Shop OBEX P1 Docs P2 Docs Learn Events
GPS Module PMB-688 Time Sync Problem — Parallax Forums

GPS Module PMB-688 Time Sync Problem

DaveGDaveG Posts: 84
edited 2010-05-31 04:30 in General Discussion
I recently purchased two Polstar PMB-688 GPS Modules w/external antennas from Parallax. They are working beautifully and I am decoding the sentences from them using a pair of SX28's. The modules routinely pick up 6-9 satellites.
·
My eventual goal is to place these GPS modules at two different locations, 60 miles apart, and use them to generate accurate synchronized time-stamps for some data. Actually I would be happy if the time-stamps were synchronous within +-10 ms.
·
My initial thought was to use the 1 per second sentences from the GPS modules to reset accurate 100Hz counters. The GPS modules would give me the hrs/min/secs and the counters would provide the sub-second count.
·
Today I have two breadboards, each containing a GPS module and an SX28, setting side-by-side on the bench. Each SX28 watches the 4800 Hz data stream from its module and puts out a short sync pulse when the S/W sees the start bit of the "$" in the $GPGGA sentence. A 2 channel scope is used to compare the sync pulses from the two breadboards.
·
Problem: I see a jitter and/or offset of +-50ms between the two sync pulses. Since each byte is about 2ms, that would be equal to 50/2 = 25 bytes? What is going on? I see lots of commercial units that claim to have a +-100ns accuracy, but they are too expensive for me.
·
Is there a different place in the data stream that is in perfect synchronization?
What am I missing?
·
Thanks.
·
Dave G.
Apple Valley, CA

Comments

  • TobiasTobias Posts: 95
    edited 2010-05-23 02:01
    Hi Dave, Sorry I can't help But would you have any advice for me with the pmb-688, I have got the same unit and am having difficultys with the code, data, I can get my BS2e to read, debug the nmea strings, but can't take out the data that I need, would like to have time and speed,,.

    Thanks, Toby
  • FranklinFranklin Posts: 4,747
    edited 2010-05-23 02:13
    Toby, go back to your original post on this and ask these questions there, If you have an attachment of your program it would help debug your problem (in your other post, please)

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Stephen
  • Martin HodgeMartin Hodge Posts: 1,246
    edited 2010-05-23 03:42
    I've tried several GPS modules, and none of them have had very precise time output. The time stamp accuracy seems to be secondary to spatial accuracy. it's insane, because the spatial data is calculated from the time data. But the cheap units just don't seem to want to share that info with the outside world. I found an old Rockwell module that has 1pps and a hundredths field in the GMA sentence. With some simple math on the SX I've been able to get what looks to the naked eye as accurate time (compared to a WWVB atomic clock). The serial data of these modern units doesn't seem to be synced with anything but the previous burst. Look for a unit with a 1pps output and give that a try.

    Or maybe some else will chime in with some tricks?
  • DaveGDaveG Posts: 84
    edited 2010-05-31 03:35
    Martin,

    Thanks for your advice regarding 1pps output devices.

    After a little research I found just what I needed and ordered a pair of Garmin GPS 18x LVC receivers. I got them from Amazon, for $68 each, in two days.

    These units put out the usual sentences but they also have a separate wire that outputs a pulse that is precisely sync'd to the atomic clocks of the GPS system, within +-1us. Good enough for my application.

    I have them hooked up on the bench and the scope shows the pulses to be well within spec. Very sweet!

    Thanks again.

    Dave G.
  • Martin HodgeMartin Hodge Posts: 1,246
    edited 2010-05-31 04:30
    Nice (and fast) work! Glad to be able to help.
Sign In or Register to comment.