Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
H
HDL
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Iterations
Wiki
Requirements
Jira
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Locked files
Build
Pipelines
Jobs
Pipeline schedules
Test cases
Artifacts
Deploy
Releases
Container registry
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Code review analytics
Issue analytics
Insights
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
RTSD
HDL
Commits
b69a5929
Commit
b69a5929
authored
6 years ago
by
Eric Kooistra
Browse files
Options
Downloads
Patches
Plain Diff
Improved description.
parent
ff746bd2
No related branches found
No related tags found
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
applications/apertif/matlab/correlator_null_clip_or_wrap.m
+15
-5
15 additions, 5 deletions
applications/apertif/matlab/correlator_null_clip_or_wrap.m
with
15 additions
and
5 deletions
applications/apertif/matlab/correlator_null_clip_or_wrap.m
+
15
−
5
View file @
b69a5929
...
...
@@ -24,9 +24,14 @@
%
% Description :
% Clipping seems preferrable because it does still contribute somewhat
% constuctive to the correlation. Even a little bit (ampl = 1.1) of
% wrapping already gives seemingly random output, so one could even
% argue that wrapping does not cause decorrelation.
% constuctive to the correlation. Clipping preserves the phase
% information (zero crossings) of the signal and a clipped signal still
% resembles the original overflow signal.
% Even a little bit (ampl = 1.1) of wrapping already gives seemingly
% random output. For severe overflow (ampl > 2) the wrapping cause extra
% zero crossings. If the wrapped signal becomes sufficiently random
% then the contribution of wrapped parts of the signal to the
% correlation will be close to zero.
% Using noise input it shows that for clipped input the correlator
% output still follows the correlator output of the full scale input
% that is used as reference (ref).
...
...
@@ -34,10 +39,15 @@
% correlator output can be normalized by N or by the number of non-zero
% contributions N-cnt. The plots show that nulling and scaling yield
% somewhat destructive correlator output.
% Even without RFI the signal may occasionally (e.g. ampl = 1.01)
% overflow due to the variation of the noise. For this clipping seems
% the most graceful choice.
%
% Conclusion: best choose input clipping for a correlator.
%
% Remark:
% Remarks:
% . This model does not investigate the effect of the channel filterbank
% on the data.
% . Delay tracking and fringe stopping will take care that the input
% b will be in phase with input a, so should we in the plots only
% look at the correlator output for phase difference close to 0 or
...
...
@@ -59,7 +69,7 @@ close all;
%rng(0); % random seed for repeatable result
% Try different amplitudes
ampl
=
1.5
;
% for input rand() or sin() use > 1 for overflow
ampl
=
8
;
% for input rand() or sin() use > 1 for overflow
% for input randn() ampl = sigma, so use >~ 0.3 for overflow
N
=
1000
;
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment