"Supply airflow ±5% of design" shows up in plenty of AHU specifications, and as written nobody can prove it or disprove it.
On one system, a fan inlet reading from the manufacturer's pressure taps, a Pitot traverse in the main duct and a flow hood at the terminal filters can disagree by more than 5%. Duct leakage sits between the first two. The instruments add their own error, and a flow hood or anemometer grid can use up a real share of a ±5% band on its own.
Filter condition moves the number again. A unit tested with clean filters delivers more air than it will in two years. If the tolerance is meant to hold at the design final pressure drop, the FAT has to simulate that drop, usually by throttling, and the spec has to ask for it.
Leave all of that out and the vendor measures where the unit reads best, usually at the fan. The site team measures at the room. Both readings are honest, they do not match, and the argument lands at commissioning, when the program has least room for it.
We would rather write the number with its test: measurement location, method, instrument class and filter condition in the same clause as the ±5%. AeroSano units go through a witnessed FAT against the approved specification, and with a spec written this way the factory and the site measure the same thing.




