![]() |
|
|||||||
|
|
Thread Tools | Display Modes |
|
|
#21 |
|
o.O
Join Date: Sep 2007
City & State: Duisburg
My Country: Germany
Line Voltage: 230VAC 50Hz
Posts: 2,616
|
^ I know. We had a couple of those a while back and they all failed.
When the last one died (flickering and went out), I took it apart out of curiousity. One of the tiny 0.25W resistors was open. Replaced all of them, replaced the jumperwire labeled as F1 with a fuse and replaced the aishi capacitor (400V 2.2µF IIRC).Worked well enough as a nightlight in the hallway, as it was already somewhat dim. Now, half a year later, it's useless even as a nightlight. My LED flashlight with almost dead batteries is brighter.. |
|
|
|
|
|
#22 |
|
Badcaps Veteran
Join Date: Apr 2011
City & State: University of Leeds (MEng ElecEng undergrad)
My Country: UK
Line Voltage: 230Vac 50Hz
I'm a: Student Tech
Posts: 13,538
|
I was under the impression that most of those bulbs didn't even use a SMPS, just a large X-cap in series. This (fake uncertified) X-cap likes to catch fire so I consider the cheap LED bulbs as worthless crap. Also, the efficiency of 5mm LEDs is about the same as CFLs - probably worse. It's only when you get to high power LEDs that efficiency becomes really good.
Here's a schematic of a very cheap LED driver, but it works surprisingly well. It's efficient (83%), it's current regulated, doesn't use electrolytic caps and won't get hot pumping 4W into LEDs... hottest part is the FET at 180mW. And it reaches full brightness in ~3.5 milliseconds. It is a modification of the cheap two transistor circuit, but made more reliable by avoiding several electrolytics and modified so that the current through the LED is limited and proportional to the current supply voltage.
__________________
** Begin Signature ** I fix TVs and electronics as a hobby and to save nice things from the dump. 40 LCD TVs, 4 monitors, 13 plasma TVs, and a 6.1 system! Free service manuals database: http://www.tgohome.com/ServiceManuals - Have a manual? Need a manual? Please PM me. I have successfully fixed (from best --> worst build quality): 3 Panasonic Plasmas, 1 Yamaha HTS, 1 NEC Plasma, 2 Sharp LCDs, 2 Toshiba LCDs, 5 Philips Plasma/LCD, 1 Hitachi Plasma, 5 LG LCD/Plasma, 10 Samsung LCD/Plasma, 1 Thomson Plasma, 1 Atec LCD, 1 Hanspree LCD, 1 Xerox LCD, 1 Harwa LCD, 2 Proview LCD, 2 Hyundai LCD, 1 "Onn" LCD, 1 Dell LCD, 1 iiyama LCD, 1 Logik LCD, 5 Bush LCD, 10 Vestel LCDs [Please do not PM me for help: I am a busy student, and we prefer input from all sides on this forum.] Last edited by tom66; 06-14-2012 at 09:25 AM.. |
|
|
|
|
|
#23 | |
|
Badcaps Veteran
Join Date: Jan 2011
City & State: Trenton, NJ
My Country: USA
Line Voltage: 240v-120v 60Hz 200A service drop
I'm a: Knowledge Seeker
Posts: 1,959
|
Quote:
__________________
Hi-pot test: FAIL |
|
|
|
|
|
|
#24 |
|
Badcaps Veteran
Join Date: Apr 2011
City & State: University of Leeds (MEng ElecEng undergrad)
My Country: UK
Line Voltage: 230Vac 50Hz
I'm a: Student Tech
Posts: 13,538
|
Yeah, I think it's unique to me. I've just cut the feedback from the two-transistor circuit (ringing choke converter) and adjusted the components for maximum efficiency and current regulation instead of voltage regulation. The clincher was realising that the current through the LED didn't need to be constant. After all, we don't notice the flicker from incandescents. So you can get away with basically no filter cap, and that's the number one failure. There's also no need to isolate the LEDs so you can get rid of y-caps and optos, the overvoltage protection is simply the resistor tied from the secondary thru the current limit transistor. It would cost less than $1 to produce in volume.
|
|
|
|
![]() |
| Thread Tools | |
| Display Modes | |
|
|