Data Type Conversion does not convert to boolean when attached to transfer function block

17 visualizaciones (últimos 30 días)
Hello,
I attached a "Data Type Conversion" block to a PT1 low pass filter (transfer function) trying to convert the ouptut signal from double to boolean. Unfortunately this did not work as expected - the output was constantly '1' or '0'.
The same issue appears when a logic block is used after the transfer function. (I assume the same type of conversion is made internally.)
I have to convert the double value to an integer first, and than to boolean. This does the trick but in principle it should not matter in which block a double value is created. Converting e.g. a pulse source double output directly to boolean works certainly.
This strange behavior has been encountered in 2018a and 2019a and with all possible conversion settings.
Any suggestions?
Thank you,
Timo
Here comes some additional information:
Block diagram
Simulation result at Conv1_out: directly converted to boolean
The direct conversion delivers a constant signal, only.
Simulation result at Conv2_out with first conversion block set to 'int8':
Now, the output looks as expected.

Respuesta aceptada

Fangjun Jiang
Fangjun Jiang el 9 de Jun. de 2020
Your filter coefficient 50e-6 makes it respond faster but the output will "creep" for a long time.
The boolean conversion or the boolean(a) or logical(a) function basically is to check if value a is equal to zero. For double precision data, "true zero" is hard to get. see "help realmin".
You "Filter_out" value never reaches "zero" for double precision data. Log the data to workspce and check a value that seems to be 0 but the real value is around 2e-22 (depending on your solver and time step).
The rest makes sence if you add a "convert to integer" in the middle. To do it directly, use a Comparator block and see "help eps".
  1 comentario
Timo Dietz
Timo Dietz el 10 de Jun. de 2020
This is embarrassing. I was totally misled by the cursor readings of '0.0'.
You are certainly right - many thanks!

Iniciar sesión para comentar.

Más respuestas (0)

Categorías

Más información sobre Array and Matrix Mathematics en Help Center y File Exchange.

Productos

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by