#HA Yellow 1.3 Dead after plugging in a 5V power (correct polarity)

36 messages ยท Page 1 of 1 (latest)

vivid shoal
#

Hey everyone... I bricked my HAYellow... it is my fault but I need to fix it... Im able to surface mount solder (and I think I probably just need to replace a diode)... but Im seeing differences between the schematic (https://github.com/NabuCasa/yellow/releases/download/v1.3c/Yellow_v1.3c_Schematic_git.8957b440adda.pdf) and what my PCB actually looks like.... they are 'GENERALLY' correct, but my PCB is not stuffed with several components.... specifically with Q1, D3, C52, R26 and some of the circuitry around there seems like the PCB must be different then.... my hunch is that I have an open D2... here is a pic of my PCB.... https://photos.app.goo.gl/fZTuHWmDJHXm2FaZ6.

thank you!

scarlet schooner
#

Have you tried to power it with a 12V power supply, yet?

vivid shoal
#

Yes, I did.... It is basically just dead, no power.... Im actually thinking that maybe D2 is just open since I don't have Q1

#

I just pulled the diode now... hoping D2 is OPEN... ๐Ÿ™‚

#

Okay, so D2 appears good actually.

#

When I plug in the 12V brick, and I measure across the barrel jack, I get 0V.. so something on the HAYellow is shorting the supply... this is where it started after me accidentally putting my 5V brick (correct polarity) onto the HAYellow. Normally a lower voltage should be okay, but I could see the current draw being higher and popping something on the HAYellow trying to compenstate.

#

I have another 12V brick, I used that one to eliminate the original 12V supply being the issue. Both 12V bricks (3A each)... both pull down to 0V when plugged into the HAYellow.

#

Im going to plug in the HAYellow now with D2 out of circuit because that should get me at least showing a solid 12V on the hot side of D2... if for some reason I don't have 12V there, then I have an issue up front in the filtering components, which doesnt' make a whole lot of sense but hey, beggars can't be choosers.

#

D2 and Q1 are not on PCB currently... still no 12V on DC_IN_F tho....

#

Okay found DC_IN_F on the schematic... just turns on the LED on the optoisolator as a power good signal it seems.... that would not be pulling my 12V to 0....

#

If I put a voltmeter across my barrel jack with no power attached, I do read 3.3K ohms (which is R44) so that indicates no short... weird.

#

Time to double check some things... Im moving sorta fast. ๐Ÿ™‚

scarlet schooner
#

You have the PoE variant, right?

vivid shoal
#

yessir

scarlet schooner
#

Do you have a PoE power source to test, if that is working?

vivid shoal
#

Jorg, slight tangent before I test your question... I learned 2 important things here in retracing my steps...

#

so, a bit of a reset now... after using a proper GROUND, I DO have 12V on the ANODE (left) side of D2

#

Im going to put D2 back in and see what I have on the CATHODE side (right/output)

#

With D2 back in.... I have a short on the cathod side of D2...

#

Trying with PoE.... seems that doesn't work either which isn't surprising if the 12V supply is shorted somewhere.

#

I suspect that it is either U16 (+5V regulator) or U17 (+3.3V regulator) and between those two, more than likely it is the +5V regulator since it would have been cranking the hardest to try and maintain a +5V output with only +5V going in.

#

Luckily my power brick is switched, so i can see it 'pulse' the input supply and then overload current protect itself. I should in theory maybe see this pulse on the output of U16 or U17 if that component is good.

#

The other thing I could do but will wait until maybe a last resort, is to put the HAYellow on a linear 12V supply (or high current) and the bad component should become obvious quite quickly, but the chances of that 12V getting into other components may be high... maybe I will try this back on a high current 5V supply and see what happens, that shouldn't damage anything that is shorted but may release the magic smoke.

#

Oddly, U16 (+5V) is outputting a voltage pulse on pin 6, but U17 (+3.3V) is NOT outputting anything on the output pin 6 resembling a pulse. so I may assume that the 3.3V regulator is the one pulling the current down... Im going to lift pin 6 on each regulator and see what happens.

vivid shoal
#

so I just pulled both U16 and U17.... strange... still overloading the 12V power supply. the only thing I see that is also connected to the +12V supply is the PoE output, so that is a bit odd if that can be pulling it down... need to look over things a bit more.

#

PS - I have taken out the compute module at this point.

vivid shoal
#

Okay.... I figured this out.... here are some more pics... it is running...

#

Notice that I pulled out the main Tantalum Capacitor which I think is C95.

#

After looking closely... it was quite discolored on one end and I hadn't noticed it before...

#

Once I pulled it, my 12V on the anode of D2 was a solid 12V... I resoldered in U16 and U17 and it booted right up (without C95).

#

So... at this point, I am actually doubting that me plugging in the +5V supply caused the issue. My HA had been running for the last two months. Yesterday I unplugged it to put it into a new location... once in location, I put it on the +5V supply and noticed that it wasn't booting... within about 30 seconds I unplugged it and noticed that I had the switch power supply into it (same polarity).... so I put in the correct 12V supply. My thought is that something happened to this cap maybe even while the HA was running originally and it had failed. So, curious if there is a known issue with these caps? Im pretty sure this is a tant and I know they have issues with 'fingering' and shorting, but this HA is only about 9 months old.

#

Either way, there is a long story for someone to read late at night. Hope it helps in general troubleshooting. ๐Ÿ™‚

scarlet schooner
#

My Yellow is running since the first deliveries of the crowdfunded PoE ones. I only had a defective NVMe, which wasn't the Yellow's fault.

#

This is the first time I read about that kind of defect here on Discord. And there has been no hardware recall, yet.