UHD source output double precision

GNU Radio
1

Dear All,

I know that the gr_complex data type is equivalent to
std::vector<std::complex > but for my work with GRC, I need to
use
UHD sources with double floating point precision. Is this possible or it
is
hardware limitation?

Best regards,

Zohair

View this message in context:
http://old.nabble.com/UHD-source-output-double-precision-tp29213228p29213228.html
Sent from the GnuRadio mailing list archive at Nabble.com.

2

On 07/20/2010 06:00 AM, Zohair wrote:

Dear All,

I know that the gr_complex data type is equivalent to
std::vector<std::complex > but for my work with GRC, I need to use
UHD sources with double floating point precision. Is this possible or it is
hardware limitation?

It should be possible, but I think you will need to make some software
changes in:

http://ettus-apps.sourcerepo.com/redmine/ettus/projects/uhd/repository/revisions/master/entry/host/lib/transport/convert_types_impl.hpp

Philip

3

On 07/20/2010 03:00 AM, Zohair wrote:

Zohair
You could do that, but the underlying signals do not have that kind of
dynamic range, so it would be a waste.

If you really need doubles, the easiest thing would be to just add a
float to double block outside the UHD source.

Matt

4

So if I leave the source float but in other block that I am writing the
precision is important and the output is necessarily double rather than
float. Will GRC recognise this? I checked the list for types from the
help
menu and double data type is not there!

Any hint, please?

Cheers,

Zohair

View this message in context:
http://old.nabble.com/UHD-source-output-double-precision-tp29213228p29226000.html
Sent from the GnuRadio mailing list archive at Nabble.com.

5

There are no double or complex double blocks in gnuradio, so the support
is not added into grc for such types.

If you make custom blocks with double types, you can set the type to
byte in grc xml file, and set the vector length to be the number of
bytes in your type.

Only the the io signature must match, so the length in bytes must be
identical when connecting, gnuradio does not know anything about the
data type.

Out of curiosity, why the need for double precision?

-Josh