It has taken me a while to decide, but I finally made up my mind on how I wanted to test my speakers. If my budget wasn't limited, the choice would be a lot easier (should I go for MLSSA, or maybe Linearx's LMS?). But it is limted, so I had to spend my money well. I did a lot of reading on webpages, tested some demo software and also asked around a lot.
Finally, the choice came down to this: should I go for a dedicated measurement option, or buy soundcard measurement software. The only dedicated measurement solution I could afford was DLSA pro, from RCM akustik. Although it is a decently performing system, the functionality is somewhat limited. For the same amount of money, there are also some pretty good software packages available, like sample champion from purebits, hobbybox, loudspeakerlab and soundeasy. They offer more options, but use a soundcard as the measurement device. After reading some more and gathering even more info through some private emails, I finally reached the conclusion that a soundcard measurement system would perform equally well as a hardware system.
I chose Sample Champion from Purebits. Mainly because it is a measurement only package, it has a good waterfall plugin and the Thiele/Small plugin also seemed pretty good. I already have LspCad std, so I wouldn't need the the speakerdesign features that other measurement packages offer. Because I'm planning to upgrade to LspCad pro in the future, I also had planned the purchase of a better soundcard than I currently own. To do some crossover simulating with LspCad pro, you need a multichannel soundcard. I've decided on the M-audio delta 410. This is a decent soundcard, with 8 analog outputs, with which you can emulate a stereo 4way crossover. More than enough, it would seem. This card also has 24 bit/96khz ad/da convertors so I can also do some pretty serious testing with it. One advantage of going the software route is that you can build your setup in small steps. Which is a lot more do-able than buying a complete measurement package in one go. My upgrade plans are: Make a test jig, buy Sample Champion, buy calibrated microphone, buy soundcard and finally: purchase LspCad pro. At the moment of writing, I own the jig and Sample champion (and an uncalibrated panasonic capsule); the soundcard and microphone are underway. LspCad pro will have to wait a few months.
For my testjig, I made a big list of demands and whishes, because I wanted it to be "all-inclusive". This is the list:
- Flexible usage (suitable for more than one software package)
- Built-in microphone pre-amp
- Built-in poweramp
- BNC connectors
- Safety features
So I crawled behind my pc, and produced this circuit:
The microphone preamp is the one from Eric Wallin.
The volumepot has 5 discrete steps. I did this, because I wanted volumesettings that are repeatable. The steps are: off, -30dB, -24dB, -18dB and 0dB.
The amplifier is based on the TDA2040and comes directly out of the datasheet. The only thing I did is to lower the gain, because lower gain means better frequency response. I did a quick test with a signal generator and my scope and saw that the peaks of the sinus-signal began distorting from 60khz and up. When distortion is visible on the scope, it means that the signal began distorting at a much lower value, but I hope that the amplifier doesn't distort below 48khz, which is the highest frequency that can be measured with my new 96khz soundcard. The minimum is 24dB, and this can be obtained by raising feedback resistor R2 from 680ohm to 1360ohm (I used a 1K2 resistor)
This is a small PCB that I designed for this amplifier. Print with 300dpi to get the correct size, and check the correct orientation. I accidentally made a mirrorimage of the pcb, which I only found out later. It did explain, however, why the amp didn't work the first time :) Check the datasheet for component values.
A quick explanation of all the switches:
Switch 1 Switches between left and right input. This way, you can connect both left and right output from the soundcard. When I use different programs, switching between left and right goes faster with a switch than changing cables.
Switch 2 Takes the amp completely out of the circuit.
Switch 3 Chooses between impedance and frequency measurements.
Switch 4 Is the +20dB switch for the microphone preamp.
Switch 5 Enables or disbles the probes (as suggested by the JustMls manual)
Here is an image of the jig from the back, with the cover taken off:
Visible are the powersupplies, the microphone preamp and the switches. The switches are from apem (comparable to miyama) and are the single/dual/quad pole toggle type switches. Not yet installed is the shield between the powersupply and the the rest. It is a simple piece of double sided pcb, I don't know if it really works, but since there is no humm on both amp and preamp it doesn't seem to hurt either.
And finally: the front.