LFRX not picking up close to DC


#1

I am trying to use the LFRX daughterboard to record very low frequency
signals in the range of ~DC to about 10Hz. To test this I hooked up a
function generator to the USRP with an output impedance of 50 ohm. Since
I
wanted to get 4 channels simultaneously I loaded the “std_4rx_0tx.rbf”
image
and looked at the signals after recording to a file. The daughterboard
seems
to be having a DC blocking behaviour where the amplitude of received
signals
decreases with lower frequencies. My rx gain is set to 0 and Vin = 1Vpp
with
0 DC offset. At 10Hz I get an amplitude of 400 from the ADC. At 1 Hz it
drops to 40 and I get nothing at DC. I have tried fiddling around with
the
gain, but it doesn’t help much.

This seems to be weird behaviour since the LFRX is supposed to work upto
DC.
Am I missing something? I have read all the previous posts on this issue
and
they weren’t helpful for my problem. Any hints/directions would be
greatly
appreciated. I have spent quite a bit of time on this can’t figure out
the
problem.

Karthik


#2

On Wed, Oct 29, 2008 at 03:41:19AM -0700, Karthik Vijayraghavan wrote:

This seems to be weird behaviour since the LFRX is supposed to work upto DC.
Am I missing something? I have read all the previous posts on this issue and
they weren’t helpful for my problem. Any hints/directions would be greatly
appreciated. I have spent quite a bit of time on this can’t figure out the
problem.

Karthik

Karthik,

you’ll want to turn off the ADC dc offset correction loop
that’s running in the FPGA. (From usrp_basic.h)

/*!

  • \brief Enable/disable automatic DC offset removal control loop in
    FPGA
  • \param bits which control loops to enable
  • \param mask which \p bits to pay attention to
  • If the corresponding bit is set, enable the automatic DC
  • offset correction control loop.
  • 
    
  • The 4 low bits are significant:
  • ADC0 = (1 << 0)
  • ADC1 = (1 << 1)
  • ADC2 = (1 << 2)
  • ADC3 = (1 << 3)
  • By default the control loop is enabled on all ADC’s.
    */
    bool set_dc_offset_cl_enable(int bits, int mask);

In python, it’s a method of usrp_source_*.

Eric


#3

On Wed, Oct 29, 2008 at 3:55 AM, Eric B. removed_email_address@domain.invalid wrote:

signals
and
you’ll want to turn off the ADC dc offset correction loop
*
*/
bool set_dc_offset_cl_enable(int bits, int mask);

In python, it’s a method of usrp_source_*.

Eric

Eric,

Thanks, that worked! For archival purposes here is the piece of code
that
worked for me. It disables the DC offset correction from all 4 ADCs

self.rx_src = usrp.source_c()
self.rx_src.set_dc_offset_cl_enable(int(0),int(15))

I hooked up a 2Hz sine wave with no DC offset to the USRP and I see a
rather
large offset after the signal comes into the USRP. This offset value
changes
every time I run my program and remains constant during the course of
the
program. The offset is different for different ADCs / channels. I also
measured an open circuit voltage of 106mV between the inputs of all the
ADCs. Is this variable offset due to some registers not getting cleared
when
a program finishes running?

Thanks,
Karthik


#4

also measured an open circuit voltage of 106mV between the inputs of all the
ADCs. Is this variable offset due to some registers not getting cleared when
a program finishes running?

Thanks,
Karthik

After measure the signal offset I adjusted the adc offset from python
till
the offset went away. When I use something like
self.rx_src.set_adc_offset(3,int(273)) I was able to get the offset
completely removed. Also, for some reason after putting this line in,
the
variable offset that I was seeing went away. I checked with restarts,
USRP
power off etc and this seemed to work perfectly.

Karthik