Philips Hue Go Light Repair Attempt

[KYBOSH]

I got my hands on a Philips Hue Go lamp that was not working properly.

Model# 71460

The symptom was slow flashing purple LED (error code?) but it would not stay lit nor are you able to connect to it via BT or Zigbee.

It did not come with a charger but when I probed the power pads (12V) I was getting a slight spark so I knew there was a short somewhere.

I was able to find a hotspot on an IC but I am having trouble figuring out what it is as there is absolutely no mention of the markings anywhere on the web.

Attached are some photos of the PCB and the IC in question.
The IC has the marking "BGW7NB". I have circled the spot where the short seems to be located (based on where it gets the hottest).

Now i am not actually sure if this hot spot is a short. It was just the hottest part on the board. I have no idea if it is supposed to run hot bcos I am not sure what it is. It maybe a proprietary power IC that has other functions integrated in it which makes it run warm even the light is off.



Any help, tips or ideas would be greatly appreciated.

[stj]

Re: Philips Hue Go Light Repair Attempt

thats a buck/boost convertor - check all the ceramic caps for shorts before blaming it

[KYBOSH]

Re: Philips Hue Go Light Repair Attempt

thats a buck/boost convertor - check all the ceramic caps for shorts before blaming it

Wow... you know i was thinking of testing the components around it again today. I figured that this IC would heat up bcos it would have the thinnest connection but that doesn't necessarily mean its the source of the short.

This reply sort of validates that I was on the right path of thinking.

Thank you very much for that!

I will read up on Buck–boost converters (first time hearing of such a device).
I'll check my caps tonight.

[stj]

Re: Philips Hue Go Light Repair Attempt

it's a buck or a boost - wont be both.
they operate at high frequencies and have a habit of killing ceramic caps

[KYBOSH]

Re: Philips Hue Go Light Repair Attempt

Well I was up last night and again this morning probing and looking about.
It was long and tedious as I was looking for shorts but it turns out that (with the exception of possibly the BGW7NB IC) I couldnt find a single one.

And while I didnt find any shorts but i found a few caps where were reading OPEN.

Initially I used Diode Mode on my multimeter to check the voltage drops of each cap in circuit. Most caps had a drop of 0.3-0.5v. Any cap that was an outlier was pulled out of circuit and checked for capacitance and resistance. For instance a few measured ~1.3v while a few didn't register any drop at all.

I found 2 totally open caps and 1 cap on its way out (only had a resistance up to 46kΩ). Some caps with outlier voltage drops actually looked okay when i tested them out of circuit though I cannot know if their capacitance values are within spec as I have no way of knowing for certain what their capacitance was supposed to be. The best I can do is look for similarly sized caps nearby to get a hint of what they might be. Thats my plan for replacing the dead ones.
For example: C2 is dead but C3 direct across the card measures 114nF.

I have checked the resistors and they seem fine.
I have checked a few of the transistors near the bulk of the caps and they also seem okay.

Most of the caps I have lifted are ~100, ~500 and ~1k nF in value.
In order to move forward I need to order replacements.

[KYBOSH]

Re: Philips Hue Go Light Repair Attempt

So i spent the late evening going thru and replacing any and all caps they looked funny to me. In all i replaced 4 of them but no joy. After replacing C2 it read as shorted. I removed it and tested the pads and they both read shorted as well. I probed all around and couldnt find anything else that could be causing this. All the resistors were good and within specifcations. The transistors nearby were okay too. I am leaning towards the theory that the short/fault is within that big Hue Engine chip. The flashing purple light is most likely an error signal that the chip's hardware or firmware is faulty. If there was an issue on the rails then the device would/should probably not startup in the dirst place as any built-in protections would have prevented a stsrtup.

No way tor me to know wothout more intel. I was also not able to get even the flashing light to come on after swapping out the bad caps. No signs of life and my watt meter shows it was only drawing 0.05 watts of power when cinnected to the mains. I willl look over my soldering job la little later with fresh eyes.