DcibeL - My experience with an SB34NRX75-6 in a sealed box is similar to yours. I did not keep the data, but the response was the same at 0.1 V (nearfield) and at 4 V (outdoor ground plane). Both measurements matched the predicted response to within +/- 1 dB. The prediction was based on measured T/S parameters using DATSv3. The DATS system uses about 0.1 V.
I know that with large signal levels, Fs changes, Qes changes, Vas changes. So therefore I can only assume that the changes somehow counteract each other such that the response stays constant.
I know that with large signal levels, Fs changes, Qes changes, Vas changes. So therefore I can only assume that the changes somehow counteract each other such that the response stays constant.
It turns out I did keep the data. See below. I am showing the nearfield low-level signal which is adjusted to 4-pi, which very closely matched the predicted response. The VituixCad predicted response is based on the small signal T/S parameters measured by DATSv3. I am also showing the outdoor ground plane measurement which is equivalent to anechoic, and it requires a high signal level. I was quite pleased with how closely the two measurements matched, and I feel any differences are within the measurement tolerance.
Now we should keep in mind that this is a sealed box system, which means it is more tolerant of driver variations than a vented box or PR system.... but still, it is a nice match.
j.
Now we should keep in mind that this is a sealed box system, which means it is more tolerant of driver variations than a vented box or PR system.... but still, it is a nice match.
j.
Attachments
Member
Joined 2003
Precisely. You must compare the box model completely to determine if the driver is truly “off spec” rather than focus on any individual parameter. I’ve mentioned enough times, T/S is defined as low level parameters, so if your goal is to determine T/S I would expect the low level result to be the more correct approach even if it’s not what the manufacturer did to create the data sheet.
There are perfectly good reasons to measure a driver at elevated voltage and compare performance at various amplitudes, but determining the cabinet model isn’t one of them for any driver I’ve tested.
Member
Joined 2003
So I managed to get a hold of a couple RND samples from SB Acoustics. They are very similar to SB17NBAC35-4 (4 ohm), but came with test data from SB for the specific driver. So what I have is not just a datasheet for a specific model, but manufacturer test data for this exact driver.
Reviewing the data and running a few sweeps, I was able to reproduce the manufacturer impedance curve with 1V sine sweep. I measured again at 100mV, and observed a change in Fs by 5Hz - 33Hz at 1V, 38Hz at 100mV.
I wanted to obtain a full set of parameters at 1V, but at the elevated excursion, my added mass method wasn't holding, and I didn't want to leave any permanent marks on the cone, so I will place some confidence in the manufacturer data, after all, the data is for this exact driver, not just a generic datasheet.
I completed T/S at only 100mV. Comparing to the manufacturer data, Fs was higher, Qts higher, and VAS lower. So what does that mean for our box model ? Measured data is shown as the dashed blue line, factory data is shown as solid blue.
Reviewing the data and running a few sweeps, I was able to reproduce the manufacturer impedance curve with 1V sine sweep. I measured again at 100mV, and observed a change in Fs by 5Hz - 33Hz at 1V, 38Hz at 100mV.
I wanted to obtain a full set of parameters at 1V, but at the elevated excursion, my added mass method wasn't holding, and I didn't want to leave any permanent marks on the cone, so I will place some confidence in the manufacturer data, after all, the data is for this exact driver, not just a generic datasheet.
I completed T/S at only 100mV. Comparing to the manufacturer data, Fs was higher, Qts higher, and VAS lower. So what does that mean for our box model ? Measured data is shown as the dashed blue line, factory data is shown as solid blue.
Hi all, quick question...
When saving an Overlay of a FR2 dual-channel measurement, i cannot get the phase trace to save.
(This is trying to save a stabilized phase trace, after having accepted the Generalized Cross-correlation delay.)
Is there a way to get the phase trace saved as part of the Overlay?
Thanks, Mark
When saving an Overlay of a FR2 dual-channel measurement, i cannot get the phase trace to save.
(This is trying to save a stabilized phase trace, after having accepted the Generalized Cross-correlation delay.)
Is there a way to get the phase trace saved as part of the Overlay?
Thanks, Mark
Member
Joined 2003
I have a quick question regarding the cursor position at the start of FFT, if I use semi dual loopback and hit home key should I leave the left cursor at that(300 sample) and the other at where the gate time is optimal or should the left cursor be set most close to the start of the impulse response? If left at ref point it seems to somewhat enhance the the resolution of it...
Bluetak for mass? Or closed box method. No thing to mark up the cone on this method.
Member
Joined 2003
Do you want the entire time of flight included in the phase response or not? Generally I would place the start cursor just behind the impulse to minimize the excess phase. Key is to make sure mic distance and start location remain the same when measuring multiple drivers so that the difference if phase / time of flight is captured.
If you’re measuring for use with VituixCAD, just save the impulse response file directly and process through VituixCAD IR-FR which has a few more features than ARTA.
Last edited:
Member
Joined 2003
Just Blue tack, I wasn’t going to build a box just for this simple comparison, maybe eventually when I put the drivers to use, but I think I already know what the result of that comparison will be.
The point of the post was simply to illustrate that the voltage divider method of measuring impedance and T/S is “good enough” for the cabinet model, even if the individual T/S parameters aren’t exactly what’s on the manufacturer sheet.
I made this mistake the first time I conducted spinorama measurements. I set the start of the window just ahead of the impulse, but as I rotated through 0-90-180, the window no longer captured the peak.
To set the start of the window, I usually measure impulse responses at 0, 45, 90, 135 degrees. Then I find the impulse response that arrives earliest, and I set my window based on that earliest impulse response... Then I go through the full rotation measurements.
j.
I use different sized magnets when I'm working on some woofers. Ferrite is best. The Neodymium ones are to tenacious. Placed on points of the clock on opposite sides of the cone. They don't come loose and they don'e leave a mark. If you are worried about marks a thin fabric bonded to one side of the magnet is insurance for sure. They make zero difference in motor strength. Believe me. I do a lot of motor design. It has no bearing on the flux inside the gap being that far away.
Last edited:
Member
Joined 2003
That's why I suggest processing the data in VituixCAD, so you have a left window included and the problem you describe will not occur. Processing the full set of spacial data becomes as simple as pressing the "FAR 1" button, adjust right window and save all.
Without a left window, one must be diligent to set the window start behind the impulse that is closest to the mic.
Hi all!
Not sure this request for input belongs here or in a different thread, but here it goes.
I've been running ARTA with great success for some years now. Initially, with a modified EMU0404 (cleaning up its left channel), which has been a brilliant interface for this. But once I accidentally upgraded all my boot partitions to Win 10 on the computer doing duty at my bench, I've lost the ability to use the EMU. I think some users have had success with it running with some legacy drivers under Win 10, but not me. I've actually sold the interface in the meantime, and I am coming to regret it. But continued support in the future is a pretty important aspect for me.
After looking for a replacement - tried a few USB solutions - settled for an internal Lynx interface. It's pretty clean, and I thought I'd like the less bulk etc., but it's giving me some heartburn.
I am unable to use it decently with the external generator feature. I'm getting very very long settling times, to the point in some measurements, it never settles. At least not usefully.
The way I practically use it is hooked on the monitor outputs of my AP S1. Even though it has DSP, the ARTA FFT is so much better in some types of actual use due to its live reading. So the AP does all measurements in all scenarios, while the Lynx grabs live FFT readings from the monitor outputs at that normalized level. But to do this properly, I feel I need to use "external generator" when running FFT, and it just won't practically work. With the EMU, it was trivial.
Do I just need a different interface - and if yes, what's the state of art these days? I've been looking at a Clarett2+, but it's a bit spendy - or maybe I have my settings tweaked wrong in ARTA?
Thanks all for your input in advance.
Not sure this request for input belongs here or in a different thread, but here it goes.
I've been running ARTA with great success for some years now. Initially, with a modified EMU0404 (cleaning up its left channel), which has been a brilliant interface for this. But once I accidentally upgraded all my boot partitions to Win 10 on the computer doing duty at my bench, I've lost the ability to use the EMU. I think some users have had success with it running with some legacy drivers under Win 10, but not me. I've actually sold the interface in the meantime, and I am coming to regret it. But continued support in the future is a pretty important aspect for me.
After looking for a replacement - tried a few USB solutions - settled for an internal Lynx interface. It's pretty clean, and I thought I'd like the less bulk etc., but it's giving me some heartburn.
I am unable to use it decently with the external generator feature. I'm getting very very long settling times, to the point in some measurements, it never settles. At least not usefully.
The way I practically use it is hooked on the monitor outputs of my AP S1. Even though it has DSP, the ARTA FFT is so much better in some types of actual use due to its live reading. So the AP does all measurements in all scenarios, while the Lynx grabs live FFT readings from the monitor outputs at that normalized level. But to do this properly, I feel I need to use "external generator" when running FFT, and it just won't practically work. With the EMU, it was trivial.
Do I just need a different interface - and if yes, what's the state of art these days? I've been looking at a Clarett2+, but it's a bit spendy - or maybe I have my settings tweaked wrong in ARTA?
Thanks all for your input in advance.
Hi b_force,
Yes, I am using the ASIO drivers. They show up as "ASIO Lynx." They are the only ASIO option I am seeing in the ARTA devices setup.
I am actually registering some success by playing with the Device Control Panel. For instance, disabling "double buffer output" seems to allow for quicker settling. I also set the "maximum channels" control to 2. The buffer size doesn't seem to have a lot of sway.
I am doing this by having the FFT run on essentially a closed loop (GEN MON on the AP). I guess it remains to be seen if this makes it workable in an actual measurement situation.
Thank you for your input.
Yes, I am using the ASIO drivers. They show up as "ASIO Lynx." They are the only ASIO option I am seeing in the ARTA devices setup.
I am actually registering some success by playing with the Device Control Panel. For instance, disabling "double buffer output" seems to allow for quicker settling. I also set the "maximum channels" control to 2. The buffer size doesn't seem to have a lot of sway.
I am doing this by having the FFT run on essentially a closed loop (GEN MON on the AP). I guess it remains to be seen if this makes it workable in an actual measurement situation.
Thank you for your input.
A question. I use ARTA together with a Xonar U7 USB soundcard (192kHz). Most of the times I do THD measurements at 1kHz, that works fine.
These days I used STEPS for an analysis up to 80kHz, and got some weird results after 20kHz. Then I connected my scope to the input of the Xonar and saw that after 20kHz there was no signal output anymore, still Arta was still working, stopping a bit later, when reaching the 80kHz.
I think there must be some steep HP filter in place, at about 20kHz. Someone knows that, and specially, someone knows how to remove that filter?
many thanks!
Erik
These days I used STEPS for an analysis up to 80kHz, and got some weird results after 20kHz. Then I connected my scope to the input of the Xonar and saw that after 20kHz there was no signal output anymore, still Arta was still working, stopping a bit later, when reaching the 80kHz.
I think there must be some steep HP filter in place, at about 20kHz. Someone knows that, and specially, someone knows how to remove that filter?
many thanks!
Erik