Dadcat said:
Dalite said:
Very likely a dead Battery Backed RAM battery on the GPS board.
This would also cause the real time clock calibrations to be performed at the beginning of each flight, as well as almanac data being lost whenever the flight battery is not energized.
That data should be retained in the GPS module by the backup battery. If it is below 1.8v, it is not reliable to retain the data.
That's enlightening.
I doubt this is the current version of the GPS board. Looks like a battery on the right hand edge of the board. Is that right?
Please forgive me if this has been discussed before, but could it be that we've all been shipped brand new Phantom 2s with dead GPS ram batteries and that's why we have a J-hook or TBE etc.?
Couldn't be that simple. Maybe I should be posting these questions in the "Other bad ideas" thread. People have been returning Phantoms to DJI for months for this problem and getting replacements that do the same thing.
Has somebody already tried changing the battery on the GPS board and found it doesn't do anything about the compass problem?
I cannot say positively either way. Here is where I am on the subject...
When I got my first Phantom I was determined to find out things that could cause uncontrolled flight.
The first think I noticed was the longest cable went from the lowest point to the highest point, and it was attached to the compass on one end and the GPS board on the other. it was also multi conductor and a fairly effective antenna for induced interference.
In trying to find ways to introduce shielding. I fell back on the oldest trick in the book; which was twisting the cable. In doing so, I uncovered the GPS board and battery. Naturally, all batteries must be tested to see if they are good. Mine measured .025 VDC.
I wrote down the info from the GPS module and started doing some research. That is where I learned about the battery function. I was concerned that the TCXO ( temperature controlled Xtal oscillator) was drawing the battery down, as the conventional approach used to be to keep a crystal time base at a constant temperature to maintain the most accurate frequency synthesis. I contacted an engineering source familiar with the design, where I learned that the TCXO was used in maintaining the real time clock, and was not active when the main power was not available. But, the time calibrating the real time clock as the TCXO came up to temp was supposed to generate data stored in battery backed ram. So, it is possible that a dead battery requires total recalibration at every flight. It is also a possibility that GPS almanac data download further introduces delay, as it has to be downloaded each time. If the battery was above the minimum voltage, the calibration data would be present at power up to keep the RTC accurate from the get go, and the almanac data would be there to aid the GPS by taking the RTC info to determine which sats shoud be in view; a warm start situation.
All of this is no more than hypothesis on my part. I don't know is this is a proper answer to your questions, but maybe it will provide some useful hints.