Having trouble with speaker workshop

[noodle_snacks]

I have not had any trouble getting frequency response measurements (behringer ub802 and ecm8000 mic) and other types of acoustic measurements with a microphone. Now i have come to impedance measurements and i am having difficulty getting any sort of result.

instead of building a Wallin Jig or similar i decided to just build an implementation of the circuit by Claudio for measuring impedance. As far as i can tell i have put together the circuit correctly, and when i play music with it plugged in i can hear sound comming out of the speaker.

For my reference resistor i used what i had availible for 8 ohms. I am really not sure what i am doing wrong. When i take an impedance measurement for my two peerless HDS woofers in parallel i get the graph attached. Any suggestions as to what I am doing wrong, or haven't done? Reversing channels does not help.

[jomor]

1rst of all you should re-run all the calibration procedures as described in claudio's website very very carefully. SW is so comlicated when calibrating that its too easy to skip something.
Make sure you dont skip the Interchannel Time Delay calibration procedure.
You should use 1% tolerance resistors for the impedance jig calibration (after calibrating the sound card input impedance and capacitance). It is critical to know the exact values of the calibrating resistors ( its helpful to measure them with a high accuracy multimeter- otherwise you cant get accurate impedance plots, thus accurate crossover simulating). One calibrating resistor should be close to double, and the other should be close to half value of the reference resistor, so the calibration will be more accurate. Then try to measure various passive components to examine the results. If measuring some inductors give you a resistor result, then your calibration or the cable is not correct and need to be checked again. If you measure passive components correctly, then measure a single woofer to get the impedance graph.
Make sure that the MLS Signal ~ Use pre-emphasis is unchecked.


Good luck.

[noodle_snacks]

ok i think i have the calibration correct, passive components seem to measure correctly etc. when i try a woofer (well actually two in parrallel). the graph is started to look a fair bit better but is a complete mess from about 50hz down.

Any suggeestions as to what i might need to change to fix this problem? Could it be a result of my cheap and nasty onboard soundcard?

[HJWeedon]

Hi

I measure speaker impedances on a regular basis. I have been using my own method that seems to work fine for me.

My method is to connect the output of my Audio oscillator (Heatkit IG-5218) to an Audio power amplifier (Heathkit AA-1505). I put a 1k resistor in series with the output of the power amplifier and the DUT back to the amplifier return. I connect an AC DVM across the DUT (Divice Under Test) connections. I then substitute a known resistor of around 8 ohms for my speaker under test. I have measured the test resistor say 8.33 ohms, and adjust the sinewave input to my power amplifier at a frequency of about 400Hz to read 83.3 milli Volts for a load of 8.33 Ohms. I take notice of the output of my Audio oscillator.

To make sure of the schematic: The + output of the Power Amplifier is connected to one end of the 1kOhm resistor the second lead of the resistor connects to one terminal of the SUT. The second terminal of the SUT is connected back to the signal return -terminal of the Power amplifier. The DVM connects across the SUT.

Make sure that your DVM can read the 83.3 mV accurately to at least 20kHz some really cheap ones do not do that very well. I do that test with the 8.33 ohm test resistor connected.

I now connect the SUT (Speaker Under Test) instead of the 8.33 Ohm resistor. I set my Audio oscillator to 10Hz and measure the voltage across the SUT. If that reads 40mV it is a 4 ohm speaker at 10Hz. I carefully adjust the AC voltage to my power amplifier as constant as I can by eye.

I now increase the frequency of my Audio oscillator in about 10 to 20 steps per frequency decade and write down the frequency and the measured resistance of the speaker as 51mV => 5.1 Ohms and so on.

I know that this method does not measure the phase of the impedance and is no more accurate than about 5%, but the method is relatively quick and gets me good and consistent results.

Hans J Weedon

Salem MA USA.

[HJWeedon]

Hi.

I think you are suffering from low frequency rolloff of your audio board. Both the A/D channel (input channel) and the Output channel have AC coupling. This AC coupling limits the low frequency response of your system. I think I can see the double resonance around 20Hz or so. That I believe is your 2 woofers mechanical resonance. The two speakers couple, and produce a double peaked resonance.

Below that frequency, I agree that all is nonsense. That is most probably due to insufficient accuracy of your system due to the low frequency rolloff of the measurement system I do not believe that a more expensive audio board will help much, they probably all roll off below 30Hz to avoid rumble and other nastiness.

The poor measurement quality below 20Hz is an unfortunate result of using equipment not specifically designed for the application. Hope you have better luck with an improved setup. I would fully trust your setup from 80Hz and up.

If you have any question I will be glad to try to answer. I have been doing this sort of stuff for about 55years (age 67 by now), inherited from my father, who started around 1925 doing Audio stuff.

Hans J Weedon

Salem MA. USA.

[noodle_snacks]

I have done some tweaking and the measurements i am getting now show the double impedance peak for my ported box slightly above the expected tuning frequency (the dip in the middle is about 50hz, -3db point is 55 or so from winISD simulations). So assuming i ignore low freqency data from say 30hz down, this impedance data should be ok for crossover simulation shouldn't it?