Re: Win11 ??

Here’s the glitch, seen in the 48 kHz record:



It lasts 6 ms. The separation of the peak signal on either side of the glitch is 16 ms, and the nominal peak-period should be close to 10 ms.

I have to wonder if the 76 minute period corresponds to the update period from the GPSD-OCXO? And maybe the Siglent generator does not handle 10 MHz reference clock slips very gracefully? You can sort of see that the generator is trying to restart on the recovery side of the glitch.

I’m on the Agilent generator now, for the past 30 minutes. Still waiting to see if it also glitches.

- DM



> On Sep 3, 2025, at 11:35, David McClain <[email protected]> wrote:
> 
> Very interesting…
> 
> The spectral display in Adobe Audition, once zoomed into the region of the glitch, makes the glitch really stand out. It is some kind of impulse in the RF channel, and does not affect the IRIG-B channel at all. Hence my keeping lock on IRIG while these glitches happen.
> 
> The problem really isn’t a fault of the data sampling system after all.
> 
> So what kind of RFI happens on a 76 minute schedule, very short impulse, lasting all of maybe 0.1 sec?  
> 
> Furthermore, it looks like my injected RF reference carrier has gone through a cycle slip, producing a huge discontinuity in the framing sideband signals. I use a Chinese Siglent Function Generator SDG 2042, to plant a DSB pair of reference sidebands at +/- 1400.5 Hz from the central WWV carrier. WWV at 15 MHz, and receiver nominal tuning at 14,985 kHz. Sidebands at audio frequencies of 2900.5 and 99.5.
> 
> I see no corresponding cycle slips in the IRIG-B channel, which tells me that my 48 kHz / 8 kHz sampling arrangement is okay. Something odd about that Siglent Generator - every 76 minutes, it slips backward by almost a whole second, based on the 0.5 Hz beatnote pattern showing up. The beatnote is perhaps generated by my 99.5 Hz sideband and the 100 Hz subcarrier modulation on WWV (??)
> 
> But I definitely see a cycle slip of about 1 sec from the reference carrier generator. The spectral display shows a very bright emphasis on both injected DSB sidebands.
> 
> And the same glitch does show up in my initial (ASIO) 48 kHz recording via Reaper.
> 
> 48 kHz record
> 
> <PastedGraphic-1.png>
> 
> 8 kHz record
> 
> <PastedGraphic-3.png>
> 
> The 8 kHz record has some meta-data modulated at the highest frequencies, between 3 kHz and 4 kHz. That is the black empty space in the spectrum shown for the 48 kHz record.
> 
> But it looks like my Siglent generator is causing the glitches. I was using it because it was much newer, and presumably cleaner output (re spurs) than my ancient Agilent 33522A Function Generator. 
> 
> I’ll go back to the Agilent Generator and see if the record clears up, re 76 minute spikes…
> 
> Very interesting…
> 
> - DM
> 
>> On Sep 3, 2025, at 10:53, David McClain (as dbm at refined-audiometrics dot com) <[email protected]> wrote:
>> 
>> Yes, that 1 kHz signal is what I have injected into the right channel from my AM Modulated IRIG-B signal coming directly out of my GPSD-OCXO. The signal in the right channel is completely unaffected by external RF events.
>> 
>> I’m looking closely at the recorded 8 kHz downsampled data vs Reaper recording at 48 kHz, both using the M$ MMS system. 
>> 
>> Normally, Reaper wants to use ASIO. But that compares apples and oranges against my data acquisition system.
>> 
>> Just have to wait to capture a glitch…
>> 
>> - DM
>> 
>>> On Sep 3, 2025, at 10:36, Yuri Davidovsky <[email protected]> wrote:
>>> 
>>> 
>>>> On 3 Sep 2025, at 18:47, David McClain <[email protected]> wrote:
>>>> 
>>>> I also thought about the possibility of an external RF impulse.
>>> 
>>> Another troubleshooting approach that I would try to really narrow it down to the device once and for all would be:
>>> 
>>> 1. Get an audio source of some simple signal, like a 1kHz sine wave. Just an old laptop with a basic software synth would do (VCV Rack comes into the mind as a free option).
>>> 2. Feed the 1kHz sine into the device, monitoring for the event. If it takes place, we can probably rule RF interference out. If the recorded signal remains pristine, then the problem is likely higher up the processing chain and there are no issues with the machine used.
>>> 3. If the event occurs on the USB device with the test signal, next step is to reproduce it on another device. We do not need to get very inventive and can simply use the inbuilt mic input on the machine and do another recording session to confirm if the event occurs with the test signal there too.
>>> 
>>> If the event happens on the USB device but not on the inbuilt device, we can be sure that there is something off about the USB interface. An interim solution in this case would be to try use the inbuilt audio for signal acquisition, you might need a preamp for it for this particular use case, but there should be a high chance that the problem is identified/solved at that stage.
>>> 
>>> If the issue is reproduced on both audio devices, then yes, there may be something off about the OS audio stack, or there may be something up with the machine.
>> 
>> 
>> _______________________________________________
>> Lisp Hug - the mailing list for LispWorks users
>> [email protected]
>> http://www.lispworks.com/support/lisp-hug.html
>