RPi 3 boots and runs fine but red power led won't light

[Rhassinger]

Can anyone give me some insight into what is happening with the RPi status LED's? It almost seems like the status LED's are a weak point in the design as it makes no sense that they are so susceptible to damage while the rest of the RPi runs properly. I have three RPi 3's, an RPi 2 B+, an RPi B+, numerous power supplies for the RPi's, many different sized Sandisk micro SD cards and have been experimenting with the Rpi's for over a year.

One RPi 3 is brand new and used with a brand new official RPi 2.5 amp output wall wort power supply, and the RPi was initially running with proper status LED's. While attempting to get an Adafruit 2.8 inch GPIO-mounted Pitft display running for the last couple weeks I eventually noticed that the red power LED was no longer working. Completely dead. The green activity LED worked properly. I removed the Adafruit display from the GPIO connector resulting in a bare, stock RPi 3 and found that the RPi 3 runs just fine, just no red power led.

The RPi Model B+ runs just fine but has no status LED's at all, either red power LED or green activity LED.

That's the symptom. The RPi's run just fine but the red power LED on the RPi 3 and both red and green status LED's on the RPi Model B+ are dead. Any ideas?

[Burngate]

The circuit driving the red power LED on the B+, 2B and 3B (there's no 2B+) is quite simple - a APX803-46SAG chip to detect the 5v rail voltage, fet, led and a few resistors.
One dying could be called unfortunate; more than one invites suspicion. Does the screen show the lightning bolt symbol?

One possibility is that your power supply can't provide enough current at 5v and its output is dropping too low; another is that you have half-blown the polyfuse.

The green ACT LED is driven directly from a GPIO on the B+ and 2B; on the 3B there's another chip in the way.
You may have done something odd with the GPIO setup?

[Rhassinger]

Thank you for the reply. The Pi is brand new, as was the official 2.5 amp plug in power transformer. It was housed in a C4 Zebra Case with the top plate GPIO cut out enlarged to accept the extra GPIO pins on the Adafruit 2.8 inch Pitft display. The capacitive touch display, Product ID 2324, is also brand new and plugs onto the Pi's GPIO pins, so the Pi is well protected and also operated properly initially with Raspian Jessie. While trying to get it running with Raspbian Stretch, which I'm still having problems with, is when I noticed the red power LED is out. I've tried numerous micro SD cards and a new 3 amp plug in power transformer. It's just frustrating that the power LED is dead but the Pi runs fine, especially in light of also having another Pi B+ with both LED's dead but it also runs fine. Is the poly fuse changeable at all? Thanks again for taking the time to reply.

[Rhassinger]

Just a thought.....since the Pitft display is very new, is it possible that display is drawing a lot of current and pushing the demand up near the polyfuse limit? I still don't understand how that would result in the symptoms I'm seeing: the red power LED is now dead but the Pi still runs fine without the display plugged into the GPIO. That is with the caveat that I still can't get the display to work with Raspbian Stretch.

[Rhassinger]

* OK, I thought I would document a couple most interesting symptoms. Using the Pi 3 in a C4 Labs Zebra case so it is protected, I tested the voltage between GPIO pins 2 and 6 while using the official Raspberry Pi 2.5 amp power transformer. Voltage was 5.15 - 5.2 volts measured with an analog Micronta (from Radio Shack) VOM. Pin 1 has 3.3 volts. No red power LED. THEN I switched to a non-Raspberry Pi 3 amp power transformer, and after about a minute the red power LED came on! Voltage was 5.30 - 5.35 volts and 3.3 volts. I switched back to the official Raspberry Pi 2.5 amp power transformer and the red power LED again worked normally. Now the red power LED works normally with either power transformer.
* I think I found a problem with the C4 Zebra case as well. When you assemble the case if you push the Pi toward the micro SD end of the case the bottom 'L' shaped clear plastic separator seems to catch on the top edge of the Pi's status LED's. I'm sure it puts a lot of pressure on the edge of the LED's. When I loosened the screws on the case for the Model B+ the power LED started working. I got one little flash out of the green activity LED by poking it with a plastic rod but it is still dead. I used a Dremel tool to grind a 45 degree notch over the LED's in that bottom 'L' shaped separator so it doesn't contact the LED's. I think I'll let C4 Labs know about this.

Separator touching LEDs

IMG_0003.JPG (39.1 KiB) Viewed 1393 times

Notched separator over LEDs

IMG_0002.JPG (38.6 KiB) Viewed 1393 times

[Rhassinger]

Screwed up last post. IMG_0002.JPG shows separator touching LED edges, IMG_0003.JPG shows the separator with the notched area over the LEDs. Sorry.

[Rhassinger]

The RPi 3 has close to the same problem with the C4 Labs Zebra cases as far as the bottom 'L' shaped separator being what looks like up against the two LED's, as well as up against the on-board blue tooth wifi chip. The tolerances of the bottom separator that holds the RPi 3 in place seems to be tighter as the RPi3 is locked in place. While I couldn't get the RPi 3 to slide around at all, it appears that the 'L' shaped separator is right up against the LED's and blue tooth/wifi chip. If a user really torques down the four mounting screws there is a possibility of putting force on those three components. I cut a slanted notch in the bottom separator which keeps that from happening. See the images.

[Burngate]

You seem to have found the problem!
My original thought, that the circuit is too simple to be problematical, is obviously too simplistic. Physical damage didn't occur to me.
I haven't put my Pis in cases for a couple of reasons - one: I'm too cheapskate to buy a case, and two: they'd never stay in it long enough to be worth while, since I'm always moving things about.

Curing it would seem to be down to reworking the soldering of the LEDs and the tracks to them.
Long-term, cutting away the plastic of the case, before putting the Pi in it, seems to be the cure.
I wonder what the tolerances the manufacturers of the cases work to.

[Rhassinger]

I just heard back from C4 Labs who makes the Zebra case for the Pi. Their Customer Service said the problem was reported before and the case design changed. I have five of the cases as I'm pretty anal about protecting electronics, and I also really like the design.

They are sending the modified parts so I can get my cases up to spec. I just sent an email to dustin@c4labs.net when I reported the problem and they got back to me.

[blackcorvo]

I had the same issue with the LED not turning on, but in my case it seems like Q1 has simply fallen off from the board. Not sure if I knocked it out myself while messing with mounting the board in an old project or what, but since I just wanted a power indicator, I put a jumper between the pads for the gate and drain of Q1. It works now.

Thought it might be a good head's up for folks who have this issue. Check if the transistor is in place.