If you've been wanting to move to a >1Hz GPS receiver but were holding out until a retail version of the awesome U-blox chipset was available, today's your lucky day. Sparkfun has just started selling the 4Hz GS406, which uses a helical antenna, which in our experience beats the usual ceramic patch. Note that it's 3.3 v, not 5v like the 1Hz EM406 that we're currently using for ArduPilot, so if we go with this we'll mount it on a voltage-converting daughterboard.
Specs follow:
* Sarantel omni-directional Geo-helix S-type active antenna
* 4Hz Update rate
* Fifty satellite tracking channels for fast acquisition
* Supports UBX data protocol at 9600bps
* Filtering effect of the antenna gives high immunity to RF interference
* Wide antenna Beam width
* Low power: 3.3V @ 75mA
* Time to fix:
o Hot start: <1 second
o Warm start: 29
o Cold start: 29
Dimensions:
* Length (with antenna) 54.5mm
* Width: 22.9mm
* Weight: 16g
When I ordered multi-hundred Qty of LS20033 and LS20031, I got a test result sheet showing that each unit was tested and the lock time reported. All units are tested before Locosys sends them out.
The LS20033 defaults to 9600 baud, 8N1. I assume your max232 interface is setup for this?
First power up it might take many minutes to get a lock. In fact it could take 12.5 minutes to download the current ephemeris. Believe it or not, each satellite only sends data at something ridiculous like 50 baud (not 50k baud... thats 50 baud!). Part of the data message is ephemeris, and only a little at a time is sent from any single satellite. 1 sat can send entire epheris in 12.5 minutes.
What I do sometimes is put a 3V lithium CR123A cell on my GPs and let it sit with clear sky view for 20 minutes. The Locosys GPS in larger form factor (like LS20031) have a status LED that flashes when locked. even if it locks in a few minutes, let it sit for 15 or 20 minutes.
Folks,
I have here a locosys LS20033 that I want to use to feed my osd applications with gps info.
I receive it today, connected to my pc serial port through a max232 interface.
the little receiver don't lock/fix sats, nor produce any nmea sentences. This is my first gps module receiver, so I don't know what to do with it.
Could the module be broken from the start? but minigps_1.32 show in it's first applet that it has no fix with the rest of the fields with zeros. the second applet don't show any sentences with nmea code, can I conclude that the module isn't producing any?
Sorry to bump in this blog, but after reading the entry above, I decide to ask you folks for your help and knowledge
Thanks In advance
JAlves
So... get the LS60033 and poof like magic you have what looks like LS20033 with the ceramic patch wanted by Jhon but Atmel chipset inside. Also, for those that want same form factor, Locosystech even makes SirfStarIII chipset GPS in their LS2002 series: http://www.locosystech.com/product.php?zln=en&id=4
Locosys does not = Mediatek by default.
I have EXTENSIVE testing of the 5Hz MediaTek LS20033 and LS20031 in AttoPilot... The larger LS20031 outperforms the smaller antenna version (LS20033) hands down in controlled experiments for solidity of GPS lock and retention of lock in steep banking turns (45+ degrees)... you only add 7 grams going from LS20033 (9g) to the LS20031 (16g). Key for both Media Tek units (and the ETek version) is smoothest possible power filtering to the GPS.. it is a serious requirement. Give this unit its own voltage regulator, and appropriate filter caps, and maybe even a ferrite toroid to wrap the GPS cable around 4x times.
The current Locosys GPSs (LS2003X series) are based on the MediaTek chipset, which do not support binary mode. The one you linked to is an older SiRFIII model.
But if you can write an efficient 5Hz NMEA parser we'd love to use it! We're tried most of the tricks, including serial interrupt, and it's still a bit of a processor hog.
I'm also looking into a chip to embed on a dsPIC based autopilot board.
What I read on the Locosys website is that most of their chips do support UBX or sirf binary mode: http://www.locosystech.com/product.php?zln=en&id=1
If not, I have no problem in parsing NMEA in a few clock cycles on serial interrupt. Just do it one character at the time in a state machine. Only converting the ASCII to float takes some time, but then again the navigational computation is an order of magnitude more intensive so it doesn't really matter..
Does anyone have experience in obtaining a quote for these units in small volume? I doubt Digikey will carry them. Today I work with an old GPS mouse I had laying around for testing.
The problem with the Locosys chipset is that it doesn't have a binary mode, so you've got to run an ASCII NMEA parser at 5HZ, which takes up a huge amount of processing power. I don't think we can do it with a single-core Arduino and still have enough power to do everything else, which is why we use binary mode on the EM406 and will probably do the same with the U-blox.
Comments
I want to say that I manage to put the lycosys working, the increase of Vcc to 4,5V did the trick.
The LS20033 defaults to 9600 baud, 8N1. I assume your max232 interface is setup for this?
First power up it might take many minutes to get a lock. In fact it could take 12.5 minutes to download the current ephemeris. Believe it or not, each satellite only sends data at something ridiculous like 50 baud (not 50k baud... thats 50 baud!). Part of the data message is ephemeris, and only a little at a time is sent from any single satellite. 1 sat can send entire epheris in 12.5 minutes.
What I do sometimes is put a 3V lithium CR123A cell on my GPs and let it sit with clear sky view for 20 minutes. The Locosys GPS in larger form factor (like LS20031) have a status LED that flashes when locked. even if it locks in a few minutes, let it sit for 15 or 20 minutes.
I have here a locosys LS20033 that I want to use to feed my osd applications with gps info.
I receive it today, connected to my pc serial port through a max232 interface.
the little receiver don't lock/fix sats, nor produce any nmea sentences. This is my first gps module receiver, so I don't know what to do with it.
Could the module be broken from the start? but minigps_1.32 show in it's first applet that it has no fix with the rest of the fields with zeros. the second applet don't show any sentences with nmea code, can I conclude that the module isn't producing any?
Sorry to bump in this blog, but after reading the entry above, I decide to ask you folks for your help and knowledge
Thanks In advance
JAlves
Locosys does not = Mediatek by default.
I have EXTENSIVE testing of the 5Hz MediaTek LS20033 and LS20031 in AttoPilot... The larger LS20031 outperforms the smaller antenna version (LS20033) hands down in controlled experiments for solidity of GPS lock and retention of lock in steep banking turns (45+ degrees)... you only add 7 grams going from LS20033 (9g) to the LS20031 (16g). Key for both Media Tek units (and the ETek version) is smoothest possible power filtering to the GPS.. it is a serious requirement. Give this unit its own voltage regulator, and appropriate filter caps, and maybe even a ferrite toroid to wrap the GPS cable around 4x times.
http://diydrones.com/forum/topic/show?id=705844%3ATopic%3A49807
But if you can write an efficient 5Hz NMEA parser we'd love to use it! We're tried most of the tricks, including serial interrupt, and it's still a bit of a processor hog.
What I read on the Locosys website is that most of their chips do support UBX or sirf binary mode: http://www.locosystech.com/product.php?zln=en&id=1
If not, I have no problem in parsing NMEA in a few clock cycles on serial interrupt. Just do it one character at the time in a state machine. Only converting the ASCII to float takes some time, but then again the navigational computation is an order of magnitude more intensive so it doesn't really matter..
Does anyone have experience in obtaining a quote for these units in small volume? I doubt Digikey will carry them. Today I work with an old GPS mouse I had laying around for testing.
-
1
-
2
of 2 Next