QuantAsylum QA400 and QA401

Jump to Latest
#1,741
Alan0354 said:
Thanks guys, I have to read it. But I managed to get it going by just monkeying around. I feed the output into input and read the spectrum. I notice, I put in a 1KHz and read back. I see a small peak at 2KHz about 100dBV down from the main peak. I try using both gen1 and gen2 and see the same thing.

I tried raising and lowering the output level of the generator, the 2nd harmonic follows at -100dB down.

Is it something to be concern? I thought the sine wave should be pure from the generator, or else how am I going to judge? 100dB is not that good, I got better in LTSpice with the IPS and VAS.

I tried higher frequencies, like 4KHz and 3KHz, it got better.
Click to expand...

LTSpice is a simulator. Hardware never works as well as a simulator. -100 dB harmonics is in fact very good. To get better instrumentation you will be spending a whole lot more money. An analog distortion analyzer with at best a 6 dB improvement will set you back at leat $400 probably more like $1K. The next step is AP, R&S, Shibasoku, Panasonic etc. You could easily spend $20K and maybe get to -110dB consistently, -120 dB when you are very careful.

There are other ways that are more affordable but if this is your first foray learn what you can get from the QA400. it may be all you need.
 
#1,742
Twin T notch filter. Variable. 250$ scratch built.
Oscillator $450.00 not including R&D and construction.
Shibasoku 725C. $900.00 + shipping, GST, brokerage fees and FX $1495.00 CAD.
Upgrades to 725C. Lost track. At least 100 hours+.

Okay I can get -120.9 THD+N.
 
#1,743
I am not talking about the analyzer, if I don't have any input, the noise floor is -130dBV, it's only when I connect the output to input and look at the 1KHz, when I adjust the 1KHz to 0dBV, then I can see 2KHz at -108dBV and 3KHz and -102dBV.

I think it's the generator that has the distortion, nothing to do with the analyzer.

In all fairness, I use my function generator ( cheap) and try, it was so bad, the 2nd and 3rd harmonics are only like 40dB down!!! It was ugly.

Thanks
 
Last edited:
#1,744
Alan, David and Demian are good. They were extremely helpful getting me
up and running with the gear that I assembled. Also Dick Moore and Richard Marsh
too. I'm green I'm sure I pissed em off a few times also, and I'm sure ther read
my stuff and said WTF.... They all stuck in there with me while I pulled my head
out and learned.

I also learned for frequency accuracy with the QA400 you need to have at least 32000 resolution.

We agreed for measurements using "Flat Head" so that all plots can be compaired
to other plots for reference.

These should help you get up the curve very fast.
 
#1,748
Alan0354 said:
I am not talking about the analyzer, if I don't have any input, the noise floor is -130dBV, it's only when I connect the output to input and look at the 1KHz, when I adjust the 1KHz to 0dBV, then I can see 2KHz at -108dBV and 3KHz and -102dBV.

I think it's the generator that has the distortion, nothing to do with the analyzer.
Click to expand...

The sweet spot for the QA400's input is about -10 dBFS.

The reason for the distortion is likely not the generator. The reason is the inherent increase in distortion in the ADC as the input approaches +3 dBV (i.e. 0 dBFS "full scale" or 1.41 Vrms).

In other words, to see the QA400's baseline distortion in loopback, set the generator(s) to -10 dBFS.

Attached is a screenshot of mine in loopback with both channels overlaid each other, tested one after the other from Gen 1. Notice the discrepancy between the left and right channel THD. This apparently is due to layout parasitics, according to QA themselves.
 

Attachments

  • QA400 Loopback 50 Avg 48K 65536 pts.PNG
    QA400 Loopback 50 Avg 48K 65536 pts.PNG
    415.8 KB · Views: 334
#1,749
Hi All

Sorry I did not reply sooner. I have been reading the manual and learning the software. I am sure I will have more questions, but I don't want to sound stupid asking questions that are clearly in the manual. I still have half to go and I have been typing the instructions in short forms for my own understanding.

Ha ha, I never use audio control software except Studio One of Personus. This is very different ( not that I am good at the Studio One by any stretch). The volume pot really tricked me.

I'll be back

Thanks
 
#1,751
I finished going through the manual, surprise surprise, I don't have much operating question at this moment. I am sure I'll have more when I really start using it later.

One question I have is the "Windowing". The manual said to use Flat Top. My experience with spectrum analyzer is always running on "rect" type display. What are the 4 options for?

Also, I notice when I tap the top of the unit, the noise floor jumped about 20dBV and settle back. This is when I use BNC coax to connect Rout to Rin and Lout to Lin. It happened to both channel. I remove the coax and tap, the noise floor still jump about 7dBV. Either case, it recover back.
 
#1,754
I just downloaded and installed the SW as linked to in post #1734, looks nice enough. I'm not ready to pull the trigger on this item, as I have no need, but perhaps I could find an excuse...

Alan0354 said:
I finished going through the manual, surprise surprise, I don't have much operating question at this moment. I am sure I'll have more when I really start using it later.

One question I have is the "Windowing". The manual said to use Flat Top. My experience with spectrum analyzer is always running on "rect" type display. What are the 4 options for?
Click to expand...
Was this an RF spectrum analyzer? Those things effectively use superhet radio received design, a llocal oscillator is swept with the horizontal, and the vertical is signal level out of the IF strip. This is a very different desgn than the FFT type analysis usually used in audio. It's this FFT that uses the various window types.

Googling fft windowing tutorial gives some good stuff:

Understanding FFTs and Windowing - National Instruments

The Bartlett and lots of other windows appear i the second link:

http://www.physik.uni-wuerzburg.de/~praktiku/Anleitung/Fremde/ANO14.pdf

Bartless is basically a "triangle," and so doesn't seem that useful, offhand.
Hamm and Hann (sometimes called Hamming and Hanning) can be confusing because of the similar names - they, like most others, are named after the guys who invented them:

https://en.wikipedia.org/wiki/Richard_Hamming
https://en.wikipedia.org/wiki/Hann_function

To make things even more confusing (and I presume Hamming was one of the first to come up with this windowing idea), these window functions in general are sometimes generically called "hamming windows."

In general, flattop will have the greatest amplitude accuracy, and not-as-good frequency accuracy, but usually in this kind of testing we already know the exact frequencies involved, and the amplitude is what we want to know.

If you're still reading and have technical interest in this, there's this free DSP book online:

The Scientist and Engineer's Guide to Digital Signal Processing

Also, I notice when I tap the top of the unit, the noise floor jumped about 20dBV and settle back. This is when I use BNC coax to connect Rout to Rin and Lout to Lin. It happened to both channel. I remove the coax and tap, the noise floor still jump about 7dBV. Either case, it recover back.
Click to expand...
Is there a microphonic component in the box???
 
You must log in or register to reply here.
Top Bottom