USRP2 scaling of short and float samples

GNU Radio
1

Hi GRers,

What is the scaling of USRP2 TX/RX samples for float and short type? I
noticed that the absolute values of received float complex samples of
USRP2
are below 1. I remeber in USRP the values are +/-several thousand at
maximum
which match the resolution of ADC/DAC. But how are the samples of USRP2
are
scaled?
I also noticed that in gnuradio/usrp2/host/apps/tx_samples.cc there is a
FPGA scaling using set_tx_scale_iq(). How does it work?

Thanks!

Bests,
Hanwen

2

hanwen wrote:

Hi GRers,

What is the scaling of USRP2 TX/RX samples for float and short type? I
noticed that the absolute values of received float complex samples of
USRP2 are below 1. I remeber in USRP the values are +/-several thousand
at maximum which match the resolution of ADC/DAC. But how are the
samples of USRP2 are scaled?
I also noticed that in gnuradio/usrp2/host/apps/tx_samples.cc there is a
FPGA scaling using set_tx_scale_iq(). How does it work?

In the USRP1 we use +/- 32767 as the floating point range. In the USRP2
we use +/- 1.0

Matt

3

Thanks Matt.
And how does the FPGA scaling with set_tx_scale_iq() work?

2009/9/22 Matt E. [email protected]

4

On Wed, Sep 23, 2009 at 02:04:35PM +0200, hanwen wrote:

Thanks Matt.
And how does the FPGA scaling with set_tx_scale_iq() work?

Don’t use it. It’s a bug that we expose it.
(It’s a final scaling that the library may someday use.)

Eric

scaled?
I also noticed that in gnuradio/usrp2/host/apps/tx_samples.cc there is a
FPGA scaling using set_tx_scale_iq(). How does it work?

In the USRP1 we use +/- 32767 as the floating point range. In the USRP2 we
use +/- 1.0

Matt