ARTA impedance measurements - help needed!!

[zobsky]

I built a simplified ARTA jig (http://soundhobby.com/measureimpedance.htm) . Whenever I try and calibrate ARTA,, I get a message (difference between channels is more than 2dB ...., check your sound card, cables etc ) or something like that. If you've ever had this message, you'll know what it is.

My configuration was
PC line out -> jig
jig->PC line in
jig->tweeter (I can hear hiss on the tweeter)

I used an 8 ohm resistor because thats what I had around. I know they suggested greated than 10 ohm, but I can't believe this is the issue.

I've checked all my cables, tried patching the line out directly to line in via a miniplug-miniplug connector etc but this issue persists.

Trying the impedance measurements with ARTA uncalibrated , I ended up with a sinusoidal mess.

My line in / line out are not not mute and the levels are equalized. I've tried this on 2 different PCs with the same result.

Help !!
Any ideas.

[Svante]

I cannot find a circuit diagram on the page you link to, could you post it?

[zobsky]

sorry, It's on the preceding page

http://soundhobby.com/measurementsystem.htm.

I get the same error with / without the jig.

[Henckel]

Are you sure that you are taking the signal to the two input channels from exactly the same source/point ?

Another option is that one of your input channels are defect or has a build in gain mismatch of more than 2 dB.

You could use RMAA for verification of the soundcard and channel ballance with a simple loopback cable

[zobsky]

I tried on two seperate PCs with different sound cards. I'll try RMAA and report back.

I'm convinced the fault is mine, .. just need to find out what the fault is.

[ente]

...... please check again all your wiring. Even if it's simple, you know sh.. happens.

Whats about the basic calibration of your soundcard?

Sometimes you can buy cables where left and right channel are mixed up. Please check it or try to change the reference channel.

Regards
Heinrich

[Svante]

Originally posted by zobsky
sorry, It's on the preceding page

http://soundhobby.com/measurementsystem.htm.

I get the same error with / without the jig.

Ok, I don't know the software and what calibration procedure they recommend, but I could think of a few things thjat could go wrong:

1. Do you calibrate WITHOUT the speaker connected? You should (probably) have nothing connected when you do calibration.

2. Do you really use only one of the outputs? Using one output for the feedback and the other for driving the loudspeaker could cause strange effects.

3. Is there a balance control on the input of the soundcard. Look in the Windows recording mixer.

4. Cheap soundcards can sometimes be surprisingly poor. Otoh you say that you have tried more than one soundcard so this might not be so likely.

[zobsky]

Okay,
RMAA wasn't recognizing the line-in, so I figured something was wrong in software.

I re-installed the drivers on one of the computers (creative audigy 2 ZS platinum) and managed to get RMAA working. I ran the tests via a lineout-linein loopback and everything seemed fine.



I went back to the calibration of the jig and it worked fine this time.

However, when running an impedance plot, I still end up with a mess (see fig). I thought maybe the (8 ohm) resistor was too low so I soldered in a 507 ohm resistor I had (and adjusted the value in the config and re-calibrated accordingly).
Same issue with an graph varying all over the place (I'm testing a cheap 8 ohm tweeter) . The driver is not connected via any amplifier.

[Henckel]

Originally posted by zobsky


However, when running an impedance plot, I still end up with a mess (see fig). I thought maybe the (8 ohm) resistor was too low so I soldered in a 507 ohm resistor I had (and adjusted the value in the config and re-calibrated accordingly).
Same issue with an graph varying all over the place (I'm testing a cheap 8 ohm tweeter) . The driver is not connected via any amplifier.

[

Try using an amplifier to drive the unit under test - I belive the 8 ohm resistor is much better than the 507 resistor you have tried

What happens if you select "stepped sine " as the generator ?

Sampling rates are correct ie soundcard and limp is set to the same values ?

[zobsky]

I built the "official" ARTA jig. I'm still having issues (zig zag impedance curves with both pink noise or stepped sine wave), even after re-calibrating the jig (I believe the reference resistor is 27 ohm). Are there any resistance measurements I can make on the jig to verify that I constructed it seperately?

For that matter, I get the crazy curves even when measuring impedance without a driver connected.

Here's a photo of what I constructed (used a cheap $1 plastic box with some ugly holes pre-drilled into it).

With this jig, I'm using my sonic impact amplifier to power the driver under test (a tweeter)