en-AU.vtt

WEBVTT
Kind: captions
Language: en

00:00:00.000 --> 00:00:01.018
English (AU) (Spoken) [Manually Transcribed Captions]
github.com/WizardTim/WizardTim-captions

00:00:01.018 --> 00:00:07.440
This here is a 5.5 watt OSRAM LED 
bulb, it's a bayonet style bulb  

00:00:07.440 --> 00:00:12.880
and it's been in service for a couple of years 
now it's been on an automated timer that goes on  

00:00:12.880 --> 00:00:20.320
every night for a couple hours and unfortunately 
it no longer works so I thought I'd try and fix it  

00:00:20.320 --> 00:00:30.206
so as is usual with LED bulbs the diffuser is made 
of plastic so I think I can just lever it off like this.

00:00:30.790 --> 00:00:37.360
And after doing that for a while it came 
loose like this and we can see the diffuser is  

00:00:37.360 --> 00:00:43.920
adhesived in with what looks like silicone on 
these tabs here and the diffuser is indeed made  

00:00:43.920 --> 00:00:50.560
of plastic because it's all bendy now looking 
at this aluminium core PCB we can see there's  

00:00:50.560 --> 00:00:56.480
eight LEDs on board and if we look closely at them 
we can see that there's three pieces of silicon  

00:00:56.480 --> 00:01:03.520
in each chip so these are probably nine and a bit 
forward voltage LEDs which allows them to use so  

00:01:03.520 --> 00:01:11.920
few LED packages in probably what operates at 
a reasonably high voltage and we can see that  

00:01:11.920 --> 00:01:18.480
these LEDs are connected to rather large copper 
areas on the PCB which presumably helps with  

00:01:19.200 --> 00:01:24.880
heat dissipation although interestingly they 
don't use the same side of the LED package it's  

00:01:24.880 --> 00:01:30.480
a so i'd imagine some of these would fail before 
the others but let's test them to see whether or  

00:01:30.480 --> 00:01:39.840
not they're good they don't have the black spot 
of death on them so i think they might be good  

00:01:40.560 --> 00:01:47.040
and testing this first one it's probably 
around the wrong way this one's good yep  

00:01:48.640 --> 00:01:56.960
nice warm white this one's 
good this one's good good good

00:02:01.200 --> 00:02:07.840
and round the other way this one's good this 
one's good and this one's good so they're all good  

00:02:08.640 --> 00:02:14.720
so it's probably a problem with the power supply 
so let me get this silicone out of the way

00:02:18.480 --> 00:02:24.240
and i'm not quite sure how to get this 
aluminium board out it's uh it's really  

00:02:24.240 --> 00:02:31.600
wedged in there they seem to have press fit 
it into the aluminium uh case of the LED thing  

00:02:32.800 --> 00:02:38.000
yeah so this thing was pretty difficult to get 
out so i fast forwarded it here and i ended up  

00:02:38.000 --> 00:02:44.560
having to just stuff a metal spudger in the side 
and lever it out which actually broke the outer  

00:02:44.560 --> 00:02:50.480
plastic bit but i was just able to super glue 
that back together and of course the bayonet style  

00:02:51.040 --> 00:02:55.920
contacts on the bottom we can just lever these 
out they just press fit sometimes they're soldered  

00:02:57.520 --> 00:03:02.000
so now that we've got the power supply out the 
first thing that we can see is that it's rather  

00:03:02.000 --> 00:03:08.960
discolored so this thing certainly gets very hot 
during operation which is no surprise seeing as  

00:03:08.960 --> 00:03:16.800
it's a dirt cheap product but there's no obvious 
skid marks nothing's catastrophically exploded  

00:03:17.680 --> 00:03:25.120
everything looks reasonably good no broken 
solar joints that i can immediately see

00:03:29.440 --> 00:03:31.840
on the other side there's 
some pretty basic components  

00:03:32.960 --> 00:03:36.960
here we see the capacitor is actually 125 degrees  

00:03:36.960 --> 00:03:45.840
rated so that's good no bulging the capacitor 
is an AiSHi brand which is pretty good actually

00:03:49.360 --> 00:03:55.920
this inductor is marked 4.0 brackets 
0.6 so i imagine that's probably four  

00:03:55.920 --> 00:04:03.280
micro-henries at 600 milli-amps that 
seems a bit much but it's probably correct  

00:04:04.880 --> 00:04:09.360
so i think the very first thing to do 
will be to test this here fusible resistor  

00:04:10.560 --> 00:04:18.480
just to make sure it has continuity and testing 
it yep it has continuity it's a couple of ohms  

00:04:20.320 --> 00:04:27.840
so i'd say this is good and that means something 
hasn't catastrophically failed short circuit  

00:04:28.800 --> 00:04:34.880
now testing this diode bridge we expect to see 
some diode drops in some directions and in other  

00:04:34.880 --> 00:04:43.760
directions not so testing that there's a diode 
drop and a diode drop and the other way around  

00:04:45.360 --> 00:04:51.760
from the negative side dia drop yep now the 
next thing in the circuit before that large bulk  

00:04:51.760 --> 00:04:57.280
400 volt capacitor is this here inductor which 
probably has something to do with the power factor  

00:04:57.840 --> 00:05:04.000
and measuring that for continuity pretty sure i've 
got a good contact here but it's it's measuring  

00:05:04.000 --> 00:05:12.240
like millions of ohms um yeah i'm definitely 
making good contact so i think this inductor might  

00:05:12.240 --> 00:05:17.840
be bad now looking at the inductor we can actually 
see that there's some cracking on the top here  

00:05:18.880 --> 00:05:25.920
so i'd imagine this thing has been thermal cycled 
a few too many times and something inside of it  

00:05:25.920 --> 00:05:34.880
has gone wrong or maybe it's just underrated and 
something has melted inside and we can also see on  

00:05:34.880 --> 00:05:41.360
the bottom here there's some more cracking so yeah 
this inductor isn't happy so i'm just  

00:05:41.360 --> 00:05:48.080
going to remove that now the color code on this 
inductor says it's a three milli-henry inductor  

00:05:48.080 --> 00:05:55.440
that's 10% rated i had a look through my junk box 
and i couldn't find a similar thing although i  

00:05:55.440 --> 00:06:01.040
had this larger 10 milli-henry inductor and it's 
physically larger as well so i reckon that'll do  

00:06:01.600 --> 00:06:06.240
i don't really know what this thing's rated for 
in terms of current but i imagine it's probably  

00:06:06.240 --> 00:06:12.080
good enough for this application and just 
putting that back in place soldering it together  

00:06:12.080 --> 00:06:19.840
and then putting this thing 
back together not too difficult

00:06:24.080 --> 00:06:29.200
and then i put it back in the fixture and turned 
it on and we can see it works perfectly now  

00:06:29.200 --> 00:06:37.520
so that inductor was definitely faulty and i also 
had a look at it with the thermal camera and it  

00:06:37.520 --> 00:06:44.800
didn't look too bad i mean it gets pretty hot but 
i imagine it's designed like that so i just put  

00:06:44.800 --> 00:06:53.200
the diffuser back on and it should be good to go 
now this inductor here that's failed i'm actually  

00:06:53.200 --> 00:07:00.560
quite interested to see how it's failed i'm not 
sure if some of the windings have overheated and  

00:07:00.560 --> 00:07:08.480
fused together or something and it's i don't know 
exploded inside maybe or if it's just a thermal  

00:07:08.480 --> 00:07:16.640
cycle thing and some of the wires have broken or 
something so i think it's a good opportunity to  

00:07:16.640 --> 00:07:24.640
do some more advanced failure analysis on this 
so i'm going to prepare some two-part epoxy here  

00:07:24.640 --> 00:07:29.920
and i'm just going to mix that together mix it 
mix and mix it and then of course once i think  

00:07:29.920 --> 00:07:36.320
i've mixed it i mix it some more otherwise it 
turns out rubbish and then putting the inductor  

00:07:36.320 --> 00:07:43.200
in and mixing it around to make sure all of 
these surfaces are wetted and then i put that  

00:07:43.200 --> 00:07:51.120
on a little hot plate i made for a couple hours 
and here i'm just going to crank it up to max

00:07:55.760 --> 00:08:02.320
all right it's been about three 
hours and i've let this thing 

00:08:02.320 --> 00:08:09.280
bake for a while as well as cool down and in 
hindsight i think i roasted it a little bit  

00:08:09.280 --> 00:08:14.080
too much as it's formed quite a few bubbles and 
has discolored a little bit however that doesn't  

00:08:14.080 --> 00:08:18.240
really matter in this application we can also 
see actually that there's some bubbles that have  

00:08:18.240 --> 00:08:24.240
formed around the cracks so yeah i wonder if 
some of the epoxy is gone in there and some of  

00:08:24.240 --> 00:08:28.800
the air is escaped and that's where some of these 
bubbles are coming from as well now to start off  

00:08:28.800 --> 00:08:34.880
i'm just going to use some 220 grit sandpaper 
and i'm just going to go to town on it and i'm  

00:08:34.880 --> 00:08:39.840
going to get through the first layer and here we 
can start to see some of the construction of it

00:08:44.560 --> 00:08:53.840
looks to be a whole heap of fine 
wire wound around a ferrite core

00:09:27.680 --> 00:09:33.040
now that we've got more of it off we can 
start to see some more detail in how this  

00:09:33.040 --> 00:09:49.840
thing's manufactured can't see any obvious 
melted sections it all looks pretty good

00:09:56.480 --> 00:10:01.840
now i'm going to grind it down to the very center 
so that we can have a look at how this thing has  

00:10:01.840 --> 00:10:18.240
its leads constructed because i'm not actually 
sure how they put the leads in this thing  

00:10:18.240 --> 00:10:23.600
and now i'm at about the center so i'm going 
to go to 2000 grit paper and wet sand it

00:10:27.840 --> 00:10:31.840
and then go to 3000 and sort of polish it  

00:10:35.280 --> 00:10:42.320
so now it's nice and polished we can take a look 
at it and oh yeah so this is how they do it so it  

00:10:42.320 --> 00:10:48.160
looks like they wrapped the wires around the lead 
here and solder it together and then they sort of  

00:10:48.160 --> 00:10:54.480
glue the uh lead into the end and the end of it 
has been formed and we can see the interface here  

00:10:54.480 --> 00:11:04.560
between the ferrite and the lead and we can also 
see right here where that crack is and we can see  

00:11:04.560 --> 00:11:10.800
on the outside of the device as well where that 
crack was so it looks like that crack has formed  

00:11:10.800 --> 00:11:20.400
around where the wire comes off the lead and goes 
around to the actual windings so i suspect one of  

00:11:20.400 --> 00:11:27.280
these cracks has broken that wire and that's 
why it's gone open circuit so that pretty  

00:11:27.280 --> 00:11:32.560
much solves the mystery as to why this thing's 
broken unfortunately i didn't get a smoking gun  

00:11:32.560 --> 00:11:38.880
picture of exactly where it's cracked but i think 
that shows pretty much where it's happened so i'll  

00:11:38.880 --> 00:11:55.840
just leave you with some nice pictures of that 
cross section so as always thanks for watching