DMX Decoder - Channel output stuck at 100%

Device: DMX 512 RGB decoder

Issue: DMX decoder outputs red at 100% constantly and does not respond to incoming DMX control signal.

Decoder continually outputs Red at 100%

The intended operation of this decoder is to receive DMX512 signal packets from a controller and, if the relevant command arrives, output corresponding RGB values to a strip of LED tape. 

PX24506 bought from eBay

..and over the course of about a year, it did just that. But then one day, after testing a new lighting controller and getting some strange results, it became apparent that the red channel was not working correctly. No matter what values of RGB was sent to the device, red would always seem to be being sent to 100%.

Tracking down this issue consisted of breaking down the signal path and ruling each stage out until the faulty agent was discovered. 

Starting at the controller I disconnected the decoder and plugged straight into another type of LED RGB fixture, only to find that the red channel was this time functioning correctly. This left either the decoder or the LED tape. Further deduction was made by simply cross plugging the red channel of the decoder into the green channel of the LED tape, and visa versa. This time when powering up the green LEDs instantly come on and sending DMX data for the green channel allowed full control over the red LEDs. This proves to us that all channels of the LED tape are working correctly and the problem must come from the decoder's output.

Breaking out the screwdriver we open up the decoder to get a better look at what's going on.

Exposing the circuit board

And on the surface, everything appears fine. From the image above there are no obvious signs that anything is wrong. In order to figure this one out we need to learn a little theory on how LED tape works.

LED tape shares one common line in for voltage to all LEDs which then returns to ground via either the red, green or blue channels. Since the LEDs draw in the range of 20mA of current, far too much for the tiny micro-controller which likely operates in the range of 3 - 5V, transistors are commonly used to act as a gate which the micro-controller can then open connecting each channel to ground and allowing light to emit.

These transistors seemed like a good place to continue the investigation. It made sense at this point to try and figure out which transistor belongs to which colour. Fortunately studying the circuit board gives us this information.

Identifying the transistors

Once identified it was worth checking that the pins on the red transistor were making good contact with the circuit board. This comprised of both a simple visual inspection and a continuity test. I traced the copper track from each pin to its next component and checked there was no resistance. Everything seemed in order so far.

Looking closely at the components we can see that they are Alpha & Omega D514 MOSFET transistors. Within this datasheet we can see the arrangement of the Source, Gate and Drain pins which now allows allow us to do some basic continuity testing. 

Signal route through the component

In order to see if the transistor is correctly closed by default we check for the absence of continuity between the Drain and Source pin. When measuring this appears to be the case for the blue and green transistors however continuity was present for red. 

This suggests that by default the transistor is open for the red channel allowing current to flow through when power is applied. No matter what level the micro-controller sends out to the transistor it remains open which corresponds with our problem exactly.  

The solution therefore is to simply replace the faulty component. A quick browse of eBay found them for sale at £5.59 for a pack of 8.

I haven't quite cracked de-soldering these types of components off of a circuit board yet as all three pins need to be completely free of solder before they can be lifted off as a whole. As the transistor was broken anyway, I instead opted to just cut the pins of and remove them one at a time.

The new transistor soldered on

With the new transistor fitted I gave it a quick test and to my delight observed correct functionality once again. There's something to be said about spending £5 to fix a £20 decoder, but that's one for another time.

Issue solved!

The red channel is no longer stuck at 100% output