I remember someone mentioning that a Bestec ATX-250 12Z was acting flakey, it quit working, then a few days later it ran fine again. I drew a schematic of the 5vsb section to see what I could find, also comparing it with the problem ridden "12E". So to begin, here is the schematic.
I was pleasantly surprised that it had some protection features I was totally unaware of. The 5vsb pwm chip has a crowbar circuit (politically correct term is "latch"). There is also a SCR circuit for 5vsb overvoltage protection.
Looking at the drawing labeled "crowbar", the left side shows the basic parts of the regulation circuit for a normally operating power supply. On the right shows the SCR protection circuit. When the 5vsb bus begins going much above 5 volts, the zener ZD3 begins conducting causing current flow through the LED section of the optoisolator PC3. This causes the output side (pins 3 & 4) to reduce resistance to the point that current flows into the scr gate causing the scr to turn on. Now there is a 1000 ohm resistor between the 16 volt line and the feedback pin, pin 4, rather than the 6800 ohm resistance plus any resistance on the output side of optoisolator PC2. End result is more current at pin 4 and the pwm chip crowbars.
To test this feature, I jumpered the 12 volt bus to the 5vsb (no motherboard hooked up to power supply!). It instantly crowbared and shut down the power supply. The voltage on the 16 volt line dropped to 2.14 volts and stayed. The only way to release the crowbar was to unplug the power supply cord from the wall socket. After a couple of minutes to drain off the electrons, I turned the power supply back on. Next I used a jewelers screwdriver to touch between pins 3 and 4 of optoisolator PC3. This also turned the SCR on, and the power supply crowbared. The power supply instantly turned off. Same voltage readings. Unplug the power supply cord from the wall socket, wait a couple of minutes, then turn power supply back on. Looks like the problems with 5vsb overvoltage are solved with this circuit.
For further analysis, I turned on the power supply, and using a small jumper wire, shorted the 12 volt bus. Instant crowbar and power supply off, 16 volt line dropped to 15.50 volts and stayed. After taking power supply cord from wall socket, voltage bled down to 14.45 volts, then quickly dropped. At this point I turned power supply back on. If it works for shorts on 12 volt bus ....... I shorted the 5 volt bus. Same result. Then I tried shorting the 3.3 volt bus. Same result. This looks like better short circuit protection than I have seen on most computer power supplies.
If you are unaware of this crowbar effect, I'm sure it will seem like a flakey power supply. But now I hope some of you will dig out that flakey "12Z" and have some fun playing with it.
I was pleasantly surprised that it had some protection features I was totally unaware of. The 5vsb pwm chip has a crowbar circuit (politically correct term is "latch"). There is also a SCR circuit for 5vsb overvoltage protection.
Looking at the drawing labeled "crowbar", the left side shows the basic parts of the regulation circuit for a normally operating power supply. On the right shows the SCR protection circuit. When the 5vsb bus begins going much above 5 volts, the zener ZD3 begins conducting causing current flow through the LED section of the optoisolator PC3. This causes the output side (pins 3 & 4) to reduce resistance to the point that current flows into the scr gate causing the scr to turn on. Now there is a 1000 ohm resistor between the 16 volt line and the feedback pin, pin 4, rather than the 6800 ohm resistance plus any resistance on the output side of optoisolator PC2. End result is more current at pin 4 and the pwm chip crowbars.
To test this feature, I jumpered the 12 volt bus to the 5vsb (no motherboard hooked up to power supply!). It instantly crowbared and shut down the power supply. The voltage on the 16 volt line dropped to 2.14 volts and stayed. The only way to release the crowbar was to unplug the power supply cord from the wall socket. After a couple of minutes to drain off the electrons, I turned the power supply back on. Next I used a jewelers screwdriver to touch between pins 3 and 4 of optoisolator PC3. This also turned the SCR on, and the power supply crowbared. The power supply instantly turned off. Same voltage readings. Unplug the power supply cord from the wall socket, wait a couple of minutes, then turn power supply back on. Looks like the problems with 5vsb overvoltage are solved with this circuit.
For further analysis, I turned on the power supply, and using a small jumper wire, shorted the 12 volt bus. Instant crowbar and power supply off, 16 volt line dropped to 15.50 volts and stayed. After taking power supply cord from wall socket, voltage bled down to 14.45 volts, then quickly dropped. At this point I turned power supply back on. If it works for shorts on 12 volt bus ....... I shorted the 5 volt bus. Same result. Then I tried shorting the 3.3 volt bus. Same result. This looks like better short circuit protection than I have seen on most computer power supplies.
If you are unaware of this crowbar effect, I'm sure it will seem like a flakey power supply. But now I hope some of you will dig out that flakey "12Z" and have some fun playing with it.
Comment