Chemistry Electronics Services
ARCHIVE OF REPAIR TRICKS, USEFUL INFO, ETC.

5415 Eppendorf Lab Centrifuge: randomly fails to power up. Power supply fail?

DANGER: THE ENTIRE BOARD IS EITHER LIVE 120V, OR FLOATS AT 170VDC!
If you don't work with such things daily, don't mess with this pcb
while it's plugged in. Only the "isolated" section marked in
the photo isn't live-hot-dangerous. Note that over 1min is required
for the large 560uF capacitor to discharge down from 170VDC.

For non-chiller units, remove the four side screws, and the entire housing should come right off. If not, turn it upside down and note the rusty crust that's gluing the metal frame to the plastic! Slide a knife, or pry to break the bond. Then invert again, lift the housing carefully and reach inside to remove the ribbon cable from the PCB (pinch the plastic connector sides to release.) There are also two ground cables and the release-string to remove. (Also note, if you want to remove the main PCB, must unscrew the four screws on the BOTTOM of the centrifuge, not the four screws on the PCB itself.

I've now seen more than one Eppendorf lab centrifuge with the same problem: sudden death, as if a fuse has blown, but the fuses are fine. The LCD remains blank. Simple cure: increase a 2.2uF capacitor value to 10uF or higher (but stay below 100uF.) This fixes an apparent problem of slow-aging of the "boot-circuit" in the main switching supply.

SEE ALSO: photos on similar repair, 5417C Eppendorf Centrifuge

DETAILS

The problem turns out to be ...possible bad design! Internal startup voltages are borderline, and any small change will keep it from waking up when power is first applied. If I unplug the LCD ribbon cable before turning on the main power, the switcher starts up OK, and the main 5VDC supply voltage appears. The 5VDC supply for the ribbon conn. is found on that isolated LM7805 regulator, the one near the rightmost edge of the pcb in the photo below. Turn on AC power with ribbon plugged in and it fails, only rising to 2.4V, and the UC2844 supply-control chip goes into constant restarting. But, but ...the LCD board is only a ~60mA load at 5V. A slight increase in a 1/3watt load shouldn't bring a power supply to it's knees!

Next I look at the circuitry around the UC2844 (which is live, lethal. Floating-at-170VDC-plus-120v). I use a (electrically floating, dangerous!) scope to monitor the 16-volt input to the Vcc pin of the UC2844 controller chip. (This Vcc is also found at the input pin of that other, non-isolated LM7805 regulator in the center of the board. The regulator tab is also the chip's common.) A 2nd scope channel monitors the voltage generated by the high-freq transformer, found at the cathode of a SMT diode next to the 220uF cap. During each restart, the 2.2uF cap must rise to 16V, then the chip turns on, and the little VOGT 50KHz transformer should take over, charging up the adjacent, larger 220uF cap, and power everything. On the scope, the 2.2uF does rise to 16V, turning on the UC2844. That voltage now decreases rapidly, since the 2.2uF is acting as the entire power supply. The 50KHz switcher runs for a moment, and it's output voltage (across the 220uF) rises rapidly. But before it has time to rise to 10V, the voltage on the 2.2uF has fallen below 10V, and the UC2844 shuts down. Then repeat the whole sequence. But, it just missed normal startup by about half a volt! One cap voltage falls, the other rises, and they must pass each other in order to turn on the supply. There's not enough millijoules in the 2.2uF to last for the whole millisecs needed to boot up the switching supply.

So, maybe there's a shorted turn in that VOGT high-freq transformer block. Or maybe the LCD front panel is now drawing a few mA extra during the rising of the isolated 5V supply. Or maybe it was always supposed to be a 22uF capacitor, not 2.2uF? The voltages and RC periods are very close to the edge of failure, and any little change can push it over the threshold. Could I try replacing the little brown VOGT high-freq transformer? Nope, not available. So, kludge it: swap the 2.2uF capacitor with 10uF 25V tantalum. Yep, that did it. Starts up now just fine. (Hmm, in hindsight I probably should have used a 22uF or 47uF rather than a 10uF, just to keep the same failure from happening again in a couple of years!) ANOTHER TRICK: stand up the capacitor on 1cm leads, then to keep it cool, bend it away from the hot TO-220.

SEE ALSO: