Monday, December 1, 2014

No GPS (again)

Drove down to field to fly. Clear sky day, winds 10mph, Kp-Index is 2 (good). I flew my smaller v262 quad and then got the Nova out.

I couldn't get a GPS-3D-Lock at all. DroidPlanner was reporting sats=0/hdop=100 after 15 minutes on flight line. Tried power-cycling and even a new battery.

Apparently, with GeoFence turned on ... if you have no GPS-3D-Lock you can't Arm at all or even fly in manual (I couldn't Arm in Stabilize either). You can't change settings like that with DroidPlanner-1. Without my laptop at field (to turn it off) ... no flying for me today. I packed it up and drove home.



In MP/Terminal/test/gps it gave me the response "ublox ok". ScottFlys (from RCGroups forum) helped me and pointed-out that my Logs revealed the GPS-Module seemed to be communicating with APM-FC but there was no satellite data. GPS Status codes were all "1" (trying to acquire sat data) after many minutes. It was acting like maybe the ceramic gps-antenna was bad or broken.

I considered turning-off GeoFence for now in case problem re-surfaces while at field. However, instead I will just fix problem. As long as I have a minimal GPS-Lock, it will still Arm. On second-thought, just turn GeoFence off while trouble-shooting GPS problems (on less variable).


BeStar BS-126 with ubx-g6010-st U-Blox chip

Upon opening the Nova, I found everything looking visually ok. The GPS-Module's blue-led was flashing (so it was getting power). The gps-module cable was plugged in all the way on both sides (the Molex PicoBlade side at gps-module and RC-Plug side at FC). I removed cable and did a rudimentary continuity test (using only exposed outer contacts) with my ohm meter ... all 4 wires passed. It's a simple cable... GND-TX-RX-VCC.

 
I re-connected it (didn't really do anything to it) and GPS-3D-Locked in seconds after powering Nova on. I was getting sats=11 and hdop=1.3 very easily. However, if I wiggle the wires and connectors it goes to sat=0/hdop=100 again. Appears to be a good old-fashioned intermittent bad connector or wire. I'm thinking on the RX or TX lines since the GPS-Module is getting power the whole time (it's blue-led is always flashing).

Inside APM FC
Let's start at the beginning. I opened the FC and checked the 4 gps header pins. All clean and shiny metal. Their solders to PCB also look ok (not great, but acceptable). All these examinations are under heavy magnification with lots of light.

Cable: RC-Plug/Dupont (FC side)
After documenting the whole cable's pin-outs, I carefully backed-out the pins and examined the crimps under magnification. They looked fine. I didn't see any corrosion or tarnishing (contact points were shiny). I couldn't really find anything wrong so I just adjusted (very slightly bent contacts) the pin-sockets in this end of the cable to be tighter and cleaned with contact cleaner. Pins do appear to be fairly low-quality metal but ok.

Cable: Molex PicoBlade (BeStar BS-126 GPS-Module side)

First thing I notice is that the white nylon plug is poorly made/moulded and not a genuine Molex part. It is also missing the Molex patent marks. I carefully backed-out the pins. Second thing I notice is that the pin-sockets are very low quality clones. They are obviously not genuine Molex parts either. They also appear to also be poorly Chrome-Plated ? Since when are contacts chrome plated? Shouldn't it be something like gold or even tin ? There is also a black tarnish inside them where the contact points actually touch the socket's pins (appears to be a bad "un-like metals" reaction). Surprised it was making any contact at all. Looks like I have found the root problem.

Zoom-in the see spots of black tarnish on pin blades.
It's the same on other side of pins.
BeStar BS-126 GPS-Module (PicoBlade socket)
Found similar black tarnish on both sides of pin "blades". But only on the 4 pins that the cable uses (so it's obviously a unlike metals reaction). Comparing to some new genuine Molex PicoBlade sockets (direct from DigiKey), it's obviously a cheap clone. The pin's actual shape is a little different and you can let the blade-pin's composition metal is not the same as the genuine part.

**************************************


I have ordered parts to make a new cable. I will try to clean (poorly cloned) PicoBlade socket's blade-pins on GPS-Module and salvage it.

**************************************

I only have about 8 flight-sessions on my Nova. During that time I have experienced 2 complete failures of the GPS-Module (to find ANY satellites) and 2 attempted fly-aways during GPS dependent RTLs. At the time, I chalked them up to sun-spots, CME, and/or poor Kp-Index (general GPS glitches). One I even thought might have been me flipping the wrong switch (and/or incorrect FrSky radio switch programming) while trying to invoke FailSafe condition. Due to the intermittent nature of this problem, it's taken a while ... but it looks like I have finally found the root cause of those failures.

 
GPS Module Trouble-Shooting:
Here are some notes. I will flesh-out these notes if I ever have to do more GPS-Module TS or decide to replace/upgrade mine completely.
 
Prerequisites:
  • Test GPS-Module outside (on a clear-to-partly-cloudy day) so it's ceramic antenna has a clear view of sky (and satellites). One less variable to worry about.
  • Give it at least 30 minutes to acquire a signal before you give up.
  • Verify cables and pins are good with a multi-meter (voltage and ohms)
  • I think GPS works on radio frequencies 1.2ghz and 1.5ghz . Nothing else should be running on those frequencies or you will get interference.
GPS Module's Blue-Led: Means GPS-Module is getting power and I think (maybe) also initialized and ready/working. Can't go much further without this.
GPS Module's Red-Led: This led is mirrored by Nova's external Green-LED. Flashing for "trying to obtain a GPS 3D-Fix" and solid for a proper 3D-Fix.
 
You can connect Nova's stock uBlox GPS-Module directly to laptop with an Arduino FTDI-to-USB adapter. Use uBlox's u-Center for Windows software. This will verify GPS-Module is good without any help from Flight-Controller. uBlox GPS-Module must have APM config file loaded (only needs to be done once ... unless it gets corrupted somehow).
 
If you ever connect VCC power to GPS-Module backwards (reverse polarity) you have likely fried it completely. You might try removing the battery (or is it a super-capacitor) or otherwise master-reset it's memory.
 
To estimate what GPS you should be getting:
 

3 comments:

  1. Hi, Great site! I've been looking for some information on this. I think i fried mine by connecting it the wrong way round when trying to rehouse the GPS.
    Though I do have the blue and red flashing LEDs on the GPS I dont have any green light on the Nova or any GPS info in APM. Should I try getting a new module or do you think it could still be the cable?

    Thanks,
    Mike

    ReplyDelete
    Replies
    1. Use MissionPlanner and Terminal Mode (if running v3.1.5-custom or lower). See what hdop/sats you get out in the yard. If it still can't detect or work 1. Try a new one. 2. Connect questionable uBlox GPS-Module to laptop for diags and possible reset/reprogramming.

      Delete
  2. Thanks for the quick reply.

    I saw in missionPlanner there was no hdop/sats. I've ordered a new GPS unit but it will take a while to get here.

    Is there a chance it could be the flight controller? I can still bing and start the machine I just dont have the GPS.

    ReplyDelete