Yes. If you search there was discussion about not using exclusive since it never seemed to perform properly. I always stick with default and slightly lower than 0dbfs when testing. I do think the difference between REW and QA mirror is related to drivers.
I agree 100% with your assessment that it’s a driver problem. I did try the Topping windows driver, but it made no difference.
Thanks for all your help.
Hi @BrianG, one other thing to keep in mind is that 0 dBFS can be a relative measurement. For example, the a 0 dBFS signal should be Math.Sin(2PI1000). That would give you sine “tips” at +/-1.0. But another app might opt to make 0 dBFS be 0.95 * Math.Sin(2PI1000).
Also, there is the Windows volume setting to contend with. This can result in the output being much smaller than you expect for 0 dBFS.
Finally (and I don’t think this is your case) but the RMS can vary a lot based on the waveform type. We generally think of 0 dBFS is the most signal you can get. But in the case of motor drive, once you have pushed as much power as you can into the motor using a sine wave, the sine can be distorted into a trapezoid to increase the RMS further, and then into a square wave to increase the RMS further still. So, a 0 dBFS sine gives you an RMS of Sqrt(2), while a 0 dBFS give you an RMS of 1 (42% more than a sine). As I said, I don’t think this applies to your case. But just to be aware that gross distortion can increase a reading substantially.
Now, with the above said, something peculiar in the first two plots (A and B) is that when driven via Mirror, you get an RMS value of 4.179V, and when driven via REW you get 4.242V. So, when driving with the QA40x app you get slightly less amplitude AND more distortion. And then in C, when you back off 0.5 dB, the distortion improves but your RMS only dropped from 12.42 to 12.05 (instead of the expected 11.92). That suggests some gain compression in the QA40x output and suggests the tips of the sine are getting clipped. But where that is happening (QA40x software, windows, DAC) isn’t clear.
Thanks for the reply Matt. Even though I’ve been retired for 12 years, I’m still very aware of the math behind the creation of the sinewave. I was mostly working through the problem to understand the nature of Windows and if it was something there that I was missing. Moto was a big help working me through the problem.
In all my measurements, I had my DM3063 connected to the right channel and I show that in the captions. My E70 is consistently around 11 mV lower on the right channel compared to the left at those higher gain settings. Swapping channels around change the direction. It’s amazing how close and agreement between the QA403 and the DVM are.
I found the sweet spot for the dBFS setting using Mirror of -0.2dBFS. THD & THD+N are amazingly close.
The QA403 is an amazing product and I’m extremely happy with it. Good luck with your further investigation. Can’t wait to get my hands on the QA455 to augment my measurements.
One small correction – Rigol DM3068 not DM3063