GNS430W "Kernel error"

TimRF79

Pre-takeoff checklist
Joined
Dec 10, 2017
Messages
352
Location
Houston, TX
Display Name

Display name:
Tim
Last Sunday after I had my $100 Hamburger, I went back in the plane and taxied to the run-up area.
During that time the 430W had a "Kernel error", screen all black with the words in yellow.

Turned unit off and back on and it was fine.

Theory: It was noon in Texas and I do not have a windshield cover, the unit may have gotten to hot.
Anyone else have experience with such a failure?
 
Never seen that one. Does this suggest that the GNS430 is Linux-based?
 
If you’re unaware, the kernel is pretty much the lowest abstraction level of software in any box. It’s the thing that talks to the hardware directly.

So keep an eye out for hardware failures.

Also it really wouldn’t be a bad idea to talk to Garmin tech support to see if the box writes what’s known as a kernel dump (yeah I know) when it has a kernel crash that might give Garmin a clue as to what failed.

They may want a copy of that. And it might give you some peace of mind if they came back and said your unit hit a software bug that they need to fix. Or they might be able to say what hardware is flaky or possibly failing.

Another simpler possibility. How hot was it? Since we are in summertime it may have highlighted a need for an avionics fan in your particular installation and the box is warning you by becoming unstable.
 
Also it really wouldn’t be a bad idea to talk to Garmin tech support to see if the box writes what’s known as a kernel dump (yeah I know) when it has a kernel crash that might give Garmin a clue as to what failed.

They may want a copy of that. And it might give you some peace of mind if they came back and said your unit hit a software bug that they need to fix. Or they might be able to say what hardware is flaky or possibly failing.

Another simpler possibility. How hot was it? Since we are in summertime it may have highlighted a need for an avionics fan in your particular installation and the box is warning you by becoming unstable.

Ill ask Garmin and see what they say. Not really in the mood to send int he 430W and pay for a repair...
But it was toasty, I think I need to get a windshield deflector
 
If you’re unaware, the kernel is pretty much the lowest abstraction level of software in any box. It’s the thing that talks to the hardware directly.

So keep an eye out for hardware failures.

Also it really wouldn’t be a bad idea to talk to Garmin tech support to see if the box writes what’s known as a kernel dump (yeah I know) when it has a kernel crash that might give Garmin a clue as to what failed.

They may want a copy of that. And it might give you some peace of mind if they came back and said your unit hit a software bug that they need to fix. Or they might be able to say what hardware is flaky or possibly failing.

Another simpler possibility. How hot was it? Since we are in summertime it may have highlighted a need for an avionics fan in your particular installation and the box is warning you by becoming unstable.
Cosmic radiation flipped a bit or two? I wonder what a 430 does to verify firmware on startup. If it's something stupid like checksum or CRC32 that sort of thing can slip by from time to time but I'm not familiar with what they had available back then when they wrote the code.
 
Well the unit started up, then during taxi it had the kernel error.
Off/On and it was fine.
After 10 min of flight it activated the next leg and everything worked except for the right dial push button.
Then Off/On and it was normal again.

But yes, I do live in Texas, it was Mid Day and the plane had sat on the parking lot for 1 hour without a windshield cover
 
Cosmic radiation flipped a bit or two? I wonder what a 430 does to verify firmware on startup. If it's something stupid like checksum or CRC32 that sort of thing can slip by from time to time but I'm not familiar with what they had available back then when they wrote the code.

Yeah me neither. That’s why I was suggesting a call to them just to see if they think a kernel panic is a bad thing or if they even have diagnostics for it on that particular model and operating system.
 
Never seen that one. Does this suggest that the GNS430 is Linux-based?

While I do believe that Garmin does use a Linux based OS, the fact that it has a kernel does not necessarily mean that as I see denverpilot has already pointed out.
 
While I do believe that Garmin does use a Linux based OS, the fact that it has a kernel does not necessarily mean that as I see denverpilot has already pointed out.

That I know from building and configuring workstations with various OSes, but the "kernel" terminology is typically used in conjunction with a *nix system. Just curious about what the underlying software is based on. Hopefully not something as arcane as the Nicolet proprietary oddball (17-bit?) processor operating system. (This scientific instrument vendor is now out of business--I hated their hardware and software systems.)
 
I previously worked for Panasonic Avionics and Gogo. Both of which use Linux as their OS. Most of the common Linux commands work for their implementations but not all. They also have their own specific commands.
 
While I do believe that Garmin does use a Linux based OS, the fact that it has a kernel does not necessarily mean that as I see denverpilot has already pointed out.

You can assume any modern avionics are based on some kind of Microkernel RTOS, such as VxWorks, LynxOS-178 or RTOS-32...

That I know from building and configuring workstations with various OSes, but the "kernel" terminology is typically used in conjunction with a *nix system. Just curious about what the underlying software is based on. Hopefully not something as arcane as the Nicolet proprietary oddball (17-bit?) processor operating system. (This scientific instrument vendor is now out of business--I hated their hardware and software systems.)

They’ve used NT 4 on older devices, and “kernel” isn’t really OS specific.

Someone posted yesterday how to see the NT kernel boot process go by on one of their devices, upside down even. ;)

The more esoteric RTOSes are common in avionics but not ubiquitous. We also used them in high density telecom gear, and my niche long ago was OS-9, which we ended up switching to VxWorks on the successor platform hardware to that, later on.

But we also made one that ran partially on OS/2 Warp. LOL.

Always a fun day in the lab when a customer says your box failed the “answer all phone calls within 100 ms of delivery” in a full up system of multiple boxes and card cages in each, that they tossed 10,000 phone calls at. Hahaha...

Let’s see... Which engineer wins the “you broke the spec rules in your RTOS code and didn’t set the priority of your processes correctly award” today...?

I sometimes miss that lab. Lots of fun toys in it. :)
 
I would think something like the 430 would have over temp detection and some type of controlled indication and shutdown. Our SIL3 safety devices have this. The kernel errors are nasty. As it went from cooking in the sun to frying with the micro(s) up and running some peripheral (probably memory) hiccupped. It would be interesting if Garmin would share which RTOS, which SIL/PL level they meet, how many non-critical micros the run and how fpgas/asics fit into their solutions. I feel like having one of my crew do a full tear down just for the heck of it :) Anyone have a dead 430 (preferrably a 750!) they want to donate to science :)
 
I would think something like the 430 would have over temp detection and some type of controlled indication and shutdown. Our SIL3 safety devices have this. The kernel errors are nasty. As it went from cooking in the sun to frying with the micro(s) up and running some peripheral (probably memory) hiccupped. It would be interesting if Garmin would share which RTOS, which SIL/PL level they meet, how many non-critical micros the run and how fpgas/asics fit into their solutions. I feel like having one of my crew do a full tear down just for the heck of it :) Anyone have a dead 430 (preferrably a 750!) they want to donate to science :)

I wouldn’t share that info with you if I made the thing. You could, as you mentioned, but one and open it up, but if you’re not a regulator, I’m not publishing how I built it unless you’re placing a really big order of them. Then maybe. LOL

As far as the temp sensor goes, we had some gear in telecom that was considered critical enough it had to always be “best effort” so it had no thermal shutdown. It tried to run no matter what, even if it was destroying itself or lowering its lifespan in a central office that had lost cooling.

The thing it did have to do and meet a standard on, was if it caught on fire it had to self-extinguish.

The old joke error message in Unix that the line printer was on fire, was true in telecom. :)

The much more fun ratings were the ratings for outdoor gear. How many shots from various firearms they had to withstand and keep operating. Hahaha. We had a fun day in the parking lot testing that one at a different company that did outdoor plant gear. The specs even had details about such things as slugs or buckshot and such. So we obliged

Nothing more cathartic than a support team blasting away at the company’s product with firearms. Every hardware team should get to do it! Hahaha.
 
I spent a fair number of years doing RTOS with VxWorks...there were many times I wanted to shoot something...is it possible to shoot the debugger :)
 
Yep...Page 76...thanks for sharing.

For our most critical systems we also "do our own thing". Then again, maybe Garmin is just using a spin off of a Linux variant of a [Safety] RTOS and can then call it their own.

A couple that make stuff that doesn't list using a custom:

Dynon: Linux
L-3: Linux and Windows
 
Yep...Page 76...thanks for sharing.

For our most critical systems we also "do our own thing". Then again, maybe Garmin is just using a spin off of a Linux variant of a [Safety] RTOS and can then call it their own.

A couple that make stuff that doesn't list using a custom:

Dynon: Linux
L-3: Linux and Windows

Yeah I very much doubt they really built an OS from scratch. It’s a pain in the ass.
 
You can assume any modern avionics are based on some kind of Microkernel RTOS, such as VxWorks, LynxOS-178 or RTOS-32...
You can assume that, but you'd be wrong. The MX20/GMX200 runs NT4.0.
 
You can assume that, but you'd be wrong. The MX20/GMX200 runs NT4.0.

Do you remember the Navy attack ship that had to be towed back to Norfolk because its primary systems Were on a WIN NT system and it suffered a total shutdown?
 
Yes, but that was more involved than the use in the MX20.

The other amusing MX20ism is that the later ones (the ones with a rocker switch power/brightness control) and all the GMX200's all have the display running upside down.
You can watch the windows boot (if you manually turn up the brightness at power on) and if you get a blue screen, you'll see them both upside down.
The original MX20's (round brightness control), ran right side up and you could see the windows boot clearly (the later ones left the screen dimmed until the MX20 software itself started).
If you use the CF card from one unit in the other, the whole thing ends up upside down.
 
Back
Top