SG-2340/MBT-4220 GPS-Lure Stratum-1 NTP clock under pfSense?
-
Hmm, I used the generic type with a USB connected GPS previously.
Do you see the com port? You can probably just try connecting to it using cu and some baud rates etc. You should see it spitting NMEA.
Steve
-
https://forums.freebsd.org/threads/gpio-pps-driver-for-minnowboard.70768/
https://forums.freebsd.org/threads/gpio-on-turbot.58693/
-
Depending on what your time is worth (no pun intended), this $300 standalone unit with external antenna has worked well for me for years.
GPS NTP Network Time Server (TM1000A)
https://timemachinescorp.com/product/gps-time-server-tm1000a/ -
Yup you could go that route for sure... Or for like less than a $100 total and a few minutes time you could setup a pi with a hat and run that as stratum 1.. I have one on my home network, that serves to the pool and my local network for like couple of years now.
-
Thanks for the various alternative proposals.
Having a stratum-1 NTP clock isn't that important for my purposes, that I what the hassle of once more device that can break, requires an outlet, etc.
It would just be "cute" to have one, given that the device I use as firewall/router/VPN box would be able to provide one, given proper drivers.
So it's more about using a device to its fullest, than having a hardcore need for a stratum-1 clock.
-
Hear yah - they really should put a disclaimer that currently does not work on pfsense/freebsd..Until such time that there is driver that allows it to, etc.
-
I agree, I'll try to get something added.
Steve
-
I would assume most people buying off the hardware off netgate store prob wanting to use it as their pfsense box ;) So yeah before some actually bitches should prob put disclaimers on any add on sort of hardware that isn't currently easy to get working, without like writing your own driver ;) hehehe
-
Well, on the freeBSD forum (see above links), someone said there's already a working GPIO driver for the board, so all that's missing is the PPS aspect.
I'm trying to see if I can get more information on that. Maybe sending that person who wrote the GPIO driver for the board a GPS module might be an incentive for him to write the PPS driver.
That would be a win-win for everyone...
-
Any chance someone knows how the config should look like to simply get the GPS to provide time, i.e. without the PPS?
I always get that, when I look in the NTP status:
May 18 23:18:46 ntpd 34463 kernel reports TIME_ERROR: 0x41: Clock Unsynchronized May 18 23:18:46 ntpd 34463 kernel reports TIME_ERROR: 0x41: Clock Unsynchronized May 18 23:18:46 ntpd 34463 GPS_NMEA(0) serial /dev/gps0 open at 9600 bpsAnd that with an external GPS antenna, in a location where my phone has no trouble picking up a GPS signal.
So I assume it's not actually the clock/GPS not being synced up, but the settings being wrong.Under NTP "Serial GPS", I have the following:
GPS Type: MediaTek
Serial Port: cuau0, 9600 baud
NMEA Sentences: ZDA or ZDG
Flags: Prefer this clock, Display extended GPS status
GPS Initialization:$PMTK225,02B
$PMTK314,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,1,028
$PMTK301,22E
$PMTK320,02F
$PMTK330,02E
$PMTK386,023
$PMTK397,023
$PMTK251,960017So far no dice....
-
Are you seeing useful any useful output from the device at 9600? The one I had ran at a lower board rate, 2400 I think.
ntpd won't sync to what it sees as a bad clock source but you should still see it showing a time value in the ntp status.
Steve
-
Mmm, OK the docs indicate 9600 but also to use serial port 4... do you see more than one serial port?
-
I tried to use cu to check any device output, but I did't get any, but seeing that least I dealt with serial ports was back when kermit was still under active development, that doesn't mean much, I may just have invoked cu the wrong way.
I picked 9600, because that's what the Linux setup showed. The standard MediaTek setup is 4800 (if selected on the pop-up), tried that, too, but no change. Also, I updated the init strings to 9600 baud, because they showed 4800, even after increasing the rate to 9600.
I know about the port number, but cuau0 is the only device showing. Since pfSense supports the serial console, and since that supposedly uses the same pins as the ones to which the GPS lure is attached to (and not regular USB-to-serial), I figure that is the proper device.
If there's a way to use different serial interfaces, I haven't found it.
-
Does anyone have any input one what it takes to access the GPS module, simply as a clock, without PPS?
Can't seem to get that working, either. -
Ok, this looks promising.
The minnowboard has a standard com port that is brought out to the 6 pin debug header. That's where you get the serial console if you enable it, /dev/cuau0. It also has two UARTs on the low speed header that the GPS-Lure uses but it appears that in the default config FreeBSD/pfSense does not recognise and attach to them.
But! If you change the
LPSS & SCC Devices ModetoPCI modein the BIOS an additional two com ports are then detected. That's in Device Manager > System Setup > South Cluster Config > LPSS & SCC Config.
Also check that both UARTs are enabled there. Mine were here but I've played around with those settings I can't be sure they were by default.Anyway once that is done and you see cuau2 and cuau3 in /dev try those ports for NMEA signals from the GPS.
Steve
-
@stephenw10 Cool, thanks!
That brings us quite a step forward. (BTW: is there documentation anywhere on all the settings in there? Many of them are obvious, others not, managed to enable some things that resulted in the NICs no longer being recognized by the system, so I reset to factory settings, and then just re-enabled the UARTs).Anyway, that done, I now get this, which means the GPS module is seen by the system:
[2.5.0-DEVELOPMENT][admin@local]/root: cu -s 9600 -l cuau2 Connected $GPGGA,010835.800,,,,,0,0,,,M,,M,,*4F $GPGSA,A,1,,,,,,,,,,,,,,,*1E $GPRMC,010835.800,V,,,,,0.00,0.00,060180,,,N*45 $GPVTG,0.00,T,,M,0.00,N,0.00,K,N*32 $GPGGA,010836.800,,,,,0,0,,,M,,M,,*4C $GPGSA,A,1,,,,,,,,,,,,,,,*1E $GPRMC,010836.800,V,,,,,0.00,0.00,060180,,,N*46 $GPVTG,0.00,T,,M,0.00,N,0.00,K,N*32 $GPGGA,010837.800,,,,,0,0,,,M,,M,,*4D $GPGSA,A,1,,,,,,,,,,,,,,,*1E $GPRMC,010837.800,V,,,,,0.00,0.00,060180,,,N*47 $GPVTG,0.00,T,,M,0.00,N,0.00,K,N*32 $GPGGA,010838.800,,,,,0,0,,,M,,M,,*42 $GPGSA,A,1,,,,,,,,,,,,,,,*1E $GPGSV,1,1,00*79 $GPRMC,010838.800,V,,,,,0.00,0.00,060180,,,N*48 $GPVTG,0.00,T,,M,0.00,N,0.00,K,N*32 $GPGGA,010839.800,,,,,0,0,,,M,,M,,*43 $GPGSA,A,1,,,,,,,,,,,,,,,*1ESo, since is obvious from this, that the GPS module can't get a fix, that's not yet being quite there...
Figuring that the patch antenna on the module inside a metal case might not be ideal, I invested into an external GPS antenna. My iPhone gets a GPS lock where the unit is, so I'd figure with an external antenna, that should be good enough of a location. Maybe not. Maybe the system just takes much longer to get a fix, as it's not assisted by cell tower and WiFi information to acquire one.
Hopefully in a few hours or so, it will acquire a satellite fix.No matter what, at least we now know we can access the device, and that's the biggest obstacle.
-
There is this though it's a bit outdated: https://elinux.org/Minnowboard:MaxBios
Yeah that looks like no satellites in view. And, yeah, it can take a while without agps. But that looks pretty promising otherwise.
Steve
-
@stephenw10 Just had an idea: is the external antenna detected automatically, or do the init commands enable it? If it's not automatic, and the integrated patch antenna is being used, inside of what's a metal case with a few vents, in a basement, I'd 100% be not surprised it can't get a satellite lock.
On the other hand, if the switch-over is automatic, then I am surprised, because I'd expect an external antenna to have a better reception than what my iPhone gets, and that has no issue locking on to the GPS signal, basement and all.
Wonder where I can find out, if the external antenna is recognized automatically...
-
Hmm, good question. Can you move it somewhere with better reception for the patch antenna as a test?
Steve
-
@stephenw10 Not really, because of where internet comes into the house. I could just try to run the box stand-alone for a while to test, both with and without the cover of the case, and with and without the external antenna (if for some reason it should be busted), but I'd have to do that when I don't need access to the internet for a period of time...
