J-Log.eu - Forum

Re: Unexpected freezes?

2016-03-20T22:05:53+01:00

Okay. Still strange.. but that's SD.

Statistik: Verfasst von dl7uae — 20. Mär 2016, 22:05

Re: Unexpected freezes?

2016-03-19T19:51:36+01:00

Hi Tom,
Just for information.
MicroSD card re-formatted, and today four flight in row with full telemetry on card and on Spekrtum TX. No problem!
So, it seems was really card or file system issue.
Thank you!

p.s. The OpenFormat output stream also was removed in the settings

Statistik: Verfasst von 3Daemon — 19. Mär 2016, 19:51

Re: Unexpected freezes?

2016-03-16T00:51:43+01:00

Quote:

Nope! JLog stop send live data to TM1000, all values from JLog frozen! TM1000 has only one "live data", his own RX voltage (BEC voltage).

No, JLog does not stop to send on x-bus than the values of transferred items get frozen.

Quote:

Yep, that my assumption. Because problems appears only after 175+ log sessions. Will check. But... how card recording process may affect data output by X-bus?

As described: JLog's software: We have a main loop and many interrupt (event) driven software components.
Data evaluation/processing as well as log recording is done in the main loop, - whereas data acquisition and data transmission (telemetry, terminal boxes, OF live stream) are jobs of event driven software parts, - pin change interrupts, logic interrupts (UART (serial), I²C bus controller) or timer interrupts.
So the SD may hang the main loop (no processing into data items anymore) whereas the event driven processes continue to serve, - like onto the x-bus.
The result is: TM1000 continues to receive data from JLog (otherwise the displays in transmitter would fall into timeout) but that data, input of the telemetry serving routine, got frozen because the data processing in the main loop got frozen by write attempts on a faulty SD.
Just a theory...

Quote:

Mmmmm??? What's wrong? BTW, I had in use this config 170+ flights w/o troubles!

Isn't directly wrong, no problems to fear from, - just unnecessary. There is no device in your heli what receives async-serial live data in OpenFormat in parallel to the TM1000 receiving data through x-bus.

Statistik: Verfasst von dl7uae — 16. Mär 2016, 00:51

Re: Unexpected freezes?

2016-03-16T00:05:40+01:00

Hi Tomas!

dl7uae hat geschrieben:

You say: log too short but JLog continued to send data to TM1000 - only data got frozen any time? (transmitter's log)

Nope! JLog stop send live data to TM1000, all values from JLog frozen! TM1000 has only one "live data", his own RX voltage (BEC voltage).

dl7uae hat geschrieben:

- SD card has a problem. Just try if it is cured by formatting of the card with the SD formatter from sdcard.org

Yep, that my assumption. Because problems appears only after 175+ log sessions. Will check. But... how card recording process may affect data output by X-bus?

dl7uae hat geschrieben:

- JLog got undervoltage.. -> brownout condition, no chance to get back to normal processing (no POR)

NOT my case, for sure. RX voltage never drop less as 7,2V. And FBL (MSH Brain) has no issues (all on-board devices powered from same Kontronik BEC).

dl7uae hat geschrieben:

If JLog got rebooted on-the-run by undervoltage (Power On Reset) then the log file w/ the next running number is the log continuation within the same session.

NOT my case, just check logs.

dl7uae hat geschrieben:

Where does JLog's supply voltage come from? From TM1000? What is TM1000's power source and what the source of that source?

All on-board electronics powered from same Kontronik BEC

dl7uae hat geschrieben:

Just try what happens if JLog is supplied from a battery for 1..3 flights.

Hmm... will check

dl7uae hat geschrieben:

CONFIG.txt: Hey, OpenFormat live stream is activated in parallel?

Mmmmm??? What's wrong? BTW, I had in use this config 170+ flights w/o troubles!

Statistik: Verfasst von 3Daemon — 16. Mär 2016, 00:05

Re: Unexpected freezes?

2016-03-14T23:53:22+01:00

Hi Igor!

You say: log too short but JLog continued to send data to TM1000 - only data got frozen any time? (transmitter's log)

OK, two reasons conceivable

- SD card has a problem. Just try if it is cured by formatting of the card with the SD formatter from sdcard.org

- JLog got undervoltage.. -> brownout condition, no chance to get back to normal processing (no POR)

If JLog got rebooted on-the-run by undervoltage (Power On Reset) then the log file w/ the next running number is the log continuation within the same session.

Of course there is no continuing log file if JLog got frozen by undervoltage w/o POR, just brownout.
That is imaginable if a hanging SD card (could not initialize after undervoltage) gets two things hang in JLog: main loop w/ the data processing (frozen values), log recording
... whereas interrupt driven processes continue to work: telemetry, I²C bus (x-bus).

Unfortunately SD cards die a bit earlier at undervoltage than the microcontroller in JLog (only discrete settable brownout thresholds - not anyone exactly matching to the electronics in a SD card). The SD card can anyway not initialize w/o JLog's bootloader. So the brownout settings in JLog's processor are useless for these two reasons.

Where does JLog's supply voltage come from? From TM1000? What is TM1000's power source and what the source of that source?

Just try what happens if JLog is supplied from a battery for 1..3 flights.

-----
CONFIG.txt: Hey, OpenFormat live stream is activated in parallel?

Statistik: Verfasst von dl7uae — 14. Mär 2016, 23:53

Re: Unexpected freezes?

2016-03-14T22:13:00+01:00

Hi Tom,
Thank you for answer!
But problem doesn't look like related for X-bus or TM1000. Because the log file on MicroSD card shorter as logging session.
Take a look for log00185.txt and log00188.txt, they are ~03:20 and ~01:45 duration only, but logging sessions should be 5+ minutes each.
So, looks like JLog freeze during logging session, just stop receive and log ESC's data, stop transmit data to TM1000 then.
I'm already double- and triple- checked all connectors.
No, I didn't use any additional sensors. Just JLog my sensor for TM1000.
Anyway, I'll check how will JLog just log, without TM1000,
also maybe MicroSD card write issue could lead unit to freeze?

WBR! Igor

Statistik: Verfasst von 3Daemon — 14. Mär 2016, 22:13

Re: Unexpected freezes?

2016-03-13T23:54:33+01:00

Hi Igor!

Well.., the I²C bus ("x-bus") freezes, not JLog (continues to receive and log ESC's data) than the communication JLog--TM1000.

Please check

- the routing of the x-bus harness ... should not run parallel or close to high current carrying lines

- broken wire or loose contact? (Check the plugs. Crimped wires...)

- possible ground loop?

- undervoltage condition in between? (does not look like from transmitter's log)

Do you use Horizon's voltage or rpm or temperature sensor connected to the TM1000?
Don't connect the ground wire of a voltage or rpm sensor! That's the best way to build up a ground loop.

Statistik: Verfasst von dl7uae — 13. Mär 2016, 23:54

Re: Unexpected freezes?

2016-03-13T18:52:19+01:00

Corresponding, from today's four sessions Spektrum telemetry TLM file also attached.
Could be viewed by this application

Statistik: Verfasst von 3Daemon — 13. Mär 2016, 18:52

Unexpected freezes?

2016-03-13T18:42:34+01:00

I have been using JLog2.6 couple of months and was very happy.
But, unfortunately, last days I see unexpected freezes during the flight/logging session. Can't spot any dependency from temperature or current, voltage, etc... Yesterday first and second flight ok, but next two with freeze. Today backward, only last two sessions ok. Such freezes looks like exactly JLog freeze, dates stop recording into Log file on microSD, and telemetry module continue receive last values. After power-cycling - all became good, then may freeze again or may not.
Kontronic JivePro 120+ HV ESC, Spektrum TM1000 telemetry module.
File "ConfVersLogs.zip" attached.
Any suggestions or ideas "what to check?" and "how to fix?" are very appreciated

WBR! Igor

Statistik: Verfasst von 3Daemon — 13. Mär 2016, 18:42