Advertisement

Blog

Design Challenge 2: Resistor Burning Out in Motion Sensor Night Light

OK, my electronics expert readers, here is another brain-teaser for us to discuss and help this Australian gentleman who contacted me recently. I have not yet researched this product, but I think the design concepts and conditions might be interesting to examine.

Here is the email and details from our reader:

Dear Steve,

I chanced upon your website and wondered if anyone there could take a minute to ponder an electronic circuitry problem? I have several combination “Click” brand Night Light Motion Activated Flashlights* in my home (see photos). The problem is they have all failed from overheating after about a year, and I wondered if anyone there could tell me why. I'm not an electronics minded person though I have a basic understanding of components. The component that always fails (melts and disappears actually) is a 2.4k resistor.

Circuit board

Replacing with the same kind of resistor hasn't helped. I assume other components have failed, too, notably the polyester capacitor above it. Input is 240V AC 50Hz (Australian mains). Whilst the devices are only $20 to replace, it's a problem I'd like to solve.

Device in the charger

Device out of the charger

*If you aren't familiar with the device, it's incredibly handy. It plugs into mains power and:

  • Charges the flashlight
  • Provides motion sensor illumination in the dark
  • Illuminates when mains power fails
  • Illuminates when flashlight removed from cradle.

If you have any insights I'd be most grateful.

Editor : Let’s dig into this issue and help our Australian reader with this problem and maybe come up with a fault condition and maybe a better design.

31 comments on “Design Challenge 2: Resistor Burning Out in Motion Sensor Night Light

  1. Victor Lorenzo
    May 9, 2014

    It is a little bit hard to say from the picture but one more bottom side picture would help.

    The TO-92 device… what is it? One thyristor? One UJT? In the latter case the circuit could be a full bridge rectifier, using series capacitors for acting as dividers and reducing the output voltage and followed by a UJT oscillator for driving the charger coil. Is that it? In that case the UJT could be the first failing part.

     

  2. Davidled
    May 12, 2014

    Based on PCB board, package type of resistor is like through hole. It seems like charging coil draw a lot of current. Also, it is noticed that there is brown color surrounding AC2.

  3. chirshadblog
    May 13, 2014

    @Daej: Its for protection isn't it ? 

  4. samicksha
    May 13, 2014

    May be that s because there is no standard pinout for the TO-92, different manufacturers use their own configuration of pinout.

  5. Davidled
    May 13, 2014

    If I understand correctly for your concern, through-hole type gives a more heating immunity than SMT package.  In High voltage circuit application such as 240V@ 50Hz, protection component could be designed such as Diode.

  6. eafpres
    May 14, 2014

    I wonder if the resistor is actually used as the safety fuse in the event of overvoltage.  Then I wonder if something else in the circuit, which should be, say, switching on and off to limit the charging duty cycle, isn't working, leading to the failure?

    I did find this where the writer said it was the fuse that had “blown”.

  7. yaweh_godalla
    May 15, 2014

    Thanks for taking an interest in my post. You could be on the right track suggesting the resistor is for overload protection, but I'm not well enough versed in electronics to be know how to check it. Since I posted the request, I took my last remaining working charger and tested the voltage across the coil. It turned out to be full mains voltage (230v in Australia). Armed with this [a-little-knowledge-is-a-dangerous-thing] bit of information, I took a burned-out unit from the junk box and applied full mains voltage directly across the coil. A spark and smoke and a garage full of burnt plastic smell proved that wasn't so smart. It also tripped the 10 amp circuit breaker in the mains switchboard. But it made me wonder if the circuitry is there to somehow reduce the current available to go through the coil. Is that even possible?

  8. yaweh_godalla
    May 15, 2014

    I don't know what TO-92 refers to. I can certainly provide more photos. What specifically would help. I have an old unit which I can disassemble and photograph. Incidentally, on the last remaining working unit I have now drilled a pattern of ventilation holes to hopefully reduce the heating problem. They do get pretty warm so I figured a little air flow can't hurt. I also sprayed it with insecticide to deter bugs snuggling up inside.

    I must say I'm pretty impressed at the interest everyone is showing. This is the first time I've participated in a forum for anything.

  9. Victor Lorenzo
    May 15, 2014

    @yaweh_godalla >> I don't know what TO-92 refers to .

    TO-92 refers to the component's encapsulation, the one that looks like the image below. Could you please send us a picture from its plane side? It is for seen its marking and try to figure out what is it.

  10. damngoodengineer
    May 15, 2014

    I would suspect the electrolytic cap.  The top looks a little bulged.  It will take most of the stress and probably they bought the cheapest part instead of the most robust.  Buy a new product and open it up and replace this cap with same value and voltage but 105 degree rating.  Look for high MTBF spec if you can find one.  Be sure to install with correct +/- orientation or it will fail on you in seconds.  Once this part goes bad all others are compromised and likely the thyristor goes next.

  11. Mikenyob
    May 15, 2014

    I would think given it's location that it is a Voltage regulator.

    Example: 7800 series TO-92

    I agree with damngoodengineer about the cap.

    (swelled=not good=leak/short=burned current limiting resistor)

  12. Victor Lorenzo
    May 15, 2014

    @Mikenyob >> I would think given it's location that it is a Voltage regulator .

    One thing to take into account, the charging method is inductive coupling and that requires powering the charger coil with AC, not DC.

    Anothers guess… Would it be one 2N3001 Thyristor? It is possible to create an oscillator with thyristors too. I make this guess taking into account the silkscreen markings.

     

  13. Mikenyob
    May 15, 2014

    Victor,

    I believe you are correct. I did not recognize that the PCB was in the charging side of the device. I thought it was a DC supply at the front end of a battery charging circuit on the flashlight side.

  14. yaweh_godalla
    May 15, 2014

    Thanks for your suggestion. I wish it was the cap because I vaguely understand caps and have replaced them from time to time in other devices. But the electrolytic capacitor it isn't bulged. It's flat as a pancake. There's no sign of electrolyte leakage either. The caps are always the first thing I look at, having repaired a few flat screen monitors (which always seem to blow the caps). On all the blown units the caps are flat and like new. All the ceramic caps look fine too, except of course the one closest to the burnt resistor. The heat sometimes melts the encapsulation on that cap too.

  15. yaweh_godalla
    May 15, 2014

    Thanks for the clarification. The TO-92 reads “13001 S 6B”.

    I have a clear picture but I can't work out how to attach images from my hard drive. The “Insert/edit image” icon below wants a URL rather than a file pathway.

     

  16. chirshadblog
    May 15, 2014

    @yaweh: Didn't you try to create a folder in the root and copy the images to it ? 

  17. chirshadblog
    May 15, 2014

    @Mike: Those batteries are re-chargeable ones isn't it ? What sort of a voltage does it require ?  

  18. yaweh_godalla
    May 16, 2014

    I can't access the batteries, but I just measured the voltage across the 6 LED lights at the top and got 3.5 volts DC. The side lamp has 15 LEDs but I can't access those.

  19. Navelpluis
    May 16, 2014

    Assuming that the resistor is part of a R-C couple circuit to derive some energy from the 240V into the bridge diodes: A problem might be is that the resistor was not a wire wound type. Metal film just will not stand pulse loads as our mains net is full of exreme pulses: The cap will fail shortly, but will recover. But then the resistor gets the full load – shortly. That's why the resistor should be a wire wound type – but only if it is sitting in the mains energy supply chain. (often a cap, resistor and zener in series)

  20. mac_droz
    May 16, 2014

    The TO-92 part is a BJT transistor – many manufacturers make it (could be for example KSE13001 from Fairchild) – I know because I used them designing CFL ballasts. Your circuit is just a bridge rectifier with 400V capacitor (as far as I can see) and a self resonanting oscilator. When the transistor fails it normally creates a short circuit across Collector-Emitter pins – I believe your resistor must be a part of a tank circuit and gets the full 300V DC across then – that makes it go bang as the power dissipated would be around:

    300 x 300 / 2400 = 37W which is way more than 0.25W it is rated for.

    You can check if the transistor is broken (most likely a short between C-E pins).

    What you can do is replace the transistor and resistor with the same type:

    http://www.datasheetcatalog.com/datasheets_pdf/K/S/E/1/KSE13001.shtml

    or maybe go for a bigger one (they come as a family starting 13001 and going up to 13009) – I would try 13002 as the bigger the transistor the slower it gets (more capacitance) and that could affect your simple circuit (change frequency).

    The circuit above is so simple it almost should not be used (my opinion) as it does not stop oscillating once you take off the lamp from the cradle – say hi to an additional EM field generator around you (not great for EMC and health reasons).

    As for the foil capactors they can only be killed by too much DC or too high AC current – I'm confident they're fine.

    Enjoy tinkering!

    Maciek

     

     

  21. Victor Lorenzo
    May 16, 2014

    Thanks for taking the picture @yaweh.

    From this marking it seems like a high voltage NPN transistor. The circuit could be one sigle transistor oscillator directly powered from the mains line and driving the charger coil. It does not seem to have any protection device so any voltage peak coming from the power lines directly passes to the transistor.

    Best option to guess what the circuit could look like is a bottom side picture of the PCB.

    My guess could be a voltage peak caused the transistor to fail. The fix could include finding a replacement for the transistor (there are several suppliers but they don't use same pinout for every model), replace the resistor and, probably, include one Bidirectional 300V TRANSIL and one fuse (1A fuse?). The fuse would go in series with the mains line (one of them) and the transil would go between both 230V cables, after the fuse, before the rest of the circuit.

    In case of a voltage peak the transil would absorve its energy, if its energy is too high the fuse would blow.

    It is just a guess.

  22. mac_droz
    May 16, 2014

    It is a Colpitts generator with your 2.4k resistor in emitter and a coil as a load in collector. Once the transistor dies – it shorts Collector (at 300V DC potential) with Emitter so 2.4k resistor sees the whole voltage and pops. Try 13002 in the same configuration, should live a bit longer. The problem with this design is that transistor works mostly in linear mode and so the efficiency will be quite low (below 50% – with remaining power disspated by transistor itself).

    As for other suggested solutions (transils / snubbers) please note that it is not a flyback – there is no leakage inductance that creates oscillations you would like to dump as this is purely linear circuit.

    There are also no transients from mains that could destroy it – electrolytic capacitor acts as a filter and unless you have lightnings hitting ground around you this should not be a concern. Your problems come (in my opinion) purely from thermal failure – transistors have not infinite lifes – the hotter they are the shorter they live.

    As for fuses – you have fuse on the input (in series with one of the AC inputs). Fuses will never protect transistors as their only role is to avoid fire – to protect transistors you need a very fast circuit that can react before the damage is done.

    Hope that helps! Enjoy!

    Maciek

     

  23. yaweh_godalla
    May 16, 2014

    Thanks Maciek,

    I will follow your advice and that of Victor too. I'm in the mood for tinkering. The overload theory seems likely. I recall one unit failed immediately after city-wide power load shedding brought about by extreme hot weather in Perth* and too many air conditioners on high. Power came back on in fits and starts.

    Incidentally the supplier stopped supplying these units about three months ago. Hopefully they are upgrading the design.

    *I assume you are in the USA, which is no stranger to hot weather, but last summer we had a week above 40 deg C, and one day of 46 deg C (115 F).  Hottest spell I can remember in 60 years.

     

  24. mac_droz
    May 16, 2014

    No worries, actually I'm in Ireland now so it is more like 18C outside (very warm) 🙂

    In the worse case you can change all the elements – enjoy!

    Maciek

     

  25. RedDerek
    May 16, 2014

    It is amazing what a good forum can do to help a person out. And to think, this potential solution is truely international – Austrailia, Ireland, USA, etc. Do look forward to the results of the higher voltage/power rated devices.

  26. ejw0
    May 16, 2014

    I'm surprised that no one so far has mentioned the safety aspects of this design.

    It would seem to me that any design where components consistently overheat presents an increased fire hazard.

    Upgrading components to handle higher temperatures would only seem to make matters worse.

    I have on occasion found other products where components have overheated and, while it was an interesting exercise to determine the cause, I was never inclined to “give them a second chance”. Maybe next time they could start a fire rather than just failing?

    I would think twice about “fixing” a design that obviously has a significant fault.

    Just my 2 cents.

    Best Regards,

                         Eric.

    P.S. It is also interesting to note that there are no keywords related to safety for use when posting on this site.

  27. salsaman
    May 21, 2014

    I bought a clone product in the UK from Aldi about a month it has already failed. Opened up the charger cradle and found the same problem. The resistor was burnt out.

    The company who made it, replaced it for me free of charge, but I am worried about the safety of this product as the charger board is just the same as in the picture shown.

    Have highlighted this failure back to this company and am awaiting their response.

  28. alam01be
    September 22, 2014

    Hello

    Ihave repair this lamp.

     

    TO92 is a transistor (KSE13001) i replaceit by a BF871 ( i have this in my

    drawer base)

    I have also to replace the capacitors C6A(56nfd) and C5A (22Onfd).They have lost

    lost their value.

    And  of course R4A (2,4kΩ) by a 2,7KΩ

    All is fine now!

  29. neilw20
    September 8, 2016

    I have 3 of these, so I traced, and drew the schematic.

    2 have failed, both with burnt resistor (R4)

    Both have failed because C5 (5n6/250v) has gone open circuit.

    C5 should be replaced with a 400v rated 5n6 capacitor and R4 replaced

    with 2K4 (not a 2K7!!).  Two 1K2's in series is OK.

    Transistor is MJE13001 (200mA 400v, reasonably fast. Gain – 15 to 20

    There are two different pin arrangements, so TO92 package might look like it is in the wrong way. Base to emitter resistance is slightly lower than base to collector resistor, measured in circuit.

    C6 is the same make, and should probably be replaced. Measured OK.

    C4 has lost some of it's value. This is common, world wide batches of bad ones!

    A BF1 transistor may not survive very long. Only 250v 50-100mA !!

     

  30. neilw20
    September 9, 2016

    As per the previous post, I repaired two by replacing the 2K4 resistor and the 5n6 capacitor. All good.

    The good one that I had not turned on for 3 months, that was working failed after 1 minute!

    Same problem. Another capacitor went open circuit. I felt the heat after the torch came on due to power failure.

    Removed power before the 2K4 was completely fried. New capacitor, and it is all good.

    I had put a dob of hotglue between transistor and the capacitor, and the extra heat made the capacitor fail, I suspect.

    250v capacitors are just not up to it. This capacitor can get close to 300v easily.

    I used 400v rated capacitor.

    The lower 400v capacitor only gets 10 or 20 volts. It will survive.

    I am a bit brighter than when I started.  ™

  31. angeldiazg
    February 3, 2020

    How this came out?
    Worked fine for some time or blew up after a while?
    I have a brennenstuhl night light with same circuit… and same failure. I repaired some time ago changing resistor and cap… and worked for a day or two and failed again.
    Any of your repair jobs working properly after long time? I really like this torch and don’t want to spend another 30€ if is a simple repair.

    Thanks a lot guys.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.