Remote Kiwi Websdr and regular "power downs" or disconnects
Our team is running several remote Kiwi WEB Sdr's on an island in the Pacific Ocean
and one of them is dropping out on a regular basis -- We think it might be heat related -
as the shack temperature is occasionally exceeding 30 degrees C.
This particular Kiwi is in an metal enclosure -- and its "brother" is in the plastic enclosure and not giving us any issues.
Interestingly, if I "tune" into a station and sit on that frequency and set the TIME OUT's OFF, the receiver does not die or disconnect from the net or "go off the air..."
Thankfully we have a remote antenna switching array on it and a REMOTE power off/power on management model to it is easy to restart.
We've noticed that some folks are adding fans to their Kiwi -- is this because of the sensitivity to overheating? I am aware that the Kiwi really does not like temps of 40C and above.
Many thanks everyone!
Comments
you can check the temp here at the bottom of the stats page
Not for BBG/BBB based Kiwis.
I thought the the temperature display was on AI's only? I assume this is not an AI?
How many channels in use and does you holding it on one frequency stop others using it?
Extra fans don't hurt where ambient is high, even an external blowing over the end plates to move any dead air helps.
I've a passively cooled mini PC here, a large slow fan while silent takes the edge off nicely.
So - what I think I am hearing here is -- Yes, heat can cause a Kiwi to shut down..
Is that an accurate statement?
Good to know! Thank you.
Yes, over 40C ambient has been reported to be a problem. But power quality issues are much more likely. Are the two Kiwis fed from a common power supply? If not, do their supplies or cabling differ?
Separate PS units -- running a custom 5A - 5V supply (LINEAR - BEL PS)
that I have used at home shack for several years -- the one "on Island" is in an aluminum case with (we think) plenty of air flow -- so, yes it is suspect until all other possibilities exhausted.
Funny think - I can tune in to a station and leave it on for 45 minutes to an hour --
When I drop the KIWI connection and return in 5 or 10 minutes, it is gone -- and then I remote PowerSmart APP it back on.
Many thanks!
Does it have a GPS antenna?
Yes - it does have a GPS antenna with good acquisition -
Now at over 90 minutes listening on the oddly behaved Kiwi --and I will bet, without fail, that if I logoff the Kiwi and come back 10 minutes from now, it will have "gone to sleep..." - Could this possibly be a bad ethernet port that is going to sleep with no activity?
Fascinating issue. ;-)
So here's a possibility. When there are no user (SDR) connections the realtime audio process that communicates with the FPGA is suspended. This leaves more Beagle cpu cycles for the GPS acquisition FFT which is very cpu intensive. This is not only because there are more raw cpu cycles available, but because the uninterruptible FFT (~10 milliseconds) can be more frequently scheduled compared to when SDR-related realtime tasks are running (the waterfall imposes similar constraints).
The point is that even with no connections the Kiwi might actually be drawing more current and generating more heat in this condition. So a simple experiment is to disable GPS from the admin interface and see how it does. Uncheck the three "acquire" checkboxes at the top left of the GPS tab. No need to reboot. Check them again to restore acquisition.
my bad on the lack of temp on a non-BBAI...
python has a tmon module that may work for checking its temp. But it should sounds like it's other than temp. My BBAI hover just above 40C when the WX is warm without issue BTW.
The fans wore out on my Maui Kiwis.
So I removed the internal fans, took off the SMA-end metal plate, and positioned two 4" USB powered muffin fans to blow much more air through the case than was blown by the internal fans. As a extra tweak, I stood the Kiwis on their sides so the PC boards are vertical.
There is no CPU temperature readout on those BBG Kiwis, but the cases were cooler than before and I see no additional RFI.
I'd really like to know the outcome of my suggested experiment. If it helps then one thing I could do is lower the duty cycle of satellite acquisition when there are a sufficient number being tracked and there are no user connections.