[USRP-users] [Discuss-gnuradio] Late packet and phase issue
Marcus D. Leech
mleech at ripnet.com
Tue Oct 6 10:21:03 EDT 2015
On 10/06/2015 05:01 AM, irene menicagli wrote:
> Hi everyone,
> I have the same problem (late packet). I think that the problem is the
> time on packet is in the past when the command arrives to device. The
> cause can be latency between pc and usrp.
> My pc is quite powerfull so this behaviour seems strange.
> Maybe I have to change timestamp on packet but I don't know if it is
> possible because I work at python level and use UHD functions
> (usrp_sink and usrp_source).
> I also use only one USRP loopbacking TX e RX.
It's not about the performance of your PC, typically.
How are you setting the timestamps on the timed packets? If you're
using host time, that will drift relative to the USRP. Further, if you
using timing that's to tight, latency variability across the network
(or USB) stack can cause packets to arrive "late" from the perspective
of the USRP.
> 2015-10-05 21:13 GMT+02:00 Marcus D. Leech <mleech at ripnet.com
> <mailto:mleech at ripnet.com>>:
> On 10/05/2015 10:58 AM, Leonard Foxes wrote:
>> Hi list,
>> It's me again.
>> I use a USRP N210 with WBX daughterboards and loop back cable to
>> connect TX e RX. I have two problems:
>> 1. Late packet. I use set_time_now and set_start_time command
>> and after few seconds I see some Ls printed.
>> 2. Phase. Phase offset changes between executions. (I'm trying
>> to use set_time_command but nothing changes).
>> This is my python code about usrp:
>> self.samp_rate = samp_rate = 500e3
>> self.USRP_address = USRP_address = "addr=192.168.10.2"
>> self.tun_freq = tun_freq= 1.45e9
>> self.tun_gain = tun_gain= 19
>> self.rx_freq_off = rx_freq_off = 0
>> self.timeref1= timeref1=time.time()
>> self.uhd_usrp_TX = uhd.usrp_sink(
>> self.uhd_usrp_TX.set_time_now(uhd.time_spec_t(timeref1), 0)
>> self.uhd_usrp_TX.set_command_time(uhd.time_spec_t(timeref1+0.1), 0)
>> self.uhd_usrp_TX.set_center_freq(uhd.tune_request_t(tun_freq), 0)
>> self.uhd_usrp_TX.set_gain(tun_gain, 0)
>> self.uhd_usrp_RX = uhd.usrp_source(
>> self.uhd_usrp_RX.set_time_now(uhd.time_spec_t(timeref1), 0)
>> self.uhd_usrp_RX.set_command_time(uhd.time_spec_t(timeref1+0.1), 0)
>> self.uhd_usrp_RX.set_center_freq(uhd.tune_request_t(tun_freq +
>> rx_freq_off), 0)
>> self.uhd_usrp_RX.set_gain(tun_gain, 0)
>> Any suggestion will be welcome
>> Thanks in advance,
> Something to realize is that if you call set_time_now() on two
> different USRPs, they'll not agree in detail about the
> time-of-day. If you're
> using an external 10Mhz and 1PPS reference, you should probably
> use set_time_unknown_pps(), and make sure that both sides
> of this are setting the same time AT THE SAME TIME.
> Also, WBX has a 180deg phase-ambiguity. Two WBX boards with the
> same 10MHz reference, on USRPs with the same notion of time,
> will either be in-phase, or 180deg out-of-phase, due to hardware
> ambiguity in the 2XLO mixer used on WBX.
> Discuss-gnuradio mailing list
> Discuss-gnuradio at gnu.org <mailto:Discuss-gnuradio at gnu.org>
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the USRP-users