en-AU.vtt

WEBVTT
Kind: captions
Language: en

00:00:00.000 --> 00:00:01.000
English (AU) (Spoken) [Manually Transcribed Captions]
github.com/WizardTim/WizardTim-captions

00:00:01.000 --> 00:00:07.680
So in the last video we had a look at this 
here PIR sensor manufactured by HPM and we  

00:00:07.680 --> 00:00:12.300
had a close look at what's inside one of 
these as well as the failure mode it had  

00:00:12.300 --> 00:00:18.660
and then we had a closer look at this here 
metal can which is the pyroelectric sensor  

00:00:19.380 --> 00:00:25.380
since making that video I've been given this 
here PIR sensor this one is manufactured by  

00:00:25.380 --> 00:00:33.600
ROBUS and is a "PROTON UNIVERSAL 360° PIR" 
sensor designed for mounting on ceilings  

00:00:34.980 --> 00:00:41.220
and this particular unit has been given to me 
because it's broken so in this video we're going  

00:00:41.220 --> 00:00:48.300
to try and repair it but also we'll take a look 
at the actual PIR sensor as well because in the  

00:00:48.300 --> 00:00:53.880
previous video I mentioned that there was such a 
thing as a quad type pyroelectric sensor and well  

00:00:53.880 --> 00:00:59.940
judging by the shape of the detection pattern of 
this unit I suspect that we're going to see one  

00:00:59.940 --> 00:01:06.840
of these inside so let's take this thing apart 
taking the front bezel off there's no screws  

00:01:06.840 --> 00:01:12.660
underneath however there are three adjustment pots 
one for lux which is because this has day night  

00:01:12.660 --> 00:01:18.660
sensing number for sensitivity and another 
for time which is how long it spends on for  

00:01:20.340 --> 00:01:26.220
but on the back we can see there's four screws 
as well as this nice little connection block for  

00:01:26.220 --> 00:01:33.720
the wires and this is only a three pin device 
one for the active in, one for neutral and one  

00:01:33.720 --> 00:01:38.580
for the active out which is switched and of 
course there's no earth because there's no  

00:01:38.580 --> 00:01:46.140
exposed metal work on this device so taking those 
four screws out and we can see what's inside and  

00:01:46.140 --> 00:01:50.940
there's even more than the previous PIR sensor 
there's actually three of these rather large  

00:01:50.940 --> 00:01:58.800
16-pin SOIC devices plenty of capacitors probably 
for timing and bulk decoupling as well as a nice  

00:01:58.800 --> 00:02:06.060
fuse the output relay some X class suppression 
capacitors and of course the potentiometers but  

00:02:06.060 --> 00:02:13.320
taking a look at the others oh okay uh I was 
expecting a single quad type pyroelectric  

00:02:13.320 --> 00:02:19.980
sensor but actually they've gone and put three I 
would presume just normal dual type pyroelectric  

00:02:19.980 --> 00:02:28.560
sensors in at 120 degrees to get there 360 degree 
coverage but this will certainly work as well but  

00:02:28.560 --> 00:02:33.780
I guess we're not going to take a look at a quad 
type sensor today taking a look back at the datasheet

00:02:33.780 --> 00:02:39.540
now I now realize I was rather misled by 
this figure here we can see that at a ceiling  

00:02:39.540 --> 00:02:45.900
height of say three meters it's expected to have 
a spread of 18 meters I now realize that this  

00:02:45.900 --> 00:02:54.000
diagram is definitely not to scale here it is two 
scale so yeah it's definitely got a much larger 

00:02:54.000 --> 00:03:00.720
range than I thought you definitely can't achieve 
this sort of range without either some very fancy  

00:03:00.720 --> 00:03:07.380
optics or rather just shoving three of them on a 
PCB under the same optics which is certainly  

00:03:07.380 --> 00:03:11.220
going to get the job done but unfortunately 
means we're not going to be looking at a  

00:03:11.220 --> 00:03:18.300
quad type sensor today however this PIR sensor 
still needs to be repaired so let's see if we  

00:03:18.300 --> 00:03:25.500
can identify the failure mode because I was told 
this thing just behaves strangely and doesn't work  

00:03:25.500 --> 00:03:31.380
properly it's not that it doesn't work at all or 
that it's uh done weird things made a loud bang  

00:03:31.380 --> 00:03:38.340
sound it's just that it's not reliable so I've 
got some test probes on the AC in as well as  

00:03:38.340 --> 00:03:46.140
the neutral and that's connected to this plug and 
plugging it into 240 volts AC there's no explosion  

00:03:47.820 --> 00:03:54.300
I heard some clicking so it's doing 
something now let's get this flipped over  

00:03:54.300 --> 00:04:01.200
because there's a LED on the other side plugging 
it in again and yeah there's some weird clicking  

00:04:01.860 --> 00:04:06.660
uh and this red light's flashing let me 
dim the light so you can see that better  

00:04:07.680 --> 00:04:14.880
so the red LED is flashing quickly looking at the 
manual this appears to be the warm-up routine that  

00:04:14.880 --> 00:04:21.720
goes for a minute and after a minute that light 
stops flashing and if I bring my hand close to it  

00:04:21.720 --> 00:04:30.240
it does actually activate although it's it doesn't 
seem to actually switch properly and after it does  

00:04:30.240 --> 00:04:37.740
it once it doesn't seem to want to do it again so 
there's some sort of weird thing happening here  

00:04:37.740 --> 00:04:46.440
I'm wondering if maybe the relay...? is the relay 
closing drawing too much power and then browning out 

00:04:46.440 --> 00:04:52.020
the power supply that's the only thing 
I can imagine that would create this symptom  

00:04:52.980 --> 00:04:59.340
so I guess the first thing we're going to 
try is to check the power supply now this  

00:04:59.340 --> 00:05:06.540
thing uses a capacitive dropper power supply 
which is non-isolated it seems to use these two  

00:05:06.540 --> 00:05:13.380
capacitors here I'm not quite sure why there's 
two of them or actually it's probably just the  

00:05:13.380 --> 00:05:17.760
one of them's for the control electronics 
and the other one is for the relay coil  

00:05:18.360 --> 00:05:27.240
so testing the first capacitor we should get 100 
nanofarads and yea 32 is definitely not 100

00:05:28.380 --> 00:05:36.900
uh that's the right thing right? yeah that's the 
100 nanofarad capacitor yep so that uh... it could  

00:05:36.900 --> 00:05:41.880
be that it's in circuit but that doesn't 
seem right uh checking the other one this  

00:05:41.880 --> 00:05:53.340
one should be 470 nanofarads and it's 383 
yeah these capacitors have me a bit suspicious  

00:05:55.320 --> 00:06:02.220
so let me get these capacitors out and we can test 
them outside of the circuit so I've got here this  

00:06:02.220 --> 00:06:09.420
LCR meter which is better than the DMM and it's 
measuring 381 nanofarad so yeah it's definitely  

00:06:09.420 --> 00:06:16.380
not 470 there's something wrong with this one and 
the other one which should measure 100 nanofarads  

00:06:16.380 --> 00:06:24.000
is 31.8 so yeah both of these capacitors 
are just completely rooted especially this  

00:06:24.000 --> 00:06:30.240
100 nanofarad one which is actually about 32 
nanofarads this thing is going to cause all sorts  

00:06:30.240 --> 00:06:34.200
of havoc in the circuit it's just not going 
to supply enough current to run the circuit  

00:06:35.340 --> 00:06:41.160
so I reckon there's a pretty good indication that 
this is what the problem is and of course these  

00:06:41.160 --> 00:06:47.220
capacitive dropper power supplies do actually fail 
quite often you'll see a lot of YouTube videos  

00:06:47.220 --> 00:06:53.460
and Wiki articles and blog posts about people 
repairing these capacitive droppers because the  

00:06:53.460 --> 00:07:00.540
capacitor has lost its capacitance it's quite a 
common thing so I really wouldn't recommend one of  

00:07:00.540 --> 00:07:07.260
these capacitive dropper power suppliers they're 
just not very reliable although the advantage  

00:07:07.260 --> 00:07:12.960
that they have is that they're absolutely dirt 
cheap you really cannot beat the price point of a  

00:07:12.960 --> 00:07:18.420
capacitor rather than an entire switch mode power 
supply so that's why they get used all the time  

00:07:19.020 --> 00:07:24.720
but at the very least it makes for a very easy 
repair so I had a look around for some replacement  

00:07:24.720 --> 00:07:33.360
capacitors and I settled on the MKP339 X2 series 
from Vishay these are some quite nice capacitors  

00:07:33.360 --> 00:07:39.420
these are from a reputable brand and they're 
rated to higher voltages than the previous ones  

00:07:39.420 --> 00:07:48.120
the previous ones were 275 volt AC rated these 
ones are 310 volts AC rated so these should last  

00:07:48.120 --> 00:07:54.420
hopefully a lot longer than the previous ones and 
of course because these are used in a capacitive  

00:07:54.420 --> 00:08:01.440
dropper it's important to use the same values 
unlike bulk decoupling capacitors where you can  

00:08:01.440 --> 00:08:07.800
just increase the capacitance value and increase 
the voltage rating so long as they're at least a  

00:08:07.800 --> 00:08:13.560
certain number it's good enough in this case the 
capacitance value is very important because if  

00:08:13.560 --> 00:08:18.540
you make the capacitance value too large you'll 
pass too much current and you risk overheating  

00:08:18.540 --> 00:08:26.280
the shunt regulator inside so we're going to go 
for a 0.47 microfarad and 0.1 microfarad which  

00:08:26.280 --> 00:08:32.580
is exactly the same as was in here before and of 
course I chose one with the correct case size  

00:08:32.580 --> 00:08:39.000
and lead spacing so it would fit in the PCB and 
just to be safe I check them on the LCR meter when  

00:08:39.000 --> 00:08:43.680
I got them and they were indeed the correct 
value within the tolerances of manufacturing  

00:08:45.120 --> 00:08:49.260
so it's just a matter of cleaning up those 
plated through holes to get the components  

00:08:49.260 --> 00:08:54.960
in thankfully I've repaired my desoldering 
gun since the last video it's very important  

00:08:54.960 --> 00:08:58.860
you keep your desoldering gun nice and 
clean make sure that there's absolutely  

00:08:58.860 --> 00:09:02.460
nothing left in the tube otherwise it 
will clog and you'll have a bad day  

00:09:03.360 --> 00:09:07.740
and apart from that there's really nothing 
complicated about this repair it's just a matter  

00:09:07.740 --> 00:09:14.100
of soldering two components in nice and easy and 
now that I've got them in we can see that they fit  

00:09:14.100 --> 00:09:20.340
nicely and they are indeed the correct height so 
they won't interfere with the plastic enclosure

00:09:22.680 --> 00:09:26.100
and then just putting it all 
back together and plugging it in  

00:09:26.640 --> 00:09:29.820
and I'd let it go through its 
one minute warm-up procedure  

00:09:30.360 --> 00:09:35.940
and if we listen carefully we can hear 
it click when I put my hand over it

00:09:38.524 --> 00:09:39.333
*click*

00:09:40.075 --> 00:09:43.740
and I've spent several minutes 
waving my hands in front of this thing  

00:09:43.740 --> 00:09:49.260
and taking them away trying to catch this 
thing out make it fail again but I haven't  

00:09:49.260 --> 00:09:56.040
made it fail again so I think it's fixed now 
so that was a rather nice repair very easy to  

00:09:56.040 --> 00:10:02.940
fault find very obvious failure mode and very 
easy solution and these capacitors were only  

00:10:02.940 --> 00:10:08.400
about a dollar fifty each which is quite nice 
because these things are about sixty dollars new  

00:10:09.180 --> 00:10:14.520
so that's it for the repair but let's take a look 
at these capacitors a bit closer whilst I've got  

00:10:14.520 --> 00:10:21.660
them out and also because I was denied that quad 
type pyroelectric sensor so let's test these  

00:10:21.660 --> 00:10:28.740
capacitors to the extreme and uh... well actually 
let's just blow them up so I've got here  

00:10:28.740 --> 00:10:35.820
a 1.5 kilovolt peak to peak biphasic pulse this is 
a 360 joule pulse delivered over 15 milliseconds  

00:10:35.820 --> 00:10:43.020
actually let's put this into perspective 
here is a quarter watt 56 ohm resistor and

00:10:43.934 --> 00:10:44.434
*pop*

00:10:45.138 --> 00:10:51.780
yeah it just explodes yeah that's uh not a 
very happy resistor you can see here in this  

00:10:51.780 --> 00:10:57.900
microscope picture the resistive element has just 
been completely vaporized off the resistor core  

00:10:57.900 --> 00:11:07.620
but also let's have a look at a 400 volt [4.7] 
microfarad capacitor out of a USB charger and

00:11:08.534 --> 00:11:09.321
*pop*

00:11:09.962 --> 00:11:16.251
yeah that thing never stood a chance either 
so let's test our two other capacitors  

00:11:16.620 --> 00:11:21.720
this is exactly the same pulse so I'm 
hoping we get some nice fireworks although  

00:11:21.720 --> 00:11:26.220
I am hoping that the Vishay one stands 
up a little bit better and hoping that  

00:11:26.220 --> 00:11:32.260
the JIMSON brand one just catastrophically 
explodes everywhere and...

00:11:33.079 --> 00:11:33.768
*contactor thud*

00:11:34.037 --> 00:11:35.441
oh what?!

00:11:35.441 --> 00:11:39.813
oh what nah 
 uh let me do that again 

00:11:40.611 --> 00:11:41.111
*contactor thud*

00:11:41.111 --> 00:11:45.397
oh what?! who's bloody well sold me capacitors that don't explode! 
Come on!

00:11:46.560 --> 00:11:52.080
so as you're probably screaming at the screen 
right now these are of course X2 class safety  

00:11:52.080 --> 00:11:57.420
capacitors all of these funny markings on the 
side of them aren't just to make it look pretty  

00:11:57.420 --> 00:12:03.780
these are actually all sorts of compliance and 
regulatory markings that say that it complies  

00:12:03.780 --> 00:12:10.740
with all sorts of safety standards that say it can 
withstand certain sorts of abuse so in the case of  

00:12:10.740 --> 00:12:17.520
these two capacitors that are X2 class rated even 
this JIMSON I don't know I've never heard of them  

00:12:17.520 --> 00:12:24.000
before but even this one is rated to withstand 
1.2 kilovolts for a minute and this Vishay one  

00:12:24.000 --> 00:12:31.560
specifically says it withstands 24 pulses at 2.5 
kilovolts so unfortunately I don't have anything  

00:12:31.560 --> 00:12:37.440
that goes that high I did pulse this thing 15 
times and it didn't explode so yeah unfortunately  

00:12:37.440 --> 00:12:43.500
I can't make it explode but I hope that this at 
least demonstrated that these little unassuming  

00:12:43.500 --> 00:12:49.080
capacitors can actually put up a pretty good fight 
against what's on the other side of these outlets  

00:12:49.620 --> 00:12:55.920
the sort of stuff in our electrical distribution 
network is rather big and scary and can create  

00:12:55.920 --> 00:13:00.840
all sorts of weird transients that have the 
opportunity to blow things up if it weren't  

00:13:00.840 --> 00:13:06.780
for these little suppression capacitors although 
of course it's not just these capacitors that save  

00:13:06.780 --> 00:13:13.020
your gadget from being exploded there's of course 
also Y class suppression capacitors as well as  

00:13:13.020 --> 00:13:19.680
varistors and PTCs and NTCs and of course fuses 
and those are the components that will generally  

00:13:19.680 --> 00:13:27.000
be absorbing the large pulses these X class 
capacitors are more so for EMI suppression  

00:13:27.000 --> 00:13:33.300
when another device nearby emits some sort of 
interference that could interfere with your device  

00:13:34.560 --> 00:13:41.040
but that's all nice and everything but why 
is it that I just put 1.5 kilovolts through  

00:13:41.040 --> 00:13:46.680
this capacitor and it didn't explode in 
fact it didn't even lose any capacitance  

00:13:47.700 --> 00:13:55.740
uh if it wasn't line transients that made this 
100 nanofarad capacitor go to 32 nanofarads what  

00:13:55.740 --> 00:14:02.160
did because that's quite often what I see people 
will say that these things get overvolted and they  

00:14:02.160 --> 00:14:09.540
internally fuse and lose capacitance over time so 
let's cut one of these open and see if we can find  

00:14:09.540 --> 00:14:15.420
anything that's obviously wrong with it these 
capacitors inside uh just like any other normal  

00:14:15.420 --> 00:14:21.120
metallized polypropylene capacitor however they 
of course need to withstand much higher voltages  

00:14:21.120 --> 00:14:29.040
and energies so the layer thickness is higher but 
also they have much better potting in this epoxy  

00:14:29.040 --> 00:14:35.640
and plastic shell so I've got the capacitor out 
of the potting now and I've unwound a bit and we  

00:14:35.640 --> 00:14:44.280
can see uh... yea I don't know what I'm looking 
at actually um yeah there seems to be some  

00:14:44.280 --> 00:14:50.460
delamination going on here but I have no clue 
what that means however thankfully I have this  

00:14:50.460 --> 00:14:55.980
here failure diagram and this shows all sorts 
of different failure modes that can happen as  

00:14:55.980 --> 00:15:03.120
well as their causes and of course our value mode 
is the capacitance decrease one and that seems to  

00:15:03.120 --> 00:15:10.080
be caused by all sorts of different things uh but 
I'm wondering if uh maybe humidity has something  

00:15:10.080 --> 00:15:16.500
to do with it partly because looking back at 
the footage we can see that there was actually  

00:15:16.500 --> 00:15:22.320
this here void in the epoxy so I'm wondering if 
moisture was able to more easily penetrate this  

00:15:23.580 --> 00:15:30.360
but this speculation is mostly driven by a video 
I saw on the EEVblog channel in which Dave Jones  

00:15:30.360 --> 00:15:36.720
had an interview with Ron Demko from KYOCERA AVX 
in which he went through this document and had  

00:15:36.720 --> 00:15:42.000
a talk about all the sorts of different failure 
modes but specifically had a look at a capacitor  

00:15:42.000 --> 00:15:48.240
that Dave Jones had out of a heater which was a 
very similar capacitor and had some rather similar  

00:15:48.240 --> 00:15:54.660
sort of failure indicators as my capacitor so I'm 
wondering if moisture was the cause whether it was  

00:15:54.660 --> 00:16:01.920
from poor encapsulation or from the manufacturing 
process I have no clue but I do recommend watching  

00:16:01.920 --> 00:16:06.780
this video if you want to know more Ron goes 
into quite a lot of detail about these safety  

00:16:06.780 --> 00:16:12.420
capacitors and their failure modes as well as 
a number of other different capacitor types so  

00:16:12.420 --> 00:16:18.360
I'll leave a link to that in the description and 
that's it for this video so thanks for watching