The monitor worked fine at lower resolutions but the edge was compressed the last quarter inch instead of folding around as had been seen earlier on another computer. The only visible failure was a burned 1/8th-watt resistor.
Philips has a maintenance manual for this monitor but it only has "functional schematics" which do not include the detailed schematics of the horizontal output circuit. These functional schematics do include a lot of useful information including PC board ID numbers so it is useful for finding the horizontal output area. I recommend getting one of these manuals and Philips seems to have a general policy of having reasonably priced manuals on their monitors which many other makers do not.
I also got the Philips manual they used in a monitor repair course that was mentioned the monitor repair FAQ. While it had a lot of good information on monitors in general, it did not mention the totem pole type of horizontal output circuit in this monitor.
It was apparent that a detailed schematic would be very useful to guide troubleshooting but if I wanted it I would have to reverse engineer it. The alternative was to search out bad components with an ohm meter part by part and hope I had found them all. I rejected this and as it turned out, I was right, as I needed this information to understand and finally diagnose the failure and really fix it. I removed the printed circuit board and traced out the circuit over three evenings with short bursts of tedious work.
The power FETs used are each a Toshiba 2SK1544 and the data sheet was located on the Toshiba web site in Acrobat format (2sk1544.pdf) although the site no longer lists the data sheet for this device. This is a 500 volt 30 amp device that is quite hefty.
The FETs are driven by Philips BC237 bipolar PNPs, a rather unremarkable transistor. They were marked "C327" and a tiny "PH" was seen. Considering that Phillips made the monitor, a search of the Philips web site was made. It was located at BC237 Datasheet.
Fast soft-recovery controlled avalanche diodes are used in the circuit and are each a Philips BYV96E. The diode data sheets were found on the Philips web site in Acrobat format at >BYV96E Datasheet. These were useful in understanding the circuit and its limits.
(At this point I pleadingly remind the reader about all these high voltages and currents in this monitor and assure you I, as anyone should, proceeded with great care to avoid damaging myself or the monitor beyond repair. Pleading off.)
With the upper power FET shorted, the output circuit was only rated at 1000 volts rather than the 1500. At 1280x1024 a scope showed a 1200-volt peak voltage using a special 2000-volt probe. The drive voltage to the other side of the horizontal yoke went up a down with the horizontal width adjustment and as expected, so did the peak voltage on the horizontal output circuit. (The horizontal width is adjusted by changing the D.C voltage on the horizontal yoke from a switching power supply on the mail board.) Similarly, the peak voltage on the horizontal output circuit went up with the horizontal scan rate. At 1600x1200 it would fold on the edge and other components would fail soon from breakdown stress.
But this superb monitor should not be sentenced to its remaining life as a lower resolution monitor. The failed parts were replaced. It worked for a few minutes but when it was raised to the full 1600x1200 resolution it once again started folding on the left side and then suddenly R805 gave out a spectacular flash and smoke signal. After removing the P.C. board it was found that the power FET along with all the previously failed parts had failed again.
On reflection, it appears that C801 was causing the failure. This capacitor has a high frequency high voltage across it giving rise to considerable A.C. current. The MKP378 series capacitor data sheet was found on the Philips web site at MKP378 Series Capacitor Datasheet. On examination it was restricted from use as "resonance capacitors in fly-back applications" in the Application Note. The MMKP376 Datasheet on the other hand indicated that capacitor's suitability for deflection circuits.
I suspect they discovered that the MKP378, while working, had too many failures in this stressing mode and modified the design to make the MKP378 series which would not be so prone to these failures. In the construction of these capacitors a number a individual sections are bussed together in parallel. Problems in this bussing could cause some sections to disconnect due to heating from the high currents and the reduced capacitance would let the voltage raise on that section of the totem pole. When C801 was removed it measured 2 nfd. low which is consistent with sections becoming disconnected but would not explain the failure alone. Unfortunately, it is hard to generate the high currents in a test setup to demonstrate this failure theory. But the absence of failure after replacing it would seem to validate the theory.
While the Philips parts site had come up with the BYV96E diodes and the BC327 transistor they had no listings for the capacitors. In the meantime a failed monitor which had had a life of daily service was found. The capacitors were cannibalized. They were in a sequence of 18, 18, 22 nfd. sequence rather than the 15, 15, 18 nfd. sequence in the original monitor. The p.c. board had a higher revision level so the new sequence may have been related to finding a solution to failing capacitors. I would be interested in hearing from anybody who knows how to buy these capacitors especially the MKP378 series.
Since the manual does not include the detailed schematic of the horizontal output stage it is available as a hand drawn GIF file. I know the main circuit board where this section resides is also the same part number as the one for the C2082 which is a Trinitron" version of the monitor. The C2082 was OEM'd to Compaq as a Q-Vision 200 monitor. It was also OEMed to DEC with DEC markings as a FSR871CV. I also suspect it is the same if not similar to other Philips 20 and 21 inch models and probably also OEMed by Philips to others. An additional problem was encountered during this repair. To change out the power FET required removing the heat sink and replacing it. The sharp edges of the new power FET next to the silicone insulator sheet gouged enough of a hole in the sheet to cause a high voltage break down since that power FET was at 1600 volts for a short period of the horizontal sweep. It required a replacement from the cannibalized monitor. It would be wise to use a little sandpaper to round the edges of the new power FET before installing it onto the board. During assembly into the monitor be sure that the mounting holes for the P.C. board and the heat sink line up to avoid sliding the heat sink in relation to the P.C. board.
-- end V1.00 --