Regarding cooling: I had tried a Pi 3b+ in a plastic case that was also sealed in an enclosure . It was running hot and throttling due to solar gain. ( and posting anomalies to my stats page) I found that there was a lot of heated air trapped right in the Pi case itself despite the heat sinks etc.the air didn’t move and the heat just sat there. A small fan to circulate the air within the enclosure helped a great deal to cool the CPU temp.
I know that you are likely chasing voltage drop so that would consume even more power but it helped me by not having to make holes in the enclosure that would open the thing up to weather.
Looks like you have good advice on the POE gear- only thing I would add on power would be to also check your 110v supply to make sure line voltage is what it should be and that there’s nothing cycling on that circuit that could be causing a sag. ( big inductive load, motor startup etc. )
I just tried a 2.4amp POE unit (3m/10ft cable) on an RPI3 with two rtl-sdr dongles attached. (I wanted the Power unit for something else).
It kept rebooting. I had to put back the original power unit and order some new ones.
You have to get quality equipment or it just doesn’t work
I finally got around to messing with my Pi out on the roof. I didn’t quite know what I was doing measuring amps even though I thought I would so I never took note of what I was getting for amps, but after putting the splitter to 12V and then converting that to 5V (reads 5.17V) I have not gotten any messages about clock being unstable. MLAT hasn’t stopped showing up on the website either which is great. I ran the Pi at max 600MHz before today (about 4 days) and it actually worked pretty well with the previous setup where the splitter was 5V. I’m currently running the CPU at default clock speeds and it hasn’t had any issues yet as of 2 hours later. It doesn’t seem to be thermal throttled yet at 57C but we’ll see as the day goes on.
I don’t think my cable was any issue considering I almost doubled the length of it and didn’t replace any. Last time I was up there it started raining so I shoved everything in the enclosure real quick not knowing today I would find it pinched pretty good (oops ). No shorts though.
The weird thing is that I am constantly shown “throttled=0x50005” where before I got 0x50000 and a rare 0x50005 but it seems like it runs way better now. In dmesg at boot I got “Under-voltage detected! (0x00050005)” only once, but never got a voltage normalized message after it. I can’t complain though if it continues to work without issues though! So far setting the splitter to 12V and then converting it to 5V seems to work the best but I might have to play around still. I may even add ventilation to it today. Thanks everyone!
It’s not great if the Pi is constantly undervoltage. You’re running it outside specification - it might work sometimes, but it might also cause hard to diagnose errors. USB is usually the first thing to go…
The critical thing to look at is the voltage supplied to the Pi while it is under load. You’ll see a voltage drop as more current is drawn (especially if you have a long cable run involved), the important thing to look at is the voltage while the Pi is connected and running, not just the voltage when nothing is active.
There are cheap $5-10 usb voltage and current testers (check amazon or eBay) that will measure voltage under actual load. Voltages around 4.8V or less are bad with under voltage warning happening a bit below that. Around 4.6V there is a chance of not recovering from an under voltage event and these low power events will be logged on the latest PiAware software.
For reference the Raspberry pi 3B & RTL dongle uses about 900mA of power while running piaware.
The Raspberry PI 3B+ & RTL dongle uses about 1200mA of power.
RTL dongles use from 300mA to 350mA of power. The rest of the power is for the raspberry pi.
The main increase in power usage on the new raspberry pi 3B+ is the wifi chip and 1Gb/sec ethernet. It needs more power to run these faster data rate chips.
). No shorts though.