Re: Win11 ??

If we use 76 minutes as the period, then a sub-counter nudged by the difference between the incoming reference frequency and the nearest frequency that the receiver can produce, shows that the granularity of the receiver tuning is about 219 microHz. (= 1/(76 * 60)).

So the receiver DDS must have a fractional remainder counter running all the time. And the glitch that we see is due to the DDS finally inching up one tick of its granular synthesis capability, as the remainder counter finally overflows.

I guess the 6 ms garbage inside the glitch is related to the PLL time constant in the DDS synthesizer, as it responds to a 1 unit step change in the driving voltage to a VCO.

So, I guess this kind of glitchy behavior is to be expected of a DDS frequency synthesizer. If there were no fractional sub-counter, then it would probably not tune very well at any frequency.

But if this were the case, then why do I have a -19.495 mHz offset when I ask the receiver to tune to 15 MHz? Why can’t it tune to within the 0.219 mHz granularity found above? Perhaps because the lock-up time would be prohibitively long, e.g., 76 minutes?

- DM

> On Sep 4, 2025, at 03:28, David McClain (as dbm at refined-audiometrics dot com) <[email protected]> wrote:
> 
> Okay…
> 
> Here is a link to an 80 MB file on my shared public cloud drive. It is a 48/24 stereo file. The glitch is in the left channel at 3:51.831.
> 
> https://www.icloud.com/iclouddrive/010q2kpSROJdDZjeCquXDTTvQ#01-250903_1003
> 
> I have now decided that the glitches arise from the Icom IC-7700 Receiver. All the audio being recorded in the left channel comes from a SSB Demod of an RF signal centered at 15 MHz. That’s the only way to get an RF signal down to audio frequencies.
> 
> The glitch follows the radio audio signal, whether fed to Analog 6 or Analog 7 in the MOTU ADC. There is never any glitch in the IRIG-B AM signal in the other Analog channel. And the glitch is always the same 6 ms and cycle-slip, no matter which RF Reference Generator (Agilent or Siglent) is in use.
> 
> So my conclusion is that something slips inside the radio, causing its own SSB demod carrier to slip,  which produces a slip in the resulting audio phase.
> 
> And this follows your earlier notion of a 32-bit counter. I’m pretty sure the DDS inside the radio is using a 32-bit counter.
> 
> Exactly why the RF Demod reference slips is a mystery to me, and may never be answered.
> 
> - DM
> 
> 
>> On Sep 4, 2025, at 00:23, Yuri Davidovsky (as work at disclosure dot ie) <[email protected]> wrote:
>> 
>> 
>> 
>>> On 4 Sep 2025, at 02:45, David McClain (as dbm at refined-audiometrics dot com) <[email protected]> wrote:
>>> 
>>> I looked carefully at the 48 kHz recordings. The glitch shows really prominently in the Left channel, and there is no corresponding glitch in the Right channel IRIG-B audio.
>> 
>> Could you post a wav file snippet with the glitch in it including a bit of the clear signal? I will try to filter the burst out just to see what it looks like on its own by subtraction of the clean signal from the contaminated one.
>> 
>> _______________________________________________
>> Lisp Hug - the mailing list for LispWorks users
>> [email protected]
>> http://www.lispworks.com/support/lisp-hug.html
>