Three site conditions and one missing name on a maintenance schedule account for most early instrument failures here. What humidity, dust and voltage events actually do, and the short list of tasks that prevents each.
The pattern across service calls is consistent, and it is not a manufacturing pattern. Instruments in Bangladesh rarely fail because they were badly made. They fail because of three site conditions and one missing name on a maintenance schedule, and all four are cheaper to deal with than the repairs they cause.
The harm comes less from high humidity than from moving between conditions. A laboratory air conditioned during the working day and switched off at night and over the weekend spends every Sunday morning bringing warm, wet air into contact with surfaces that are still cool. Water then condenses where nobody can see it: inside a housing, on a board, on the underside of an optic.
Prevention: run the air conditioning continuously in rooms holding sensitive instruments rather than switching it off overnight, because steady and warm is kinder than cool and cycling. Keep optics and spare lenses in a dry cabinet or a sealed box with silica gel, and change the gel on a date rather than when somebody happens to notice it has changed colour. Where the manufacturer offers a dry purge for a spectrometer or an analyser, use it. And do not leave an instrument switched off for weeks during the monsoon, because many instruments survive partly on their own internal heat keeping them above the dew point.
Brick dust, road dust and construction dust are finer and more abrasive than the dust most equipment is designed around, and the dry season delivers months of it without interruption.
Prevention: a filter cleaning schedule with a name and a date against it, at an interval set by looking at your own filters rather than by copying a manual written for a European laboratory. Seal cable entries and floor penetrations. Wet mop rather than sweep. Do not put a ceiling fan directly above an instrument bench. When a site next door goes under construction, treat it as a temporary change of conditions and increase the cleaning frequency until the work finishes.
The events that matter are deep sags during load shedding and at changeover, transients as the generator picks up or drops the load, sustained over and under voltage, and a neutral to earth voltage nobody has measured. Sensitive electronics seldom fail at the moment of the event. They fail weeks or months afterwards, which is exactly why the connection is missed and the instrument gets blamed instead of the supply.
Prevention: a stabiliser sized against the actual depth and duration of sags measured at your site, an uninterruptible supply with genuine isolation for anything containing a computer or a detector, and an earth resistance that has been measured with an instrument rather than assumed from a drawing. Have the earth measured again after any building work, because that is when it changes and nobody rechecks.
Every item below takes minutes. Every one of them, left undone, ends instruments.
Write the list, put a name and an interval against each line, and review it at a fixed point every month. A log with a name in it survives staff changes. Institutional memory does not.