Climate Change

[pbxcomm]

Hello all,
I would like to take a moment out of my busy day to change things up a bit.. First, as an owner of two Phantom Pro Plus going on a third, I've experienced my share of misfortunes, firmware or otherwise. That said, I would want to read some positive experience's updating firmware and other. I am a firm believer that there is more to learn from users with successful updates and long flight times. I've pulled little from threads about firmware failures,,,, I too have sent a few. I also believe there may be a common thread with those who have had 100% success rate with updates as opposed to those who have not,, as myself. A few things come to mind:

What was the original firmware version prior to updates?
What device (IPAD ect.) with what version?
USB2 or USB3?
What OS version used, Windows 10x32 or Windows 10x64, Home or Pro?
Were successful firmware updates performed with a MAC or Windows?
Were successful firmware updates performed primarily through Assistant 2, Remote Control or both?
Was an OEM USB Cable used?
Did the end used follow DJI's update instructions verbatim, or was a different procedure used that yielded better results?

Have a missed a few?

These questions to name a few should be asked of us during creation of a repair case number, which are not but useful to DJI to narrow down failures and the cause and affect.

So, lets make an attempt to change the trend..

Thanks..

 

[GMack]

I've wondered if these messed-up updates could be caused from lost packets off the DJI server leading to a corrupted BIN/Firmware file being installed?

For fun, I pinged the DJI website and Amazon and found the DJI site will often give me "Lost packets" (25% loss) where Amazon is solid with no lost packets like below:

 

[pbxcomm]

Good stuff.. I byte of data loss on a firmware update is considered a failure. I was monitoring my network throughput I/O, noticed last time I attempted a firmware update, I received a failure to connect to FTP or connection dropped. This is where I confirmed what I suspected the firmware is coming form the DJI FTP site. If this is the case, then ANY procedure or update process will fail.
I would want to ascertain the FTP site to test connectivity on an average.

Interesting..

 

[GMack]

Fwiw, I could not get the NFZ update to install. Tried doing it through Assistant 2 as well and it failed to get to the second NFZ install screen with it too.

I decided to do a Refresh while in Assistant 2 of the same version in the P4. It seemed to do the same download and refresh of the same version, but once I completed it, the NFZ installed as it should. Was it a lost packet in the first firmware download? Dunno, but a refresh and another firmware download and install seemed to fix the failure of the NFZ update.

Maybe DJI should have Amazon host their firmware updates?

 

[anotherlab]

I've wondered if these messed-up updates could be caused from lost packets off the DJI server leading to a corrupted BIN/Firmware file being installed?

For fun, I pinged the DJI website and Amazon and found the DJI site will often give me "Lost packets" (25% loss) where Amazon is solid with no lost packets like below:

View attachment 88519

While that is possible, I don't think it's a likely cause. File transfers are pretty robust these days. If a packet gets dropped, the process that requested the file basically says "hey, I didn't get that last packet, please resend". That much is handled in low level code and the DJI programmers don't need to worry about that. What they can also do is calculate a checksum value for the file. Kind of like a fingerprint of the file. The checksum is sent over before or after the file is sent over. Then the downloaded file is compared to the fingerprint and if they don't match up, then the file is corrupt or has been altered.

What is more likely the the problem is that there is a bug in the firmware update itself or the upload to the AC is interrupted, leaving it in a fragile state.

 

[pbxcomm]

Correct on the Checksum info.. Some failures occur at the "Transfer" and some on the "Download" before reaching the "Update" segment.
I assume: The Download is to the PC, the transfer is the firmware to the AC and then the update. I'm going to also assume that the checksum is at download, however something may fail at transfer, USB, or AC not entirely ready? If this is the case then the update will likely fail. Perhaps we should be factory resetting the AC prior to any firmware.

 

[pbxcomm]

I'm going to run a checksum program during download tonight and see what I see.

 

[pbxcomm]

Anyone with a good success and attrition rate on firmware updates care to chime in here? Has the firmware process failed where you need to re-start or was the process 100% complete, 100% of the time? If so, I imagine we should be looking elsewhere.

 

[pbxcomm]

I'm thinking if there is a bug in the code or source code specific segment, then I would think that it would affect all AC that ultimately receives this code/file. That also does not explain the fact that there are as many success updated firmware as there is failures.

 

[GMack]

I'm thinking if there is a bug in the code or source code specific segment, then I would think that it would affect all AC that ultimately receives this code/file. That also does not explain the fact that there are as many success updated firmware as there is failures.

Some of the "Success stories" may be premature too. I've been disappointed in making that comment "It works for me" and then comes Day 2 and it flops. More so when I think, "Hey, I got a good DJI product for once not needing to be returned" and then it too disappoints someplace in their sloppy quality control (e.g. A tilted sensor.).

Back to the downloads, I've had some downloads report a file size of say 56.6MB and then the next one is 56.8MB. Something happened someplace for same file on the same day, and I'd think they'd be the same size.