Powernumpty

Has it been well cooled?

Things stopping after an hour or so are normally thermal or bad disk in my experience.

Assuming you've not been thrashing the on board storage I'd look at cooling.

If the cpu/board has been heat cycling strongly the solder might have crystallized and when hot there is enough expansion to disconnect a pin or two. To check I'd let it cool (to the point it started when you last got an hour run) then boot it with less cooling see if it fails sooner, they should throttle back but I be careful just reduce the cooling not risk running without.

If it is disk I think there are images for the BBAI to run from SD, disconnect the cape and test that for while.

Stu

Could you print the output of (command prompt)

"ping kiwisdr.com" (I'm looking for if it returns the IP address, and in what format)

And then

"ping 50.116.2.70" (to see if it can actually route packets to the address)

--

The common issues are DNS where the name does get linked to the right address or more recently the address returned for "somesite.com" coming up as IPv6 where these is no route using that version from your location. Both are errors but it would help to narrow it down.

Stu

I have some FrSky Smart Port voltage and current sensors, they output 3.3V serial data (inverted).

https://code.google.com/archive/p/telemetry-convert/wikis/FrSkySPortProtocol.wiki

The only thing stopping me using them is the dense bit of wet carbon between the keyboard and chair.

As far as I can tell from a web search that might be a SR400AC (giving us the FCC ID: might help to check that).
If it is that model, it's input is marked as 12V 3A which should give decent range of PSU's to test.
If a clip on or two didn't do anything then larger FT-240 with multiple turns should show if it is coming from the leads.

Personally I like a low power main router, something decent that sips electricity as it is harder to spray energy around if it is hardly using any, then get a separate WAP (or mesh?) and move those around as best suits your location and wireless coverage. One other option is to move that router away from the incoming data and bridge that gap with fibre to ethernet adapters. In the UK VDSL is ubiquitous and noisy, I have the VDSL-Ethernet modem then Eth-fibre-fibre from there, means the VDSL lead is kept short (with ferrite) and the rest of the stuff has no "long-copper" links to risk radiating (or injecting noise).

(I'm going to have to put a personal web page up somewhere to avoid posting the same "fibre, ferrite.." comments on all threads, at least I didn't mention a brand of routers this time.)

I shudder to see some of the things sold now. I forcibly get to test a new one very week or two as a close neighbour buys a budget item, kids break it then buy another, worse. The other side gets a new bit of computer kit to review on the same schedule. I've become like a wine taster now - "ah PC PSU, thick odour of Corsair with DC extensions, underlying LED strips and VDSL" - absolutely revolting!

OK so about the coax, antenna and items close by? We need to know what to discount as "recently tested good" (ideally on another system).

GPS is not a massively strong signal if something is wiping it out you would see what you do now, have you tried a phone app like "GPS Status" (Android) to check all is well locally? (bear in mind mobile phone GPS's are horribly good now). I just started the app on my three year old phone and it had about eight satellites indoors (from cold) in the time it took to type the last two lines.

From previous experience I would be surprised if it is software as there are many more physical things that can break GPS than software if it works to some extent.

I believe 5.5mm OD 2.1mm ID

2.5mm ID is fairly common and will cause issues.

I keep a few adapters from universal power supplies to check and/or adapt.

Regards

Stu

The "Quickstart" if printed (A4) is about 37 pages now.

How about an FAQ (forum) post with the most frequently asked questions that could have a short answer but then also point to the relevant QS sections or helpful forum posts.

some FAQ candidates

1. Why is it not responsive five minutes after first boot?
2. PSU?
3. I've broken the software, how to recover?
4. How to edit.....bands, markers

FYI The forum links in the quickstart are to the older forum.

I still think the mid-plate build is a good route for the AI. (if not in the original case)

Yellow for Kiwi+AI, Green for AI. Purple just to hold the lower 6x40x50mm plate on (holes all wrong for best heat transfer without distorting the plate here)

Red not used.

60mm Kapton tape over the purple holes (with screws) and entire length of the plate, 0.8/1mm fibre washers under the KiwiCape.

Only thing to watch for is the DC socket pins, might want to notch the plate there.

I just haven't worked out an inexpensive way to put that sort of plate in the proper case getting the heat out but not having to modify the shell. Obviously depending on the amount of material in contact it still needs some air movement but many times less than small heatsinks.

Plus I do believe there are noise gains to be had, don't have hard figures just gut feeling.

Stu

I have found the MLA-30 didn't work powered from the BBG USB port here but I had assumed that was some failure of the design rather than no supply to the port (will have to go get one and check..) .

As the power injector for the MLA has a 5V-12V step up I didn't try again seeing that as a bit of an issue for HF. If you want to use it with the boost I'd split out the 5V before the Kiwi and try it there, the other advantage of doing it that way is some filtering between the MLA boost switcher and the RF side.

Personally I would use another Bias-T and feed a linear voltage to the MLA avoiding the boost, it would probably run from a 9V battery that way.

73 Stu