and SPL with different measurements. With the Clio I measure different SPL levels at the same rms output voltage for swept sine and mls. I don’t know anymore which one is higher or lower. I think it is related to power level and bandwidth of noise sources. I have no answer for it at this moment. I will have a look at later and do some tests again with my Clio.
How stupid can I be, of course with a mls measurement, frequency start cannot be set to 200 Hz. I meant for the tweeter impedance measurement with a swept sine, to limit the low frequency to 200 Hz. And also not more than 1 Vrms on a tweeter.
Concerning the relation, voltage and SPL with different measurements. With the Clio I measure different SPL levels at the same rms output voltage for swept sine and mls. I don’t know anymore which one is higher or lower. I think it is related to power level and bandwidth of noise sources. I have no answer for it at this moment. I will have a look at later and do some tests again with my Clio.
REW supports 2-channel measurements now, with loopback wire for timing reference. But of course it's priorities are for room response analysis. It's impulse response export choices are not always suitable for simulators, most of which prefer MLS based formats.
Other choice is Holm Impulse (free) or OmniMic (registration fee to get full features)
Other choice is Holm Impulse (free) or OmniMic (registration fee to get full features)
Why would it matter if the impulse response was measured with an MLS or other test signal? An impulse response is an impulse response is an impulse response...
In attach a proposal for a plan to realize the Tower XL prototype.
Headlines are:
1. IEC baflle measurement of all drivers
SPL on and off axis, impedance and AC (acoustical center)
2. Make infinite baffle driver models using the IEC baffle measurements
To compare all responses with datasheet values. Calculate the AC offsets.
3. In cabinet measurement of all drivers
SPL on and off axis, impedance
4. In cabinet response simulations of all drivers
To compare with the “in cabinet measurements”. Using Leap or other software.
5. Create SPL and impedance curves (and frd’s) for X-over design
Using TSP data and measured cabinet responses and do some splicing. To be done in Leap or other software.
6. Design of X-over
To choose filter concepts some digital designs can be done at first, using miniDSP or other digital processor.
Starting with a passive filter is possible too.
To be done in Leap or other software.
7. In cabinet measurement of filtered drivers
SPL on and off axis axis
8. In cabinet measurement of the sum response
SPL on and off axis
9. Listening, tweaking with measurement checks
10. Build passive filter if not done yet.
- Check SPL on axis of each driver and the sum response, impedance at input of X-over
- listening test.
More details about the measurement conditions and contents in attached document.
Comments are welcome.
Headlines are:
1. IEC baflle measurement of all drivers
SPL on and off axis, impedance and AC (acoustical center)
2. Make infinite baffle driver models using the IEC baffle measurements
To compare all responses with datasheet values. Calculate the AC offsets.
3. In cabinet measurement of all drivers
SPL on and off axis, impedance
4. In cabinet response simulations of all drivers
To compare with the “in cabinet measurements”. Using Leap or other software.
5. Create SPL and impedance curves (and frd’s) for X-over design
Using TSP data and measured cabinet responses and do some splicing. To be done in Leap or other software.
6. Design of X-over
To choose filter concepts some digital designs can be done at first, using miniDSP or other digital processor.
Starting with a passive filter is possible too.
To be done in Leap or other software.
7. In cabinet measurement of filtered drivers
SPL on and off axis axis
8. In cabinet measurement of the sum response
SPL on and off axis
9. Listening, tweaking with measurement checks
10. Build passive filter if not done yet.
- Check SPL on axis of each driver and the sum response, impedance at input of X-over
- listening test.
More details about the measurement conditions and contents in attached document.
Comments are welcome.
Well, look at the file and it's extension name.
REW adds lots of info about recording in the first 10 or so lines. Many programs don't identify it as impulse data file, unless the data/text is edited. There are also many options about how data is manipulated before export as .txt. These parameters/settings seem to vary between programs.
I have had lots of difficulties exporting/importing measurement data between REW, ARTA, OmniMic and HolmImpulse. (REW cant make directivity in spectrogram style like ARTA and OmniMic.)
Last edited:
Aatto,
It is just a plan proposal to do it. With these measurements, we are able to design a technical good speaker. And hopefully the sound will be the same.
Do you have the plan to make the first X-over digital? I know miniDSP very well and how to program biquads. It will allow you to listen to the difference between high and low order filtering. Very interesting, I already did it in the past. And we design digital versions that can be built passive also, so no AC compensation for these.
Later on we can design more optimized digital versions too. Like 5th order elliptical for example and other stuff with AC compensation of course.
It is just a plan proposal to do it. With these measurements, we are able to design a technical good speaker. And hopefully the sound will be the same
.Do you have the plan to make the first X-over digital? I know miniDSP very well and how to program biquads. It will allow you to listen to the difference between high and low order filtering. Very interesting, I already did it in the past. And we design digital versions that can be built passive also, so no AC compensation for these.
Later on we can design more optimized digital versions too. Like 5th order elliptical for example and other stuff with AC compensation of course.
Paul
this is a good plan and makes sense, if this helps we can try some settings with miniDSP and then based on that we can go for the passive version, I will build the passive version anyways as I mentioned I want a passive system to test amps, so either way works for me.
from my limited experience and knowledge, I was thinking that we may be better to build the passive version first and build the speaker around passive XO since tweaking and changing the DSP is not a problem but not the other way around, but either way works for me.
but eventually we will have a DSP version too, I have no experience in biquads but i like to hear the elliptical XO.
this is a good plan and makes sense, if this helps we can try some settings with miniDSP and then based on that we can go for the passive version, I will build the passive version anyways as I mentioned I want a passive system to test amps, so either way works for me.
from my limited experience and knowledge, I was thinking that we may be better to build the passive version first and build the speaker around passive XO since tweaking and changing the DSP is not a problem but not the other way around, but either way works for me.
but eventually we will have a DSP version too, I have no experience in biquads but i like to hear the elliptical XO
.Yesterday afternoon was productive on this end. I picked up all the plywood for the IEC baffle, cut the 4 main panels, and cut most of the bracing/connector plates, etc. to assemble the baffle.
I cut and glued up the inserts, and will trim them to final size and prep them to be routed for their respective drivers today.
Aatto is coming down to my place later this afternoon with the drivers. We will get the cutout and flush mount routed for each driver, and add the mounting tabs to the inserts.
I am hoping we have enough time to plan and document our testing strategy, setup and procedures. I would like to have folks review that in this thread.
The engineer in me wants to have a formalized plan before executing, haha.
This is for the 175 hz IEC baffle.
Cheers,
Gable
I cut and glued up the inserts, and will trim them to final size and prep them to be routed for their respective drivers today.
Aatto is coming down to my place later this afternoon with the drivers. We will get the cutout and flush mount routed for each driver, and add the mounting tabs to the inserts.
I am hoping we have enough time to plan and document our testing strategy, setup and procedures. I would like to have folks review that in this thread.
The engineer in me wants to have a formalized plan before executing, haha.
This is for the 175 hz IEC baffle.
Cheers,
Gable
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.