measure speaker T/S parameter, what do I need?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
1. what equipment do I need to measure T/S?, been building fullranger cabinets but seems published specs are freq overblown?

what is a minimum equipment I would need? (lower cost will be good).

2. what kind of SPL meter do I need to get the the speaker frequency/SPL graph produced?, some sort of mic-software combo?

thanks,
 
Most people use a Behringer ECM8000 mic and an external sound card with phantom power (sorry i don't remember the brand of it).

Alternative, I choose to use the cheapest possible solution since i do this just for fun.

So this is my setup.

  1. Laptop pc.
  2. Usb external soundcard, from ebay for approximately 15 $.
  3. Limp software for T/S measurements. (part of ARTA software package) free. http://www.fesb.hr/~mateljan/arta/
  4. A power amplifier. (although there is an option to measure without, i found this method more stable).
  5. A few cables, alligator clips etc.
  6. Patience.
I suggest you read carefully the manual of Limp before you start.

Now, If you like to make frequency response measurements as well, you will need the following.

  1. Self powered mic (uses one AA battery) with a 3.5 mm connector (i had this in stock).
  2. HOLMImpulse for frequency response measurements Free.HOLM Acoustics
  3. Alternative you can use "ARTA" but i found Holmipulse more user friendly.
  4. Mic stand. (i use my camera stand).
Have fun.
 
more suggestions

I've been tweaking/troubleshooting my setup for a few months now. Here is what I've settled on and what didn't work.

  • A PC running Speakerworkshop, available for free here: Speaker Workshop I'm sticking with SW because I invested so much pain & suffering into learning it, I'm afraid to switch to something else. If I were to do it all over again, I don't know that I would use SW.
  • A Behringer UCA222, which is essentially an external sound card which connects to the comp's USB port. I settled on this after failing to get SW to work with my on-board sound card, and with a Rocketfish card. SW requires 2 channel recording to use all of it's capabilities. Your comp's soundcard may be good enough to get started with SW or whatever software you choose. The Behringer is also nice because it uses RCA jacks for input and output which lets me use good quality cables. And the long USB cable lets me keep the signal-carrying wires short, yet position the computer further from the driver under test. If I were to do it again I would save a little money with the Behringer UCA202. BEHRINGER: UCA202
  • A power amp stable into at least 4 ohms. Theoretically, you only need 1 watt for frequency-response measurements, and a fraction of a watt (I use ~0.3VAC for 8 ohm drivers) for impedance curves. It's good to have a few extra watts.
  • A Wallin Jig, which is not available for purchase. Instructions for building one are here: Eric Wallin's Internet Homepage It's not necessary, but is convenient, especially if you want to calibrate your sound card in SW.
  • A little jig to hold drivers steady is nice, but not necessary.
  • If you want to do frequency-response measurements, add to this a microphone like the Behringer ECM8000, or the Dayton EMM-6. At least 2 people on this forum are using the ART Phantom I phantom microphone power supply.
  • Some way to add mass to the driver being tested, and a scale accurate to about 0.1g to measure said mass. Most people use some kind of reusable adhesive putty like this: Elmer's Tacky Adhesive can be Removed and Re-Used | Sticky Tack I've found at putty does not stick well to paper cones, and loses its stickiness after a few uses, so I've moved to small (~15g) magnets. Put them on opposite sides of the driver's cone, symmetrically around the dust cap. Also, I just weighed the magnets once and wrote their mass on them - very convenient.
  • Be aware that Parts Express is selling this little gadget: Dayton Audio WT3P WT3 Bundled with Precision Scale but I don't know anyone who's used it.
Enjoy!
-Byron
 

Attachments

  • impedance test setup.jpg
    impedance test setup.jpg
    115.2 KB · Views: 1,711
Just another Moderator
Joined 2003
Paid Member
Speaker Workshop with Claudio's cable jig is probably the cheapest entry into T/S measurements assuming you have a half decent sound card. as mentioned, if you have a laptop you will need a good usb sound card. Check out Claudio's web page here Claudio Negro's home page, how to project and test a loudspeaker using Speaker Workshop

For impedance measurements (for calculating T/S params) most modern motherboards on board sound should be sufficient, so if you have a modern computer (as opposed to a laptop) you probably only need to buy some cable and resistors to be on your way :)

Tony.
 
any particular powered mic?
No, I just use a self powered mic.

As ByronInLawrence said, you also need “Some way to add mass to the driver being tested”. Instead of magnets and a scale, I use coins and bluetak. You can find the exact weight of coins from google, no need to measure them yourself. If I want extra weight, I stack two or more coins together with bluetak in order to achieve the target weight. Then stick them all together to the driver with bluetak again. Works for me.
 
After reading a few web pages about using WM-61A as a test microphone I decided to try it.
I have Sony MDR-V6 headphones that are supposed to have a "flat" response from 5 Hz to 30K.
So I plugged them into the output of my sound card (Audiotrak Prodigy HD2) and put the mic close to the center of the speaker. The response is anything but flat: severe loss at both low and high ends. Any ideas?
 
Just another Moderator
Joined 2003
Paid Member
After reading a few web pages about using WM-61A as a test microphone I decided to try it.
I have Sony MDR-V6 headphones that are supposed to have a "flat" response from 5 Hz to 30K.
So I plugged them into the output of my sound card (Audiotrak Prodigy HD2) and put the mic close to the center of the speaker. The response is anything but flat: severe loss at both low and high ends. Any ideas?

Hi Alex, what are you calling high and low ends? Headphones flat down to 5Hz sounds extremely optomistic to me ;) could be marketers taking liberties!

on the high freq side, you have effectively done a nearfield measurement which tends to be inaccurate at high frequencies... WM61A's typically have a response peak at the high end, usually approaching 10Khz.

Tony.
 
I still use a signal generator and small power amp, a resistor and a meter or scope. I don't like the added mass method and usually knock together a test box or an adapter for some existing box. The numbers I get are generally very close to manufacturers numbers (at least the good ones). The signal generator can be any cheap thing, from antique to modern and the scope can be equally as decrepit. Most meters do OK at low frequencies and it doesn't have to be true RMS. The stuff just has to work and be reasonably accurate. I have a sound card and all the rest but unless I were doing a large number of woofers it's as easy to use the dinosaur methods just like Weems and Speaker Builder described them.
 
Just another Moderator
Joined 2003
Paid Member
ok that definitely sounds a bit sus then.... I'd try pointing it at a known good quality tweeter at perhaps a 1/2 meter distance and see what you get (for the high freqencies) and a nearfield measurement of a decent woofer should show you down to well below 500Hz.... , but it sounds like you have a problem with the mic or perhaps the preamp. did you do a linkwitz mod on it? if so you may have damaged it.

Tony.
 
I've borrowed the WT3 and it was quite easy for me to install the software and be up and getting all the T/S params calculated for me automatically in less than an hour.

I also have an old Mitey mike with MM1 preamp/power module which I'm going to have to try out again now that I realized a cable I was using in the past wasn't working.
 
Tried a no name mic from a phone headset and a calibration mic that came with a high end Denon - very similar results. Apparently even great headphones are not good for this. Same steep decline after a few KHz and no low frequency response at all unless the microphone almost touches the inside of an earpiece.
 
headset mics are not meant to capture all noise in a room, instead they would be optimal if centred around vocal frequencies and blocks souds that are not close or very loud.

I'm waiting for a secondhand function generator I got on fleecebay, then I also want to get started with measurements, allready got the scope :)
 
What coupler have you used? Headphone response may be coupler dependent around below 50Hz and WM-61A microphone dependent above 10kHz. Microphone is likely to have a resonance in high frequency band, before response rolled-off. And what is IN-SERIES resistance seen by headphone? Headphone specs you have might be derived from setup with a coupler. There are a few coupler types, depending on headphone / eat bud type, examples are here Ear simulators Have you tried T/S measurements for your headphones? I have Dayton Audio WT3P WT3 Bundled with Precision Scale it is extremely nice to work with, easy Impedance Test, but often it does not allow T/S test after Impedance Test completed, I need to figure out if I am using it properly. Sometimes USB cable straightened or bent gives slightly different results... need to keep it straight all the time, Peter
 
Tried a no name mic from a phone headset and a calibration mic that came with a high end Denon - very similar results. Apparently even great headphones are not good for this. Same steep decline after a few KHz and no low frequency response at all unless the microphone almost touches the inside of an earpiece.

Don't waste your time trying to calibrate your mic using ANY headphone. There are too many variables here AND the phones generally never have absolutely flat response. Good LF also depends on very close coupling with the head and you probably cannot achieve that properly with a mic set up. If you do the enclosed volume could create all sorts of other response errors that the 'ear' can take without a hassle but not a measurement mic !

Best option would be to measure a known tweeter and a known woofer and keep that as reference. Compare everything you make with that and you know what's hapening. What you will not know is what it's like in absolute terms. In any case the WM-61A can safely be used probably from about 200Hz to 5Khz without bothering about calibration. It's pretty flat there.
Do a search and you will find plots of several WM-61A capsules compared. Most vary beyond 5Khz .
 
I still use a signal generator and small power amp, a resistor and a meter or scope. I don't like the added mass method and usually knock together a test box or an adapter for some existing box. The numbers I get are generally very close to manufacturers numbers (at least the good ones). The signal generator can be any cheap thing, from antique to modern and the scope can be equally as decrepit. Most meters do OK at low frequencies and it doesn't have to be true RMS. The stuff just has to work and be reasonably accurate. I have a sound card and all the rest but unless I were doing a large number of woofers it's as easy to use the dinosaur methods just like Weems and Speaker Builder described them.
Could you perhaps point me to an on-line resource for how to perform this test? I have the gear, but no idea how to hook it up and interpret the results (for speaker measurement).
Thanks!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.