FlightAware Discussions

Increased plane numbers after software update/upgrade

After changing the power supply on my secondary ‘combo’ for the third time, and going through a couple of +35 degrees in the garage, I think it’s safe to say the so called “RPi Power Supply” from eBay cannot handle high temperatures. It still works fine indoors, in an air conditioned room. Those were the first 2 PSs.

Also noticed, and mentioned previously, performance improved a little with the new PS. That is an expected outcome I think, with a better quality PS.

I asked in the past about software updates increasing performance, as I had observed such a thing. The consensus at the time was that it was not likely. Well… it happened again.

I updated the software on the secondary ‘combo’ through PuTTY, not via the FA page, and after that it started decoding more planes than the primary. It’s been a couple of days now. The primary had been consistently 5-10% better in the past, but the secondary is now a few planes ahead.

The primary ‘combo’ is using the standard 3.7.1 FA image, no updates. The secondary had the standard 3.7.1 FA image, before issuing the sudo apt-get update and then upgrade commands.

Maybe the power consumption was reduced somehow and the dongle is getting more juice!

Honestly though, very strange.

Swap in a 3.7.1 card again and see if it drops again.
Or you can exchange cards between your two stations to make telling them apart more challenging.

That likely accounts for the performance increase after the PS change. A few days had elapsed when I updated the software.

Actually, the update was a last hope thing, as the ‘combo’ had stopped working, again, after the power supply change, but the temperature was only in the lower 20’s this time around.

I’ll wait and observe for a few more days. I was planning to activate a third ‘combo’ using the PCB antenna that was talked about recently. That would give me 3 ‘combos’ in the same area, my garage, to experiment and compare results.

The ‘problem’ is that my attention was ‘captured’ by another project that can also use the spare RPi. It’s about building a satellite ground station as part of the SatNOGS network ( https://satnogs.org/ ).

Lots of fun, so little time.:smile:

So the actual change to the setup is that it’s now probably finally working right after not working properly before because it was too hot?

Still very doubtful that the software made any difference at all.
It could still be that the kernel changes some USB stuff or whatever and your lack of power makes that change actually relevant to performance.

Anyway unless it puts out 5.1V it’s still not a solid power supply for the application.
That’s just the way it is with the crappy RPi power design for models <4

1 Like

I really don’t know. The previous failures had occured on very hot days, when the temperature in the garage was ‘north’ of 35 degrees Celsius. As ‘chronicled’ here: RadarBox FlightStick heat stroke? , I went through 3 SD cards, 2 RPis, 2 dongles, and 2 PSs of the same type. It continue to stop working when the temperature increased past 35 degrees. If I let it cool down, it would work again, until the hardware temperature increased.

After the third PS, but of a different brand, the ‘combo’ worked fine for a few days, but suddenly stopped when the garage temperature was only in the lower 20’s. No knowing what else to try, I issued the update/upgrade commands, instead of rebuilding/replacing the SD card.

Since then, the temperature in the garage has gone over 35 degress a few times, and so far it has been rock solid, just like the primary ‘combo’.

Perhaps a combination of factors, but one thing is for sure, the secondary ‘combo’ was never ahead of the primary for the almost 1 year they worked side by side.

Maybe you moved the antenna a few inches?

Comparing anything with antennas in a garage is gonna be highly random in my opinion.

1 Like

Both antennas are in the garage. The two ‘combos’ are 4 feet apart inside the garage.

Nothing has moved, but a possible explanation just crossed my mind.

The RPi is only 10 inches away from the PS. The dongle is connected directly to the RPi USB port. The new PS appears to be of a better quality. I’m now wondering if the other 2 PSs, of the same type, were also strong RFI ‘generators’ that reached the RPi and dongle, impacting it’s sensitivity.

It still does not explain the failure in low temperature, but I’ll ‘accept’ a coincidence there.:grinning:

Get yourself something beefy like this:
https://www.amazon.com/MEANWELL-Switching-Power-Supplies-LRS-50-5/dp/B018TDA64O

Then you can have a 5V power rail and adjust it to 5.15 V on the rail.
On the RPi 5V rail you’ll have around 5.00 to 5.05 V.

Oh gotta say i had some bugs crawl into that power supply though, that might be a downside :wink:

That is an option, but I have been having good luck with re-purposed, brand name discarded consumer electronics PSs. Only the so called RPi PSs seem to cause problems for me.

Well of those i’d only recommend the genuine “Official” ones.

https://www.amazon.ca/Official-Raspberrypi-Switching-Adapter-International/dp/B07GJH4DVN

But at that price …

Very much so. A clear case of “you get what you pay for”, but doing it on the cheap is part of the fun.:rofl:

Looks like the ones I have been re-purposing. Just add the adapter from barrel to micro USB from eBay, where else.:grinning:

By the way, none of the 3 RPis are model 4. No plans to buy another RPi until they come out with the eMMC version.

I know they are not, they wouldn’t be working with the dongles otherwise.
They still need to figure out their USB driver/firmware for the model 4.

But power seems to be rock solid compared to the previous iterations.