In a power dip, a particle counter on UPS can stay powered and count nothing, because its vacuum pump is wired to the normal supply.
That is one of the quieter ways an environmental monitoring system loses the minutes QA needs most. When the AHUs stop, pressures between grades fall and may reverse and particle counts rise, until the generator picks up and the fans come back. The batch decision leans on the record of that stretch.
Putting the EMS server on UPS is the first step, and most sites have done it. The gaps we see sit further out along the chain:
- The vacuum source for the counters, central pump or one per counter, fed from a non-UPS board.
- Network switches and media converters between the sensors and the server, plugged into a local socket in a riser.
- Pressure transmitters powered from a field panel that drops with the AHU control panel.
- Clocks. If the acquisition units and the server do not share a time source, the order of events cannot be rebuilt afterwards.
A logger that buffers locally and forwards when the network returns covers a short link failure. It does nothing for a sensor with no power.
ALCOA+ asks for complete records. A gap in the middle of an excursion leaves the batch with no evidence in either direction.
When did someone last trace the supply of every item between the sensor and the historian?




