A few months ago my place of employment and enjoyment suffered a catestrophic electrical failure. We recently installed some power factor correction equipment on our three phase mains setup – presumably to save money when powering running some of our more inductive loads.
For those that dont know, power factor is the ratio of real power and apparent power. We use a large number of transformers and motors, thus presenting a large inductive load to the power supply. Inductive loads cause the current waveform to lag behind the voltage waveform. Apparent power is measured in VA and calculated by multiplying the RMS of voltage by the RMS of current. Real power is measured in Watts and calculated by xxxxxx. Typically, power companies charge residential users based on real power, whereas industrial users are charged for apparent power. This means that cost savings may be acheived by ultising power factor correction in industry.
Cutting a long story short, apparently something went wrong whereby the neutral had become disconnected, and one of the three phase lines was used as a return path – or something like that. Needless to say, work sure was fun that day compiling a list of all broken equipment. Around ninety percent of PC’s required new power supplies, all the printers requred replacing, and approx fifty percent of monitors had a nice puddle of eletrolytic sitting at the base. Annoyingly my monitor was the only one in my department to remain completely opperational so everyone else has been gifted with new 24inch screens and I have to make do with a 21inch. The only other monitor to struggle on was a Samsung SyncMaster 245B, that suffered a flickering backlight a week later. This also made it to the scrap pile, where it stayed – until today.
I cracked the case open expecting to find puddles of electrolyte or scorch marks, but found none. Inside were three PCBs – control board, power board and backlight inverter. A casual check over the boards didnt reveal any blown componentry, although some of the resistors showed signs of extensive heat. The display was in perfect condition, other than the backlight, so I instantly dismissed the control board. I didn’t fancy going near the inverter quite yet, as neither my multimeter of myself could withstand a jolt of 3kV. I checked all the obvious voltage points and all the values read as expected.
I reconnected the inverter board, and all 6 CCFL tubes so see if the flickering settled down or got worse over time – in a bit to identify a temperature dependant behaviour. Instead I noticed that something was buzzing along with the flicker of the backlight. As carefully as I dared, I moved my ear closer to the 3kV and 240V voltage sources in a bid to locate the buzzing. As it turned out, the buzzing was coming from the power board, so the inverter was eliminated for now. The power board supplies the inverter with 24V, which measured dead on 24V without the invertor connected. Once connected, this 24V showed a 4V ripple which was well in excess of anything I was expecting.
In my experience, a ripple is caused by unsuitable de-coupling. Granted, my experience is purely low power stuff, but let’s see where this goes. Taking my unsuitable de-coupling theory, and the fact that a load of input protection had been “tested”, gave me a somewhat high expectation that the component in question was the 82uF 250V capacitor. I de-soldered the capacitor and it only measured 6uF. Clearly, if this wasnt responsible for the flickering, it would still require replacing as it was somewhat out-of-spec. I managed to source a 82uF 240V replacement capacitor and soldered it in place.
Tentitively, I powered the monitor back on, and… it worked. I mean, of course it worked – how could it not work. Unfortunately, I sourced an 85deg rated capacitor as a replacement for the 105deg version, but it should hold for now. Hopefully, this replacement will be a statistical outlier, and give me a few years of use – but if the backlight starts flickering I actually ordered 2 replacements.