I use a Mettler AE240 electronic top loading analytical balance.
If you're just weighing small objects a few percent drift is nothing(i.e. measuring things only weighing less than a gram), but for analytical purpose, it means everything. I find that calibration drifts several tens of PPM after a few days.
i.e. tare 88.5526g
add water to determine non volatile component
gross: 125.0000g
allow to evaporate completely... Since I'm trying to quantify TDS from only 36g of sample, a few mg drift from balance would throw off the results significantly.
With an electronic top loader, there's an issue of corner load in addition to calibration drift, meaning that if the beaker is not placed at exactly the center, it might throw off a few digits too.
So, with this, I'm having to do this:
1.) Calibrate
2.) measure tare
3.) add sample
4.) record
5.) wait...
6.) calibrate again
7.) record and calculate
I understand that with a mechanical substitution balance like Mettler H 8, it's not as "convenient", but it's a hanger type and it does not suffer corner load effect and does not require calibration each and every time.
I couldn't locate the link right now, but I recall reading on a some university lab manual recommending using the mechanical one over an electronic one for work requiring more precision.