Re: Constellation Modulator delay calculation

[Ali G. Dezfuli]

thank you, Jared and Cinead!

Sure it helps Jared!

On Wed, Jan 11, 2023 at 7:39 AM U L <jareddpub@gmail.com> wrote:

Ali,

I think the example uses the generic_mod block inside. This block uses the polyphase arb resampler for antialiasing. The line that computes the rrc delay is here. Briefly, for your example with sps = 2, it's (11 x 2 x 2 - 2) / 2 = 21. For sps = 4, (11 x 4 x 4 - 4) / 2 = 86. I don't know if this helps.

Jared.

On Tue, Jan 10, 2023 at 6:08 PM Cinaed Simson <cinaed.simson@gmail.com> wrote:

Hi Ali - your flowchart has no device and no throttle. You need to add a
throttle.

And I don't the expertise or the time to look at the source code.

So I'm punting back to the list.

-- Cinaed


On 1/10/23 03:10, Ali G. Dezfuli wrote:
> Hi everybody,expertise
>
> I just want to know how the delay of the "Constellation Modulator"
> block in GRC is calculated.
> In fact, whether you set the block's last parameter "truncate filter
> transient" or not, a delay of 86 samples could be generated, no matter
> how many points are in the constellation.
>
> This magic number (i.e. 86) also appears in
> "linear_equalizer_compare.grc" example (in
> gr-digital/examples/equalizers in modulated_sync_word variable).
>
> I know this delay depends on the "samples per symbol" (sps) parameter
> and comes from the built-in pulse shaping with its default taps equal:
>        firdes.root_raised_cosine(32, 32, 1.0, exess_bw, 32*11*sps)
> and with the following sps to delay relation:
> (sps=2, delay=21)
> (sps=4, delay=86)
> (sps=8, delay=348)
> etc.
>
> my GR version is:   v3.11.0.0git-316-gc11667ef
> thank you all!