Crazy results using Cable or Mic attached to sound card

[John Mulcahy]

Afraid not. Perhaps best to make sure a basic loopback measurement works normally before trying the impedance rig, if you haven't already checked that.

 

[Elmojo]

I thought I had done that multiple times, but perhaps not correctly.
Is there a guide or video for that process somewhere?

 

[John Mulcahy]

There isn't a lot to explain, connect output to input and make a measurement. Attach the mdat file for review.

 

[Elmojo]

Hence my frustration.
I've found really very little actual explanation of how to do some of these steps.
For example, when connecting output to input, as you say, what cable should be used?
Is this with the regular measurement cable or a special 'calibration' cable that's wired differently?
If it's the regular cable, should the leads be shorted?
Is the sound card calibration necessary? If so, is it done with the same cable as the circuit cal done on the measurement screen?
There are a ton of little details like this that I've never seen demonstrated in a way that I can easily follow.
As much as I hate to admit it, I guess I've become part of the YouTube generation that needs to see things done to fully grasp them.
Also, when I connect to my external USB sound card, the noise floor (sorry, probably wrong term) as shown in the input graph on the measurement screen is at about -32dB. When I connect the same cable to my laptop, it's down around -61dB. Is this significant? I assume so, but can't figure out why or how to correct it.
The levels and results I seem to see on the preferences page don't appear to agree with what gets reported on the measurement page in some cases. For example, I can run through the 'check levels' process on the prefs page, and sometimes get things to look pretty good. However, when I go over to take a measurement and click the [check levels] button there, it tells me the ref in level is very low. Sure enough, if I ignore it and try to take a measurement, I get a similar message, and the graph is pretty much flatlined.
This is all addiing up to me not being able to make this work, and it's baffling me as to why. I really don't feel like it should be this hard, at least the early 'getting started' part.

 

[John Mulcahy]

There is nothing special about the cable used for a loopback, it is the same as would be used to connect anything to the interface.
The leads should not be shorted.
Soundcard calibration is not used for impedance measurements, they have their own calibration steps. For SPL measurements it can be used if the input and output are both on the same interface. It uses a loopback connection using a normal cable.

What external USB soundcard are you using?
What kind of connectors does it have?
What kind of connectors does your laptop's built-in card have? Is the input a mic/line or mic only? If mic only it isn't suitable.

Problems tend to be due to fairly mundane things, like using a mono connector in a stereo socket or stereo in mono. I suggest you begin by using only the signal generator (1 kHz tone) and the Levels meter with a loopback cable and make sure:

• that the levels shown on the inputs are sensible when the generator is not running (should be very low, typically below -80 dBFS)
• that when the generator is turned on using the left channel as output only the left input channel level increases, ditto when using the right channel
• that the input level follows the generator level. There will often be an offset, depending on the settings of the input and output level controls, but adding 10 dB to the generator output should add 10 dB to the input level

 

[Elmojo]

Thanks John, I appreciate all the info. That would have been super useful a week ago when I started this twisted journey.
My wife finally saw my frustration last night and pulled the plug, by which I mean she bought me the Dayton DATS box.
I think there's something wrong in the sound card(s) I'm using, or the way I'm trying to configure them, or both.
In any case, I just don't think this is going to work for me. Hopefully the 'plug and play' nature of the DATS box will be easier to manage.
Thanks again for your help.

To answer some of your questions, in case any of this helps someone else in the future...
I'm using this external card from Amazon: https://smile.amazon.com/gp/product/B076PC4VV4
It supposedly has a stereo Mic input, and line-in as well, which is why I bought that one.
It uses USB to interface with the computer, and all 1/8" (3.5mm) for the rest of the connections.
My laptop has separate headphone out and mic input jacks, both 3.5mm. I know you say that a mic jack isn't suitable, but I've seen it working on the video in the OP, so I know it's possible.
I'm using all stereo jacks, but the output (I think, could be the input) jack is wired only to one channel, making it effectively mono, as shown in the diagram you posted earlier.
I obviously didn't know anything about the signal gen, so I haven't tested that, but I can say that on the external card, when there are no signal inputs, the level is at about -32dB, which I find very suspicious.

 

[John Mulcahy]

My wife finally saw my frustration last night and pulled the plug, by which I mean she bought me the Dayton DATS box.

Very good of her!

I can say that on the external card, when there are no signal inputs, the level is at about -32dB, which I find very suspicious.

Sounds broken.

 

[Elmojo]

She's a good one for sure!
I agree, I think the external sound card has issues.
That doesn't explain, though, why I get similar whack-o results from the internal card on the laptop, and the Audigy 5/RX card in my desktop machine.

 

[Stollery]

So I'm having a similar problem, and am similarly frustrated. I agree it looks like I'm just recording noise however when I run the impedance sweep I do get a signal that looks like the expected impedance curve, take a look at the input graph bottom left of the impedance measurement dialogue. This measurement completed OK but resulted in the impedance and phase chart also attached.

I've tried switching input channel, changing resistors, rebuilding circuit, different speakers but still no joy. Don't understand why the level shown in 1.jpg doesn't result in a reasonable impedance curve in 2.jpg.

Anyone any ideas?

 

[John Mulcahy]

Perhaps something was not correctly set up when calibrating the impedance measurements. Try clearing the calibration using the Clear Cal button and seeing if the measurement looks more plausible then. If it does try each calibration step individually, following it by a measurement, to see which step is going wrong and narrow down where to look.

 

[Stollery]

Looks like one input channel on my sound card is dead - new card coming tomorrow - will report back. Thanks so much for your help

 

[Stollery]

No success unfortunately. New USB sound interface exhibits the same behavior. (Reason I though there was a problem is that the scope only had input on L channel, but I know realise that's by design). I've attached some REW measurements to see if they enable anyone to give any insights as to what I'm doing wrong. There are 2 loopback calibrations for the 2 soudcards followed by the impedance rig calibration process. I've also attached a screen grab of the settings.

I have the impedance rig driven by the headphone output of the UCA222 (or 202) and the output of the rig connected to the L/R line inputs.

Thanks in advance for any ideas.

Phil.

 

[John Mulcahy]

Reason I though there was a problem is that the scope only had input on L channel, but I know realise that's by design

Can you explain what you mean by that? Why should there only be input on the L channel?

I've attached some REW measurements to see if they enable anyone to give any insights as to what I'm doing wrong. There are 2 loopback calibrations for the 2 soundcards followed by the impedance rig calibration process.

The file you attached contains one impedance measurement and nothing else.

Did you try my suggestions above?

 

[jtalden]

I have the impedance rig driven by the headphone output of the UCA222 (or 202) and the output of the rig connected to the L/R line inputs.

I see your screen grab of the measurement lists the "Microphone" as the input. Did you try to select "line in" input?

 

[Bernard]

I'm using this external card from Amazon: https://smile.amazon.com/gp/product/B076PC4VV4

Late arrival...
Which OS are you using ?
Read on Amazon : OS support: win98 / XP / 2000 / Vista / Win7. NOT support win 10!

 

[Stollery]

I see your screen grab of the measurement lists the "Microphone" as the input. Did you try to select "line in" input?

Thanks, I changed it to the only other option on input "Default Input", same behavior both on L or R input channel

 

[jtalden]

Considering Bernard's post above you may want to look for a new or used audio interface that is commonly used with REW. eBay has lots of options at very low cost.

 

[RobertoG]

Hello everyone. I encountered the same problem when measuring impedance. At the beginning, I used a built-in sound card. Unfortunately, no positive effects. Then I used the UCA202 card, the results were the same - bad. The problem is that after calibrating the "reference cal" instead of getting a straight 3ohm resistance line (I used such resistor), I get a very skewed graph that does not correspond to reality. I've been sitting on it for the third night and can't get it over with. Sample charts during calibration are presented below. I combined with levels, changed channels, mounted the measuring jig twice - with no results. I am also wondering why after short circuit calibration I get the result which is a straight line of 100ohm. I am using version 5.20.
I will be grateful for your help.

 

[RobertoG]

I would like to add that if I choose SPL measurement I see curve which coresponds speaker impedance (blue one). These two charts below are normal speakers impedance measurement after calibration. There is something not well done with calibration but I cannot find it.

 

[John Mulcahy]

Is your sense resistor 100 ohm? If so you just need to swap the input channels, an indication of channels being swapped is getting a measurement result that sits on top of the sense resistor value, as was the case for your short circuit measurement.

 

[RobertoG]

I used 48ohm resistor, but I put its value to the menu where Rsense resitor is define. Shoud it be fine, am I right? Of course I tried to swap input channels, but without success. I always see noisy impedance result with huge peaks up tu hunderd ohms. Such result are displayed during Referece calibration and during normal measurments done after whole calibration process. When I select SPL measuremnet for the same hardware connection menu I see nice curved characterisctic, which looks like impedance but on Y axis I have dB.

 

[John Mulcahy]

You could try clearing the calibration and making measurements without any cal, to see if you can get valid results. That may help track down any issues with how the jig is wired or connected.

 

[Stollery]

Hello everyone. I encountered the same problem when measuring impedance. At the beginning, I used a built-in sound card. Unfortunately, no positive effects. Then I used the UCA202 card, the results were the same - bad. The problem is that after calibrating the "reference cal" instead of getting a straight 3ohm resistance line (I used such resistor), I get a very skewed graph that does not correspond to reality. I've been sitting on it for the third night and can't get it over with. Sample charts during calibration are presented below. I combined with levels, changed channels, mounted the measuring jig twice - with no results. I am also wondering why after short circuit calibration I get the result which is a straight line of 100ohm. I am using version 5.20.
I will be grateful for your help.

This is the same behavior I've been getting posted above. I'm afraid to say I couldn't resolve it and followed the thread initiator and bought a DATS box. It works fine with REW.

 

[RobertoG]

I am afraid the problem cannot be easily solved. It is definitely not the jig or the sound card. Just measuring the card and its transfer characteristics gives good results. The strange thing is that when measuring impedance, the UV meter shows sensible changes and the levels are large. Unfortunately, then a mess is displayed on the characteristics. As if there was a problem with measurement data and their reference to calibration. Are you sure that v5.20 works correctly in terms of impedance measurement with the jig suggested in the manual? Thank you for your help.

 

[John Mulcahy]

Are you sure that v5.20 works correctly in terms of impedance measurement with the jig suggested in the manual?

Yes. Have you tried clearing the calibration and measuring without it?