Test System

Project started in May 2008, and my current test system was completed in February 2009.

I use SoundEasy (SE) v14 to test drivers and speaker systems. I suppose I'll get the hang of before I die, but I'm not absolutely sure...

The program is quite useful, but required a substantial investment of my time before I was able to produce usable results. Get John Kreskovsky's Design Guide, print it out, and use it for your first project. Download all of the manuals, print, read, and use them (you'll need three thick binders). Also, join the SE users group on Yahoo. Post your unsolvable dilemmas there, John, Bohdan, and Soongsc monitor the group and can help people that have exhausted the manual and Guide resources for answers. Now I'm pretty comfortable with SE and I've developed a decent portable system to use with it.

By the way, I don't test drivers like this. I had hooked the driver up just to test the system.

When I'm running free air impedance tests, I clamp the driver in a horizontal position or mount it to the edges of a pair of 2" x 2" boards. The objectives are:

• 1. To firmly anchor the driver.

  2. Don't change the free air load on the cone (i.e., don't give the cone any appreciable surface to couple to).

  3. Try to keep it at least 2' from near-by objects to cut down on reflected sound disturbing the cone during the measurement.

When using MLS or stepped sine wave signals to test drivers, the signal levels are usually at or over 1 V_RMS. At levels this high, reflections from nearby surfaces can often be seen in the impedance curve. This is why most professional impedance testing is performed while clamping the motor structure of the driver to something solid.

One great exception is the WT3 tester form Parts Express. I don't use one, but its test signal level is low enough that you can usually set drivers on a table (as in the picture above) and get good results. By the way, be careful to not cover or obstruct any vents on the rear of the motor (pole vents etc.), it will effect your measurements.

I don't really need to do frequency response and HD tests on a large test baffle since I'm not trying to develop a library of reusable tests at this point (like Zaph has published on his site). I'm going straight to specific system designs. I mount the drivers in the speaker system's enclosure, then I measure each driver for SPL and phase data, mounted as they will be used, to develop the crossover design. The advantage to this methodology is that acoustic center off-sets and diffraction effects are in the measurements and (fairly) easy to account for in the crossover design. One disadvantage to this approach is that it is tough to see, quantify, and separate these effects individually, although they can be separated and measured with some extra effort if you need to.

The system consists of:

A used IBM R52 Laptop (nothing special, 1.6 GHz, running XP, SP3).

A used E-MU 02 cardbus soundcard (special, very high quality analog sections, with sampling rates up to 24 bits at 192 kHz).

A used E-MU m1616 dock (special, again very high quality, matches the 02 card specs).

A home built LM3886 amplifier (turned out kinda special, very clean).

A home built connector box (test jig), handy for multiple connections.

A ProTek digital multimeter (most of the quality of a Fluke at about 1/2 the cost), used to match test signal levels.

An ECM 8000 measurement mic by Behringer, calibrated by Kim Girardin to 45 kHz.

I also have an old Radio Shack SPL meter and a General Radio Type 1562-A calibration tool, so I can set things to repeatable SPLs. My absolute accuracy at 1000 Hz is probably good to ± 0.5 dB. The Radio Shack SPL meter is A weighted and stops at about 8 kHz, so it's not worth much for any frequency other than 1 kHz. I'm still looking for a way to calibrate my measurement mic for decent absolute accuracy in general SPL readings, either in SoundEasy, or I may get a copy of True RTA.

The attached PDF covers the complete system, getting SE to work with the E-MU system, performance (loop-back) tests on the E-MU and test amp, how the connector box is wired and what you can do with it...

Jay R. Taylor

02/21/2010

Home