Re: [Discuss-gnuradio] 1PPS Synchronization - File Meta Sink Data

Hi Michael,

If you are working with metadata and timestamps, I think the gr-analysis module may be of use to you. As for your particular question, my understanding is that the 1PPS provides the timestamp information present in the metadata. So I would suppose the (possibly interpolated) sample with integer seconds would be on the PPS. But I'm not 100% confident on that so perhaps someone else can chime in and help us out. I will also point out the file metadata sink is not particularly efficient at writing metadata (although quite robust) so you may need something like specrec (in gr-analysis) for higher sample rates.

https://github.com/garverp/gr-analysis

PWG

From: Michael Skaggs <mskaggs1@umbc.edu>

Subject: [Discuss-gnuradio] 1PPS Synchronization - File Meta Sink Data

Date: April 1, 2016 at 10:49:30 AM EDT

To: discuss-gnuradio@gnu.org

Hi all,

I'm doing a simple RF recording with a B200mini synchronized to an accurate GPS 1PPS. I'm attempting to determine which samples(most importantly the first one) fall on a 1PPS pulse. I'm using the File Meta Sink block in GRC to try and extract information from the recording. When I read the data from the header, here is what is shown:

fileName: 'X:\XXXXX.dat'

ifileName: 'X:\XXXXX.idx'

ioType: 'r'

totalPulseCount: 11

sampleCount: 1000000

startTime: [0 0.1120]

timestamp0: '1970-01-01 00:00:00.112'

rx_rate: 1000000

rx_time: [0 0.1120]

timestamp: '1970-01-01 00:00:00.112'

rx_freq: 500000000

sampleType: 'float'

typeFlag: 'double'

vBit: 1

impediance: 50

samples: []

So, this is a 1 MSps recording with a center frequency at 500MHz and about 11 seconds in length, so the extracted information is definitely correct. But a couple of questions from the metadata:

The "startTime" value is 0.1120. I realize that this is the time stamp of the first sample of the segment. However, does that mean that sample 112000 (1MSps*0.1120) is the first sample that falls on a PPS since the B200mini is synced to it?

Second, the "samples" field is empty. I was attempting to populate it with sample numbers that fall on a 1PPS or a header. If the header is synchronized to the 1PPS, this will be fairly easy. If not, is there a way to determine which samples fall on the PPS?

Thanks,

Michael