QuantAsylum QA400 Audio analyser

[dfuruta]

How about something like this?

A few things I don't know about:
the output pad messes up the output impedance–I'm not sure this is a problem, but if it is, I'll change it to an H pad.
the smallest resistor in the output pad is tiny, 1.2R.  I'm planning to use slide switches to keep the cost reasonable.  Even with the contact resistance of the switch, the pad should still be accurate within a couple of dB–is this likely good enough, or would it be a better idea to bump up the total resistance of the pad?

This schematic is becoming a bit of a maze.  Once everything is more or less as it should be, I'll draw it again for easier reading.


The big news:  I picked up my QA400 from the post office today and tried to get it up and running.  The computer I'm using it with runs Windows XP SP3.  The software and drivers installed without incident.  However, the calibration procedure keeps failing.  When I try loop-back measurements with the uncalibrated device, the plots look feasible and it seems to be working as it should, but for some reason it won't calibrate;  at first, the program crashed, but now the calibration window just spits out a cryptic error message.  I'm exchanging emails with the Quantasylum support, and they've been helpful so far.  But, we haven't found the solution yet.

It's possible something is just going wrong because I'm using an old computer, but I have no clue.  Hopefully I'll know more about this soon.

 

[ruffrecords]

I am not an advocate of NF or EIN as a specification since both are dependent on the gain. They are handy to know and give an indication of the noise contributed by the input section of the pramp. However, most mic pres are not used at maximum gain for any significant portion of their lives and spend most of their time at much lower gains. In this region EIN and NF are of less importance than the output noise of the preamp.

Cheers

Ian

 

[abbey road d enfer]

However noise performance is expressed, EIN or NF (or S/N ratio  ), one number is not enough. I have no problem with EIN expressed in dBu or dBv, but it's not sufficient without a value for noise current. BTW, I know that a 200r resistor produces -129.7 dBu noise voltage, just like I know pi = 3.1416 or OdBu = 0.775V
Regarding NF, I have a slight problem, since it is ambiguous. It is possible to define NF = 3dB for let's say 100r and 600r. Now I suspect a manufacturer will quote the lowest NF for one optimum value of source impedance, which again is not enough to properly describe the behaviour.
When I wrote: "Sometimes noise is measured with typical 150-200 r load and a short-circuit; I don't believe it's for particular enlightenment purpose," I meant that most people would be lost if they had more than one number for evaluationg the noise performance of a piece of gear they covet. What is the proportion of SE's who can work out  the implications of a combination of EIN and Input Noise Current, or NF at 150R + NF at 50 and 800R?
I know many SE's would just say: "Huh".
I think however that this is for me the most useful specs, because I can relatively easily deduce the value of noise current and extrapolate for any value of source impedance. I would not feel so comfortable with two values of NF.

 

[dfuruta]

It's alive!  QA support was prompt and very helpful.  The crashing problem I encountered is fixed in the next software version, apparently.  They gave me a simple workaround.  So far, the software seems easy to use.

It's a cute little guy, isn't it.

 

[ruffrecords]

It's alive!  QA support was prompt and very helpful.  The crashing problem I encountered is fixed in the next software version, apparently.  They gave me a simple workaround.  So far, the software seems easy to use.

It's a cute little guy, isn't it.

I am itching to buy myself one of these. Have you done any tests of real gear yet?

Cheers

Ian

 

[dfuruta]

Nope, not yet.  Didn't have time today, and to be honest I've been putting most of my effort towards getting this frontend together.  I can say, though, that the loopback tests seem to be as advertised, the software is user-friendly, and the support from the company has been great (although, it seems like they might not have it totally ironed out, yet).

 

[ruffrecords]

The suspense is killing me so I just went and ordered one. Even with the $49 shipping to the UK it still cost less than 170GBP.

Cheers

Ian

 

[Rochey]

For me, the quality of the software is just as important as the hardware.

I only have experience with AP, but if they can make a GUI that's just as easy, they are laughing all the way to the bank.

Any talk of spdif io?

 

[dfuruta]

The suspense is killing me so I just went and ordered one. Even with the $49 shipping to the UK it still cost less than 170GBP.

Cheers

Ian

Great!  You'll be able to give a much more qualified review than I.

 

[abbey road d enfer]

For me, the quality of the software is just as important as the hardware.

Agreed.

I only have experience with AP, but if they can make a GUI that's just as easy, they are laughing all the way to the bank.

From the example on their website, the programming language is much more complex than AP's (private void button7_Click Object sender Events arg e ?????!!!!!). What's more, AP's learn mode is even simpler.
Those guys at AP know that GPIB is an esoteric language, so they made their own.
At QuantAsylum, they use the resources of Microsoft's Visual Studio which, for me, is as esoteric as GPIB or NI's Labview.
Anyway, I've ordered a QA400, I'll have it next month (had it delivered at friend's so I saved customs). I guess I'll have the interface finished by then.

 

[Rochey]

Abbey, thanks for being our beta tester.

I've often consider simply using a behringer USB box. It would do most of the analog measurements required around here... I think one could achieve a lot by simply characterizing it.

But at $200... I'd be tempted to go directly to this.

 

[abbey road d enfer]

Abbey, thanks for being our beta tester.

Actually, I think dfutura is our beta-tester. As I said, I will pick up my box in april. My comments are based on the website (and living with AP since their beginnings - I was their french distributor for about 5 years - had to give up AP because of silly politics).

I've often consider simply using a behringer USB box. It would do most of the analog measurements required around here... I think one could achieve a lot by simply characterizing it.

My experience with "musician's" USB converters is that most of them have actual performance far worse than published figures. Gross distortion, 60dB S/N, spurious,...that makes them unsuitable for measurement purpose. Ther are probably some that have decent performance, but I couldn't really which. I'd rather have nice unbalanced connections (as in the QA400) than those noisy so-called balanced outputs I have experimented.

But at $200... I'd be tempted to go directly to this.

Talk about impulse-buying...

 

[audiomixer]

If it weren't for the silliness of getting stuff across customs here in Switzerland I would get one to.

just started to test my PCM4222EVM together with ARTA software, but windows audio is a dungeon full of pitfalls so deep you might never see the light again. I have a I2S interface that has WDM driver only. I set up the line input to 100% level in the mixer proprieties dialog, as you would expect and 96kHz / 24bit - and the signal clipped at +2dBu already and the noise floor is all over the place instead of +14dBu. windows kindly does digital gain and possibly sample rate conversion in the background as well without notifying? I hate windows so much.....
next up will be a ASIO driver - we will see how this behaves.

the best interface I used till now is the duran-audio measurement interface, bypasses the windows audio subsystem as well.  for reliable measurements this is the best way to go I am afraid.

anybody up for a QA400 group buy here in Europe ;-) ?


- cheers, michael

 

[ruffrecords]

Mine has finally arrived in the country. I just paid the duty/taxes on it and with luck it will be delivered tomorrow!!!


Cheers

Ian

 

[dfuruta]

Looking forward to reading your report!

I've almost got my front-end for my QA400 put together.  Haven't had enough time away from work in the last week to finish it up, but I should soon!

After playing with the QA400 a little more, it would be nice if it could automatically measure IMD.  I wonder if this will be in a future software release.

 

[gemini86]

So is there a calibration feature at all? I would hope that the frontend addons don't have much distortion or noise, but if they add just a little, it would be nice to zero it out.

 

[dfuruta]

I'm pretty sure the calibration procedure only affects levels (and only within a rather narrow range), and not distortion or noise or anything else.

 

[ruffrecords]

Disappointing. Can't get past the calibration procedure at the moment. I get the correct 1.29V outputs but when I do the loop back connect and continue I get a calibrate FAil#1 and windows says a serious error has occurred and shuts down the program. I am running Vista in a Samsung R700 laptop.

Cheers

Ian

 

[dfuruta]

yes, i had that problem too.  the solution is to let it run for a while, looped back into itself with the generators on, uncalibrated.  then try calibrating and it should work.  no clue why it does that;  apparently the next software release will fix this.

 

[ruffrecords]

yes, i had that problem too.  the solution is to let it run for a while, looped back into itself with the generators on, uncalibrated.  then try calibrating and it should work.  no clue why it does that;  apparently the next software release will fix this.

Thanks for the tip. I notice there is a sort of beta version on their web site - maybe that includes the fix. If this is such a common problem you would have thought they would include it in a FAQ.

Cheers

Ian