Those interested in mobile lighting applications with HP LEDs may have heard of taskleds B3flex driver: Programmable, batt-protection, multi modes for different applications, temp-control and so on. Compared to it’s Chinese competitors rather expensive, but in my eyes – well spend money.
It has one known issue: The wires connecting main switch and driver board should not be too long. “Not too long” as in not more than 4 inches (whatever that may be in the metrical system)
Well, i didn’t care and went for a long cable anyway and … to make it worse, used a 4x cable, 2lines for power, 2lines for signal, unshielded of course.
Unsurprisingly this did not pay off. One did not need to hook up an oscilloscope to see, that the “switch in” of the main driver IC was collecting a lot of rubbish and started blinking in a random manner and of course failed from reacting properly on any input from the switch.
Greg from taskled suggested to add an 100nF cap, wired parallel to the switch input. I did as told , but it didn’t help very much.
Then, thx to JNGD and microcontroller.net, I went for an additional R in line with one wire (the little white SMD brick soldered to the cap’s leg). In my case something between 3-700 Ohm helped to massively improve signal quality on a cable > 1m. Everything above 700/800Ohm just cut any signal. I guess that value may differ slightly from project to project, but it looks to me like an easy workaround for the wire issue of the b3flex.
comments welcome -