Stability for DC v3.03 and Skylark v5.5

I am new to using the Dreamcatcher… having only recently purchased the v3.03 hardware, and am running the new v5.5 software.

I am wondering what the general accepted rule for up-time is on this combination? I don’t think I’ve exceeded 48 hrs up-time without experiencing a lockup requiring a reboot. I don’t know if I should start troubleshooting my power supply, re-flash the SD, or if this is expected performance. I don’t see it discussed much… except in relative terms in the forums to the older hardware, and since I haven’t run the older versions of hw/sw, I am a little unsure how to interpret the up-time performance I’m seeing.

My USB power supply has the correct name-plate ratings (2.1A) but the output is a tad low (like 4.9V). I could try a lab supply pushed up to 5.1V if the hardware is that sensitive to low voltage. I have poked around on the SD card, trying to figure out what I thought were stuck downloads (but came to read and find out it’s repeated files sent over and over for redundancy) but the software never had immediate issues with anything I did on the SD card (like deleting a “stuck” file or zindex). News and Wikipedia entries are filling up properly it appears.

Side question hopefully not to derail the topic, when do copies to the backup SD occur? I only notice files being copied over after I do a reboot.

Thanks for any data-points or suggestions,
Will

Hi Will - - the older versions of Skylark prior to 5.5 did lock up often, but the current 5.5 runs without problems for weeks +. You point out a low supply voltage at 4.9V - - that may be an issue. If you can push it up, try that.

Does anyone else have any ideas? Ken

Thanks, Ken, for the datapoint that I should be seeing pretty high up-time (weeks+).

I checked the test-points on the Dreamcatcher itself, and it appears you lose another 0.1V across the Vusb ferrite. So the 4.9V I mentioned earlier was looking more like 4.8V at the board.

I swapped in a different 2A USB supply… it’s about 5.25V at the output, and 5.15V on the DC board (with about 1A being drawn). I’ll post in after a few days if things get better, for forum-searching sake if anyone else runs into this.

Will

I have seen months of uptime with a commercial/cheap 2A power supply.

first night i left a session open on a browser, and i’ve had a lockup. found it this morning with only the ‘busy’ led dimly flickering and DC unresponsive on the LCD and browser.
restarted DC and it’s back.

Another problem, Will, is that some folks have is their USB to micro USB cable. Even with a powerful power cube, the cables are or become flaky. Use a short quality fast charging sync/power cable. I had what I though fit the bill, but it was deficient. Replaces it, and all was back to normal. Ken

my USB cable has actual markings on the jacket - “USB2.0 CABLE AWM 2725 VW-1 80°C 30V”

Sounds great, but it doesn’t necessarily mean its a good enough cable. I know this sounds a bit crazy, but we Forum members have been thru allot over the years, and cables have become a big black hole!

So try all your cables if you have a problem with the Dreamcatcher 3.03 running Skylark 5.5 rebooting. The current configuration seems to run forever as Syed said. Ken

mine is not rebooting. i’ve been 'round-n-'round with cables, too. it’s from a reputable source, and is heavy-weight. it is no tinsel wire cable.
when my DC is unresponsive (in the way i reported), the reset button does not work, either. it takes about five hours or so before it locks up.
seems it’s related to leaving the Weather app running in my browser session. is five hours the update interval for new data?

@Syed, @Abhishek
Perhaps in a future board we will get a “power good” indicator led on the board or a selection on the screen. Until then, it is absolutely necessary to monitor your power for at least 5.0 volts on the board. Personal experience tells me that 4.9 will fail and crash, as described several times here. What @kenbarbi says here relating to power cables is gospel truth, and will rule your DreamCatcher! Incidentally, my DreamCatcher has been up continuously for 68 days as of today, no failures, no re-boots.
Jerry

Interesting experiment … leaving the Weather app open.
I have it running (open session) logged in as othernet.
the current ‘what’s-new’ say my /data/weather/2019/04/17/0000 and
I also get 2019/04/14/0000 and 2019/04/15/0000

I think is the three day look-ahead from the file /grib2/gfs.2019041400/gfs.t00z.—.grib2

I will keep it open watch what happens at the next weather file download… they seem at erratic times now.
my previous files are /04/12/000 and /04/13/0000 so 24 hours between files

i measure 5.08V between the USB_PWR shell and the USB_5V test point.

my dashboard trio is: What’s New, Tuner (Status tab), and Weather.

repeatable results so far: roughly 5 hours with trio open before lockup; roughly 10 hours (without Weather) before Tuner Status stops updating (but the clock time continues to advance).

fixes: for lockup - only pulling the power plug since front panel and reset switch are unresponsive; for Status - shutting the app and re-opening it.

@Syed
@kenbarbi
@Abhishek With such symptoms, I would refer you to @Syed, as this sounds more like a defective unit. The Weather is commonly behind, but Tuner normally is spot on until the wifi gets interrupted or the tuner app stops. This does not interfere with functioning of the board, but the heartbeat indication in the software and the tuner indications stop fluctuating, even though the DC itself is still running fine. In fact the other apps still respond and you can still run the diagnostic and read news or weather or wiki. At this point, however, a new login is required to see the Tuner application. The lockup of the receiver itself when the voltage is correct is unusual. When I had lockup problems, I did watch mine with a “Watts Up” metering device and caught a voltage sag that was not apparent in testing the on-board voltage. The on-board voltage as you measured it was 5.02 every time I checked. However, when the WattsUp was left in line continuously for a few days, it found the voltage sagging at some time to 4.9. Current draw was approximately 1.1A average. When we corrected the power supply so that the voltage maintained at 5.05, a sag to 5.02 was detected, but no more lockups. When DC version 5.5 was released and installed, the current draw reduced to .91A average and none of the previous symptoms returned.

the lockups seen here is the panel and reset unresponsive, the over-the-web also unresponsive, the various LEDs on the DC all in a constant state except for the Busy indicator (which is found dimly/quickly flashing). lockups have not happened in over 24 hours when Weather app is not left running. i did not say anything about the how current the data on display in the Weather app was or was not. the data animation continues until the taskbar fills with ‘crossed circles’ [ ⊕ ](synchronization problem? free memory depletion? something else?).

also note the Tuner app Status tab not updating was remedied by closing it and opening it again.

no, my experience indicates what i am observing is -not- a hardware problem.

Another question, then, what size of SD card are you using? Did you format it before burning in the Skylark 5.5 image?

i have twin 64GB cards in place. second card installed a day after the first. first card was auto-partitioned and formatted by DC after writing and verifying (with Etcher) and booting Skylark image.

There is a ‘clean-up’ cron type job that runs periodically. I would remove the second sd card (the one at the corner called SD2_data) and run for a while with just one sd card in slot SD0_OS

I agree - - Syed needs to take action to replace board. Ken

after processing 1.2e6 packets without incident, what happens when an API call in a second web session happens in the midst of an API call in the first session?

i observe what seems to be processing gone haywire in any session once there are at least two API calls appearing on the taskbar ( ⊕⊕ ). this time there was a heartbeat, busy flickering ‘normally’, and the packet led mostly on and blinking about once a second. reset sw active and was employed. all good now.

with the use of a wire mesh horn, my PER is 0.000. gotta catch 'em all! :smiley:

yesterday I ran with two web sessions open, and played with putty serial to the dreamcatcher. I had a similar result two API calls ended up hanging with the ( ⊕⊕ ) until I closed all the sessions. The Dreamcatcher kept ‘running’ — didn’t require a reboot – and I was able to re-establish a web session and the API calls were cleared.

I couldn’t see any processes that were hung in the log viewer