Maxhawk Posted June 4, 2016 Posted June 4, 2016 I have 3 of the 2672-222 bulbs in the house. Two are in my family room and have been working fine since I installed them about 30 months ago. The third bulb failed after about 12 months of use (bought at the same time as the others). One difference is that this bulb is on a wall switch. In normal use, I keep the bulb in the ON position through my ISY-994i and use the wall switch to turn the lamp on and off. When I go on vacation, I leave the wall switch on and have a vacation program that turns the light on and off. Since the bulb was out of warranty (purchased November 2013) I decided to open it up just to see what was inside. Upon initial inspection it was obvious what the failure was: the main bulk cap off the rectifier was bulging and removal and measurement showed it as an open circuit (zero uF). I had already relegated it as junk and wasn't very careful with my disassembly. Since I didn't know how it came apart I used my bench vise to help open it. It turns out that it should unscrew apart and I've noted such in my photos. For anyone who does attempt to repair their bulb, note that the header that connects the two boards together was extremely difficult to desolder. Normal amounts of heat and solder wick wasn't working so I had to really crank up the temperature to get the solder to melt. I could tell that it was damaging the board so I then tried my SMT hot air station to heat up the pins simultaneously. This also didn't work so I finally just clipped the header pins with my diagonal cutters. If you want to salvage this connection I suggest a vacuum desoldering station (https://www.hakko.com/english/products/hakko_474.html) to remove the solder. Since the antenna is appears perfectly coiled, trying to recoil it during reassembly might mis-tune it. Perhaps someone with more RF knowledge can comment on this. So I have 2 conclusions based on this exercise: 1. This device has a similar failure as the 2413S PLM with its cheap Chinese capacitors and use of underrated parts (http://forum.universal-devices.com/topic/13866-repair-of-2413s-plm-when-the-power-supply-fails/). 2. It's best not to operate the Insteon bulb with a wall switch because switching the AC likely caused voltage spikes which led to the premature demise of the capacitor.
stusviews Posted June 5, 2016 Posted June 5, 2016 One out of three failures is weak evidence. The fact is that you did not use the bulb as intended, that is, uninterrupted power. Just saying
Techman Posted June 5, 2016 Posted June 5, 2016 Was the bulb in an enclosed fixture? They generate heat and need air circulation.
larryllix Posted June 5, 2016 Posted June 5, 2016 One lesson I have learned is if you buy a group of LED bulbs for a multi-lamp fixture, always buy at least one extra, mark it's purpose and put it away. In a group of five bulbs in my dining room chandelier, one of the CREE (yeah the fade away type) bulbs went dim. Trying to replace it, the style of bulbs changed about a year ago and the old style was not to be found in 60W equiv. The Borg replacement bulb was larger, ugly with slotted holes in the bottom, but that was no matter. The bulb has a completely different dimming curve tan the original ones so that at 20% when all the original bulbs are very dim, this one is glowing about 50% perceived brilliance. You may get a warranty replacement free but it may not match the style, brilliance, or dimming curve of the mated bulbs. Get at lest 25% more bulbs than sockets for each LED multi-lamp fixture.
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 Was the bulb in an enclosed fixture? They generate heat and need air circulation. No, it was used in a cheap Idea lamp like this:
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 One lesson I have learned is if you buy a group of LED bulbs for a multi-lamp fixture, always buy at least one extra, mark it's purpose and put it away. I certainly appreciate what you're suggesting, but in my case it was just one bulb in one room and I did have an extra. However in the long term, I'll probably be better off installing an Insteon wall switch with a regular bulb.
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 One out of three failures is weak evidence. The fact is that you did not use the bulb as intended, that is, uninterrupted power. Just saying You're right, but I still maintain that the design is not robust. I come from an industry where our customers complain incessantly about these types of failures so we overdesign our products with margin to build in robustness to prevent this type of failure.
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 That type of enclosure actually traps heat. Yes perhaps if it were shining downwards, but I have it sideways shining horizontally towards a wall for wide indirect lighting. I still believe that transients from the wall switch killed that cap along with Chinese capacitors that may not have the proper ripple rating for the application. This is extremely common in the Chinese-made consumer market items.
Teken Posted June 5, 2016 Posted June 5, 2016 For the benefit of others could you please indicate the hardwares production date and revision number along with the ISY firmware? I know you indicated it was purchased in 2013 but would like to know the actual values.
larryllix Posted June 5, 2016 Posted June 5, 2016 Look at how many times CREE changed their bulb designs as well as the CFL companies. Next thing you know most of the bulbs have "Not for use with dimmers" listed on them, as well as "Must be mounted base down". LED lighting history is young yet, and the ugly truth will be arriving, as we learn the $hard$ way.
Teken Posted June 5, 2016 Posted June 5, 2016 I agree when the race is to the bottom in terms of price most of the time quality is the first thing to go out the window. One would figure in 2016 with the massive amounts of technological advancements in component design there would never be any sort of bad part available. Sadly, even in 2016 you can still buy poor quality resistors, capacitors, diodes, IC's etc.
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 For the benefit of others could you please indicate the hardwares production date and revision number along with the ISY firmware? I know you indicated it was purchased in 2013 but would like to know the actual values. I can't find a production date on the bulb nor the box. There is a small label on the bulb indicating "R2.5". My ISY is currently running v4.4.6.
Teken Posted June 5, 2016 Posted June 5, 2016 I can't find a production date on the bulb nor the box. There is a small label on the bulb indicating "R2.5". My ISY is currently running v4.4.6. Sorry I meant to ask what does the ISY indicate the firmware is for the LED bulb?
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 Sorry I meant to ask what does the ISY indicate the firmware is for the LED bulb? Since the original bulb is fubar, I'm doing this with the remaining 3 bulbs that were purchased at the same time. However it's not obvious to me how to query the firmware or hardware revision as I don't see such an option when right clicking the device in my admin console. However the information at the top of the screen when the device is selected indicates v.42.
Brian H Posted June 5, 2016 Posted June 5, 2016 v.42 is the firmware reported to the Admin Console when added to your ISY994i controller. I don't believe the hardware revision is available in the Admin Console. Only printed on a white sticker or maybe silk screened on the bulbs base or shell. I looked at the FCC Database photos but could not read the power supply IC part number. Could be the same LNK520 used in the PLMs and other newer hardware revisions of the Insteon devices. If it is. Then the capacitors should have been rated to be used in a switching type power supply.
Maxhawk Posted June 5, 2016 Author Posted June 5, 2016 v.42 is the firmware reported to the Admin Console when added to your ISY994i controller. I don't believe the hardware revision is available in the Admin Console. Only printed on a white sticker or maybe silk screened on the bulbs base or shell. I looked at the FCC Database photos but could not read the power supply IC part number. Could be the same LNK520 used in the PLMs and other newer hardware revisions of the Insteon devices. If it is. Then the capacitors should have been rated to be used in a switching type power supply. The only sticker I can find is "R2.5" on the base of the bulb along the the standard sticker that contains the model number, FCC ID, etc. The switching IC used on the bulb is an LNK304DN.
Brian H Posted June 5, 2016 Posted June 5, 2016 Thank you for the added information. R2.5 is probably the Hardware Revision. I have a few newer modules that start with an R. An R at the end of a part number is a different thing. The R on the end of a part number indicates Refurbished. I found the Datasheet on the LNK304DN. It runs at 66kHz oscillator frequency. So it would also need a capacitor designed for a switching power supply. I also have a correction. The 2413S/U PLM used a LNK354GN.
Teken Posted June 5, 2016 Posted June 5, 2016 Since the original bulb is fubar, I'm doing this with the remaining 3 bulbs that were purchased at the same time. However it's not obvious to me how to query the firmware or hardware revision as I don't see such an option when right clicking the device in my admin console. However the information at the top of the screen when the device is selected indicates v.42. Let me check a few I have on hand . . .
Maxhawk Posted June 6, 2016 Author Posted June 6, 2016 Thank you for the added information. R2.5 is probably the Hardware Revision. I have a few newer modules that start with an R. An R at the end of a part number is a different thing. The R on the end of a part number indicates Refurbished. I found the Datasheet on the LNK304DN. It runs at 66kHz oscillator frequency. So it would also need a capacitor designed for a switching power supply. I also have a correction. The 2413S/U PLM used a LNK354GN. I've got some more info on the circuit for those who are interested in what's inside. AC comes in with a MOV directly on the input and then to a full wave bridge rectifier. The output of the rectifier has a C-L-C-L-C filter. All those caps are 250V electrolytics and all 3 of them (10uF, 2.2uF, 2.2uF) measure open circuit. The rectified DC feeds a 9910-5 LED driver and the LNK304 buck converter. Based on the 2K and 20K resistors of the feedback circuit it looks like the LNK304 is down converting to just under 19VDC. I didn't trace it out, but I believe this 19VDC is being used to drive the MC33063 DC-DC converter. The MRF49XA ISM radio is 3.3V and the PIC16F688 micro works from 2-5V, so it makes sense that the Motorola DC-DC converter is regulating down to 3.3VDC. At the end of the day, either voltage spikes or heat killed the caps. They're rated at 105C but who knows if they have a decent ripple current rating. They are cheap Chinese capacitors that I've always stayed away from due to reliability issues. You'll find plenty of information on the internet about this subject.
Brian H Posted June 7, 2016 Posted June 7, 2016 Thank you for the added information. I don't have much faith in the Chinese Fujicon Capacitors in the PLM. I suspect what ever is in the other modules are not much better.
larryllix Posted June 25, 2016 Posted June 25, 2016 I just had another Cree bulb go bad. Used in my open bottom, shine down, lantern style, in outside garage lights on a MS for about a year of service. That makes about 6 of my 12 Cree bulbs have gone bad now. So far, with mostly Cree purchased in my early days of LED bulbs, they haven't lasted as long as decent quality incandescent bulbs do.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.