Solution Finder

Reliability of dimmers

Posted by: Jon Smirl on Wed, 08/25/2021

When building a dimmer for lighting, how can the reliability of the power supply be increased? I'm asking this because I've installed a lot of Insteon products which have a LNK304 inside. They are about 7-10 years old and the power supplies are dying in hundreds of devices forcing them to be replaced. The failure mode is exactly the same in all of devices -- they start blinking on/off. They work for a few seconds, then warm up and fail, and repeat when they cool.  The TRIAC is not involved, same failure mode if the lighting load is not turned on.

Any comment on why the Insteons are failing? I have disassembled multiple units and there is no visible sign of component failure. The failure mode is common to dimmers, on/off, and keypads.

How would the reliability of the older Insteon design compare to the new DER-865 design?

Would using an isolated supply for the MCU and a small discrete supply for the TRIAC/MOSFET drive be more reliable?




Submitted by PI-Bombadil on Mon, 09/06/2021

Hello Jon,

Thank you for your inquiry. Would you know the circuit used in the Insteons? Is this the typical buck converter? If the products are 7-10 years old, perhaps some of the parts have reached their life. Most common are the electrolytic capacitors or perhaps the optocouplers, if the circuit used requires one.

In terms of reliability, isolated vs non-isolated should not have much difference. In fact, if one used isolated supply and would need to use an optocoupler, that would be an added consideration for reliability too.

Comparing with DER-865, reliability should also be same with the proper selection of parts, specifically, the capacitors. The new thing with DER-865 is it uses LinkSwitch-TNZ, which has a zero crossing detection (ZCD) feature, so there's no need to add extra circuit just to detect and process the zero-crossing of the input voltage.



It is the typical buck converter. 

I have found one consistent failure mode. They have mounted electrolytic capacitors on their sides. Then components near the caps have radiated enough heat to melt the plastic covering on the cap. That melted plastic is pooling on the PCB below and likely shorting stuff out.

Another failure mode appears to be poor flux removal. There is noticeable flux left on the board, not sure they cleaned them at all after hand soldering components. This flux has baked until turning brown/black.

There are various revisions of the device too. Not all contain LNK chips.