2456d3 failures

[apostolakisl]

I have two lamplincs that i pulled out for my Christmas decorations and found they both were completely dead (no indications of power or anything at all but not popping a breaker). They had been plugged in since last year but no load attached and the program that operated them disabled.

 

I took them apart and both appeared to have the same failure. A small componant at the center of the board labedled "D3" next to it was blackened all the way though the board. It is a small square piece about 1.5mm by 3mm. They are rev 2.4 and 2.7 made in the 41st week of 05 and 34th week of 06.

 

Any knowledge base on this out there?

[Brian H]

Since it is heavy enough to handle the load. I doubt the fuse is popped.

But did you check the fuse that is in the black shrink tubing from the AC Line input pin to the PCB?

 

I checked both a V2.0 and V2.6 revisions. Both are basically the same.

D3 in the 2.0 is marked 30A the one in the 2.6 15/30.

They are wired as a zener diode and measure close to 30 volts across D3. The 2.0 was 29.5 volts and the 2.6 was 30.2 volts. Educated guess is 30 volt maybe a 1.5 watt from the 15 marking on the diode in the V26.

 

One point. The banded end of D3 {+30 volt connection} and the +5 Volts VCC for the electronics are directly connected to the AC Line Input. The older ApplianceLincs are the same way.

 

If D3 is fried. It may have just failed or maybe one of the other diodes in the circuit went bad and too much current went through it.

 

I do know Ken M. had an early ApplianceLinc fry D3.

[apostolakisl]

Brian,

 

Thanks for the info.

 

I tested the fuse and it is intact. From the line-in to either side of the diode (D3) is no resistance. Either side of the doide to neutral is open circuit.

 

The diode has nothing that can be read on it (burned off). So I don't know the polarity or anything else about it except the general shape. These are my only ones of this model so I can't open another one to check it. (I have several older icon ones that continue to work just fine)

 

The diode has to be fried from the appearance of it. My main question is whether it is the victim of another failure or if it is the primary (and only) failure point. I could solder in another one if I knew exactly what to buy and which way to put it in.

 

It seems odd to me that these failed over the 10 months from last Christmas to now without anything plugged into them.

 

Lou

[Brian H]

The older 2856DB Icon Dimmer Modules have the same basic PCB in them as the older 2456s. Different firmware and no AC pass through outlet.

 

The side of the D3 pads closest to the D3 screening on the PCB is the banded cathode and should be connected to the AC Line Input. The AC Line Input should also connect to pin #5 of the programming connector pads J5. Lower right edge of PCB.

 

Sounds like you may have some runs burned open when D3 failed.

[apostolakisl]

Sounds like you are quite familiar with the circuitry on these things.

 

So do you think it will be fruitless to replace the diode? I am reading your statement "it sounds like may have some runs burned open" to mean other stuff is fried.

[Brian H]

With no Line Connection to the one side of D3. At the very least the PCB run to it burned open when D3 failed.

 

Replacing D3 and not repairing the open connection would not fix the module.

 

No I am far from an expert on the hardware. I just did some early observations when I joined the Developers Group. That at the time was being sold for "Those just interested in how Insteon Hardware works".

I have done lots of X10 module mods and tests.

 

There is a partial power supply schematic of the Icon relay version of the module on efundies.com.

http://efundies.com/guides/

[apostolakisl]

I found this on cocoon website.

 

I can tell you we track failures of every returned Insteon product. Here is an example over a six month time span:

 

R1 open 27.84%

D3 shorted 23.71%

Triac burnt or shorted 20.27%

MOV burst 6.01%

Air gap failure 5.50%

Bad soldering 4.47%

Paddle issue 2.41%

Load Flutter 2.06%

Bad component 1.89%

C1 exploded 1.37%

Not available 1.37%

Bad LED 1.03%

P/S 0.86%

5V to ground 0.34%

Bad switch 0.52%

Wrong ID 0.34%

 

I guess I got my unit before they ran this stat and upgraded the first few items on the list.

 

I am going to try removing d3 and see if both contacts still show closed to hot. If the far side shows resistance, I will try finding a replacement.

[lawr1000]

From Brian H's measurements your D3 is probably the same as D1 in the schematic he posted. This is a shunt power supply configuration using the zener diode to produce approximately 30 VDC. R1 and C1 limit the current in the zener diode. So a shorted R1 or C1(in particular C1) will cause the Zener to burn up. With a 30V drop across the zener it doesn't take much current to produce power levels that exceed the device's power rating.

 

We have now way of knowing if the schematic matches your device so measuring open to either side of the diode may be normal. Given the device was working when you took it out of service my guess is the power supply was poorly designed in the first place and the zener was cooking from day one. So there is a chance that just replacing it might work (if you can discover the actual part number of the Zener Diode).

 

However, I would advise you to consider safety. These devices have high voltages inside and represent a serious hazard to your health. Also if you manage to get the device working it might be weakened and cause a fire at some point in the future.

[Brian H]

It is not normal for D3's banded end and +5 volts VCC not to connect to the line input.

 

I checked both an Icon Dimmer and LampLinc of the same vintage and it was connected to both places.

 

Can't say about the later revisions. I know my rev 4.1 ApplianceLincs are 100% different and maybe the LampLincs also. Don't have any to look at.

[lawr1000]

Based on your testing then your are probably right in guessing a PCB trace is burned too. Unless apostolakisl had a measurement error due to flux or conformal coating. I would at least take the measurement again using the meter in both directions. If it is still open then there was some other catastrophic failure and it is unlikely worth repairing.

[Brian H]

You maybe correct on the D3 dissipating a fair amount of power.

I have a LampLinc on my test stand and the back of the case gets noticeably warm.

 

There where some reports of failures in that area in the now defunct AccessHA forums.

[apostolakisl]

So I popped D3 off and tested across the two contacts. The result is open circuit. The pcb is damaged from the burn and I don't think you could solder another one on the same way. There are several locations that are electrically equal on the pcb that jumpers could be soldered to.

 

I did try powering it up with d3 out. . . still dead.

 

It appears that the diode shorted. I wish I felt confident about which diode to but back in and, of course, a place to buy 2 of them.

[Brian H]

So you would be looking for a different style that is not surface mounted to the same set of pads?

 

The markings of 15 30 I could not find. We know the voltage is 30 volts but the wattage is a guess of maybe 1.5 watts.

 

I did find a few surface mounted 30 volt 1.5 watt zener diodes on the Mouser site, but if the pads and runs are questionable?

 

Was the line input still going to pin 5 of J5? If not the run from line to D3 may also be to J5-5.

 

No easy way to tell if repairing the runs and replacing the diode would fix it.

 

I am thinking on the same lines as lawr1000. There could be more parts bad. Though the diodes are cheap more for postage than the parts.

[apostolakisl]

Yeah, I once bought a $4.00 capacitor and they charged me $14.00 for shipping. I just laughed.

 

I would only do this if I could just hit radio shack and pay $1.50 or something like that.

 

It is really more of a "can I do it" than a cost effective use of my time.

 

Ande the diode is definitely shorted. I was able to get contacts on the burnt/broken off ends and it is closed.

[nstein]

Almost all of my older LampLincs died on me; I never opened them up, but I was able to get SmartHome to exchange them all under the extended warranty e-mail, they sent out in 2008 (It included the SwitchLinc paddle problem):

 

We have identified and corrected an issue with LampLinc Dimmers shipping before January 1, 2008; these units may have a status LED that does not illuminate when plugged in to a powered outlet or does not respond to "Set button" presses. We identified a component replacement that corrected the issues and all LampLinc Dimmers shipping on or after January 1, 2008 should not observe these issues. If you have a unit that has a status LED that does not illuminate when plugged into a powered outlet or does not respond to "Set button" presses, please call our Customer Solution Center @ 800.762.7846 for a warranty replacement.

Pertains to Smarthome Product No.: 2456D3, 2456D2

Smarthome stands behind every product we sell. Rest assured that should your LampLinc experience this behavior in the future, Smarthome will extend the warranty for this issue for an additional 5 years for all LampLinc Dimmers purchased before January 1, 2008. We will extend 2-year warranties to 7 years and 7-year warranties to 12 years for this issue.

 

While it does not say dead LampLinc I will say "these units may have a status LED that does not illuminate when plugged in to a powered outlet or does not respond to "Set button" presses." qualifies.

 

-Nick

[apostolakisl]

Almost all of my older LampLincs died on me; I never opened them up, but I was able to get SmartHome to exchange them all under the extended warranty e-mail, they sent out in 2008 (It included the SwitchLinc paddle problem):

 

We have identified and corrected an issue with LampLinc Dimmers shipping before January 1, 2008; these units may have a status LED that does not illuminate when plugged in to a powered outlet or does not respond to "Set button" presses. We identified a component replacement that corrected the issues and all LampLinc Dimmers shipping on or after January 1, 2008 should not observe these issues. If you have a unit that has a status LED that does not illuminate when plugged into a powered outlet or does not respond to "Set button" presses, please call our Customer Solution Center @ 800.762.7846 for a warranty replacement.

Pertains to Smarthome Product No.: 2456D3, 2456D2

Smarthome stands behind every product we sell. Rest assured that should your LampLinc experience this behavior in the future, Smarthome will extend the warranty for this issue for an additional 5 years for all LampLinc Dimmers purchased before January 1, 2008. We will extend 2-year warranties to 7 years and 7-year warranties to 12 years for this issue.

 

While it does not say dead LampLinc I will say "these units may have a status LED that does not illuminate when plugged in to a powered outlet or does not respond to "Set button" presses." qualifies.

 

-Nick

 

Nick,

 

I really didn't think this one would be covered under warranty so I didn't try. I think I will call them tomorrow. Thanks for pointing that out.

 

I just pulled apart my working icon lamplinc and as it turns out, it is exactly the same board and components. The only difference is the grounded plug and the pass through outlet.

 

Anyway, the d3 diode is running hot on this unit as well. It has blackened the pcb though it continues to function.

 

I could read the 30 and 15 as well as identify the polarity banding. Clearly this was an undersized componant and I would be surprised if my two lamplinc icon switches don't die soon as well.

 

Lou

[Brian H]

I have a hardware 4.3 LampLinc and it has a different circuit board in it, but the power supply is basically the same configuration. One difference I did see was the Zener Diode was a different type. A large glass cased one.

 

I have no ides what the latest >5 revision ones have.

 

My revision 4.1 ApplianceLinc has the same zener diode as the later LampLinc does. Also a different PC board as they went from 120 volt AC relays to 24 volt DC ones.

 

Both of the later ones I looked at. Had a voltage of 31 volts on the zener diode.

[apostolakisl]

I RMA'd them to SH today.

 

My icon ones (which still work) are going to fail at D3 as well sometime based on the brown PCB that surrounds it.

 

I don't know where I bought them from and SH doesn't honor the warranty as a manufacturer, only as retailer (big cop-out in my opinion).

 

So, if it might be better to swap it out for a more capable diode before it fries rather than wait for it to damage the PCB traces and possibly other components as well. Then again, I may just take a working unit and kill it.

[Brian H]

I am glad you where able to get warranty replacements for the LampLincs.

I never even got the email about possible early ones with problems.

 

The warranty thing if not sold by them has been seen before. Smartlabs is the manufacturer who sells through dealers. Some dealers where saying call Smartlabs who then refer you to Smarthome. Since they didn't sell it they will not replace it and may tell you to call the seller.

 

As for finding a replacement. I have found a 30 volt 3 watt SMA surface mounted zener on the Digikey sales site. Though we don't know the originals wattage.

 

The problem maybe more of the mounting pads size as they are part of the heat sinking for the diodes itself.

 

There are some with leads also if you can fine a place to solder them and have them mechanically stable.

[apostolakisl]

I checked out digikey and they sure have a lot of choices. If I took a working one off the board is their a way to measure the specs? I doubt it. The 30 I assume stands for 30v. The 15 could be 1.5 watt. I would need to look at it closer meaning that I would need to take it apart more and since it works I didn't do that.

 

It would be easier to not surface mount the diode as there are a number of holes in the pcb that are on the same trace that would be easy to solder jumpers into. It might do better at heat disipation if it weren't on the pcb too. At least it wouldn't damage the pcb.