Jupiter-T GPS module binary protocol

[davorin]

I've setup a prototype GPSDO based in the Jupiter-T module....

But as I would like to also display accurate date/time information with a microcontroller, the only seems to rely on the UTC time mark message 1108...

From the Jupiter-T datasheet it states that it should be possible to switch over to binary protocol with @@Wb command...but after that I still can send @@ commands and it responds...

What is the proper way to be able to receive the Zodica binary messages?

[davorin]

Found it...the command should be @@Wb\x01 and not just @@Wb...

Solution found here with help from James R. Miller:

http://www.jrmiller.demon.co.uk/projects/ministd/faq.html

"How do I RESET a Jupiter GPS?"...contains a ZIP file with various messages to switch between protocols you can just pipe to serial port...

[Melt-O-Tronic]

Thanks for that!  I thought I had been all over J R Miller's site, but I missed that file somehow.  Working out the protocol is the next step in my GPSDO project.

[davorin]

Well...GPSDO works....btu if I can have accurate timing for free...why not extend it...

Did some serial capture after switching to binary mode...it sends back @@Wb and immediately after that starts sending several messages after the 1PPS pulse...

Code: [Select]

<- @@Wb\x01\x34\x0D\x0A Switch to binary mode
-> @@Wb\x01\x34\x0D\x0A Acknowledged

->
FF81
5404 Message 1108 UTC time mark
...

FF81
E803 Message 1000 geodetic position status output
...

FF81
EA03 Message 1002 channel summary
...

FF81
6F04 Message 1135 EEPROM update
...

FF81
1F04 Message 1055 timing receiver configuration output
...

FF81
2004 Message 1056 timing receiver status output
...

FF81
5404 Message 1108 UTC TIme Mark
...

FF81
E803 Message 1000 geodetic position status output
...

FF81
EA03 Message 1002 channel summary
...

FF81
6F04 Message 1135 EEPROM update
...

FF81
2004 Message 1056 timing receiver status output
...

FF81
5404 Message 1108 UTC TIme Mark
...


I just don't know yet to which 1PPS pulse the UTC time in message 1000 belongs to...

[davorin]

This is odd...

After every time mark pulse the modules starts sending messages with ID 1000 first, which contains the GPS and UTC time...
THe last message is always ID 1108, the time mark message...

But when comparing the UTC seconds (seconds from ID 1000, and UTC seconds in week), the UTC seconds in message 1108 is always 3 seconds ahead....

Need to build a quick DCF77 receiver to be able to compare it against something (o;

[ruffy91]

GPS time doesnt count switch seconds. They are separately counted. I think this is your problem.

[davorin]

EhmmmI didn't mention GPS time (o;

Message ID 1000 outputs UTC time applied with the current 16 seconds correction...

Right after mark pulse the UTC time is for example 17:23:46, whereas the UTC time mark week seconds calculate to 17:23:48 instead...

Hmm...I missed UTC nanoseconds from Epoch...whatever that is (o;

Code: [Select]

FF81 Start Flag
E803 Message ID = 1000
3100 Data word count = 49
0000
E879 Header checksum
B70E Set Time 10msec ticks, 32 bit
0000
2400 Sequence Number
2400 Satellite Measurement Sequence Number
0000 Navigation Solution Validity
1000 Navigation Solution Type
0600 Number of Measurements Used in Solution
0000 Polar Navigation
1607 GPS Week Number = 1814
A93E GPS SecondsFrom Epoch, next run = AA3E
0200 GPS SecondsFrom Epoch
0000 GPS Nanoseconds From Epoch
0000 GPS Nanoseconds From Epoch
0D00 UTC Day = 13
0A00 UTC Month = 10
DE07 UTC Year = 2014
1000 UTC Hours = 16
3300 UTC Minutes = 51
2400 UTC Seconds = 36
FFC9 UTC Nanoseconds From Epoch
9A3B UTC Nanoseconds From Epoch

Calculates to 999999999nsec...really close to 1sec (o;