Friday, September 20, 2024

Re: Timing a flowgraph execution

Hello Again,

Apologies for the extra email, but I found my problem and thought it
best to share my findings. Firstly, I had an error in my flowgraph
that was causing Overflows and Underruns as well as Sequence errors
that I hadn't detected. Fixing this fixed the problem I was having
with timing the flowgraph.

I have deleted the null source to throttle to null sink in the
flowgraph and set the "Function Probe" block to use the ID of the
"Soapy RTLSDR Source" block. It then reports the value of
nitems_written as fprobe_val, which I use in the "QT GUI Level Gauge".
I set the value of the Level Gauge to be 'int(((fprobe_val /
samp_rate) / duration) * 100)'. When the flowgraph is run a progress
bar moves across the screen and I can tell how much of the satellite
pass has occurred and adjust the antenna position accordingly.

So I consider this fixed.

All the best,

Chris

On Wed, Sep 18, 2024 at 8:40 AM Chris Gorman <chrisjohgorman@gmail.com> wrote:
>
> Hello all,
>
> I would like to set a timer to create a percent complete graph with
> the "QT GUI Level Gauge" block. I've created a flowgraph that does
> this and attached a screenshot and the flowgraph. My goal is to have
> this kind of "percent complete timer" to tell me when to move my
> antenna during a satellite pass. This means adding a piece of
> hardware and its connections to the attached flowgraph and I am
> worried about having two timers in the same flowgraph.
>
> If I use this flowgraph and add hardware and use a head block with
> "samp_rate * duration" items it doesn't complete at the same time as
> the "percent complete timer". There are more items to complete for
> the hardware when the "percent complete timer" ends.
>
> I have tried to use the "Function Probe" block to probe the source of
> my hardware piece, but that doesn't work properly as the timing is off
> for both timers.
>
> Does anyone have a suggestion on how to calculate the percentage
> complete from the timer on the hardware piece instead of using a
> separate null sink as I have in the attached flowgraph?
>
> Thanks in advance,
>
> Chris

No comments:

Post a Comment