Wolf, Regarding your new GPS calibration feature, I bought the new GARMIN GPS 18x LVC a couple of weeks ago and finally found time to hook it up last Saturday.Message 1 of 15 , Aug 29, 2011View Source
Regarding your new GPS calibration feature, I bought the new GARMIN GPS 18x LVC a couple of weeks ago and finally found time to hook it up last Saturday. Its 1 PPS signal has got to be the least expensive, and traceable, calibration source for soundcard samplerate calibrations. How will we be able to demonstrate traceability of the FFT results back to the GPS 1 PPS signal?
And speaking of soundcards, last night I ordered the newer E-MU 0204 USB 2.0. I found several on-line reviews and articles stating that is uses asynchronous USB block transfers and (like you said) the DAC sample rate is not based on the PC’s clock / USB 1 ms framerate like most USB soundcards. I didn’t see anything about the ADC samplerate, but I would assume it is also controlled only by the crystal in the 0204.
My M-Audio PCI soundcard DAC has ~30 ms Std. Dev. of jitter for a 1004 Hz tone produced by SpecLab. I assume this cycle-to-cycle jitter is due to the PC being busy doing a lot of other things. If the ADC jitter is the same, what effect would that have on millihertz FFT analysis? The jitter histogram distribution appears to be Gaussian / random around a mean samplerate error.
Another question; I have been using a low samplerate of 8000 for Frequency Measurement Tests to measure a 600 Hz beat audio tone. The E-MU 0204 won’t sample that slow, so I will use 48000. I understand that this will increase my ~600 Hz absolute measurement error for any given ADC clocking rate ppm error, but I don’t know how the 6x faster samplerate might effect the (Complex, with internal offset) FFT results.
John Stuart, KM6QX
From: SpectrumLabUsers@yahoogroups.com [mailto: SpectrumLabUsers@yahoogroups.com ] On Behalf Of wolf_dl4yhf
Sent: Monday, August 29, 2011 9:19 AM
Subject: Re: [SpectrumLabUsers] Samplerate Calibrator
I will examine this issue as soon as time permits. There will be an awful lot of changes in the next version, also (and especially) in the sampling rate calibration (in combination with a GPS).
All the best,
Am 28.08.2011 23:11, schrieb John Stuart :
I’m doing some audio jitter tests today; using the SL Test Tone Generator feeding the DAC with the ADC turned off. To output exactly 1,004 Hz, as measured by my GPSDO locked HP counter, I need to adjust the “Input calib” to calibrate my soundcard’s output sample rate. Changing the “Output calib” has no effect on the audio frequency that comes out of the soundcard.
The screen-shots below show the 2nd pass of automatic sample rate calibration, where I had selected “Calibrate Input S.R.” and then “No” to the “use same for Output” question. After this 2nd pass, my counter is reading 1,004.000 032 Hz
John Stuart, KM6QX
From: SpectrumLabUsers@yahoogroups.com [mailto:SpectrumLabUsers@yahoogroups.com] On Behalf Of wolf_dl4yhf
Sent: Sunday, August 14, 2011 12:57 AM
Subject: Re: [SpectrumLabUsers] SignaLink USB
About the seemingly swapped input- and output clock calibration : Try running the test signal generator in half-duplex mode, i.e. from SpecLab's point of view only the DAC, but not the ADC. I bet the 'output sample rate' will have an effect. In full duplex mode (*both* ADC + DAC enabled), different sampling rates for in + output were always a pain because the internal processing chain only runs at one single rate. Only the output may be resamples, but (in full duplex mode) the test signal generator still runs at the sampling rate of the *ADC* (because the ADC's sampling rate can be corrected with a reference signal added to the input).
All the best,