Filtration Upgrade Assessment Before a Media Change

Purpose

A customer asks for a better filter, a tech fits the highest MERV on the truck, and six weeks later the same house has a frozen coil, a tripping limit, or a compressor that failed on liquid return. The filter did not fail. The blower had no pressure left to give and nobody measured before deciding.

This procedure produces one number before any media is bought: the pressure drop this system can afford to spend on a filter. Everything after that is arithmetic. It also produces the answer nobody wants to hear often enough, which is that some systems have no budget at all and the honest finding is duct, not filtration.

Scope

Covers the pre-change assessment on residential and small light commercial forced-air systems: nameplate rating, test port installation, static measurement with and without the existing filter, the budget calculation, media selection against a manufacturer table, and post-change verification.

Does not cover routine like-for-like filter replacement at a maintenance visit, which is a line on the annual maintenance SOP. Does not cover electronic air cleaners, UV devices or any powered device in the airstream, which have their own installation and verification SOP. Does not cover duct renovation beyond identifying that it is needed and writing the reason.

Roles and responsibilities

Procedure

1. Read the blower's rated maximum external static off the equipment nameplate and write it down. Note the nominal tonnage or rated cfm at the same time. Acceptance: a nameplate figure in inches of water column, in writing, with the model it came from. What wrong looks like: a remembered figure, since ratings differ between a fixed-speed and a variable-speed air handler of the same tonnage. Stop rule: no legible nameplate means the manufacturer's spec sheet gets pulled before measuring, because every number after this is compared against it. Hazard: none at this step, a nameplate read with the panel closed.

2. Isolate power, then install static pressure test ports where they belong. Ports go between the filter and the blower on the return side and downstream of the coil on the supply side, so the reading brackets everything the blower has to push through. Acceptance: two ports, plugged, in those two locations. What wrong looks like: a supply port upstream of the coil, which leaves the coil out of the measurement and flatters the result. Stop rule: no port location that satisfies the bracket means the reading is not taken at all rather than taken wrong. Hazard: this is a drilling step, so prove the equipment dead per NFPA 70E-2021, 120.5, with work practices at 29 CFR 1910.333(b)(2), isolate the blower under 29 CFR 1910.147, and confirm what is behind the sheet metal before the bit turns, because a drill through a refrigerant line, a flue or a wiring harness is the same mistake three different ways.

3. Restore power and measure total external static with the existing filter still in place. Run the system in its highest airflow mode and let it stabilize. Acceptance: a total external static figure, taken with all panels on and every register and grille in its normal position. What wrong looks like: a reading taken with the blower door off, which unloads the return and reads low. Stop rule: an unstable needle means wait, not average by eye. Hazard: this is the step that puts energy back in, so covers go on first and the disconnect is closed from the hinge side rather than square in front, with the customer and pets out of the room.

4. Measure the pressure drop across the existing filter by itself. Take a reading either side of the media in its own rack or grille. Acceptance: a drop in inches of water column with the filter's nominal size, depth and MERV rating recorded beside it, MERV being the ASHRAE 52.2 rating class. What wrong looks like: a drop taken across the whole return grille assembly, which includes grille loss and overstates the media. Stop rule: no usable reading either side of the media means the drop is not estimated, it is measured after the rack is modified. Hazard: system running, so probes go through the ports and hands stay out of the filter track.

5. Pull the filter and measure total external static again, without one fitted. Bag the old filter at the rack rather than carrying it through the house. Acceptance: a measured filterless total external static figure. Do not calculate this by subtracting the filter drop from step 3: removing the filter raises airflow, which raises every other component's drop, so the subtraction gives a number the system never sees. What wrong looks like: a computed figure with no reading behind it. Stop rule: measure it or the assessment does not proceed. Hazard, and it is on the pull rather than the access: a loaded filter releases its collected dust when it is flexed or knocked, so it is drawn straight into a bag at the rack, and where the media is visibly fouled with growth or the space is small the control is a respirator under a written program per 29 CFR 1910.134, not a comfort mask.

6. Compute the filter pressure budget and hold half of it in reserve for loading. Budget equals the nameplate rating from step 1 minus the measured filterless static from step 5. Permitted clean drop is half of that budget, and the budget top becomes the change-out trigger. Acceptance: a budget and a permitted clean drop, both in inches of water column, written on the ticket. What wrong looks like: spending the whole budget on a clean filter, which puts the system over rated static the week it starts to load. Stop rule: a budget at or below zero means this system has no room for any filter and the finding is return-air restriction; no MERV upgrade is quoted. Hazard: none at this step, arithmetic done at the truck.

7. Choose media against the manufacturer's pressure drop table at your actual face velocity. Face velocity is measured airflow divided by total filter face area in square feet. Acceptance: a candidate media whose table clean drop at that face velocity is at or below the permitted clean drop from step 6, with the table value written down as read. What wrong looks like: picking by MERV number alone, since a deeper media at the same MERV can run a fraction of the drop of a 1 in panel. Stop rule: if nothing at the customer's requested MERV fits the budget, the honest answer is more face area first, and that gets quoted rather than the filter. Hazard: none at this step, a table read at the truck.

8. Fit the media, re-measure, and set the change-out trigger in the record. Acceptance: measured total external static at or below the nameplate rating, measured filter drop within 0.02 in w.c. of the table value it was selected on, and airflow back in the 350 to 400 cfm per ton band on a cooling system. What wrong looks like: a post-change static that clears the nameplate but an airflow that does not, which means the budget was right and the duct was already marginal. Stop rule: over nameplate after fitting, the new media comes back out and the old size goes in until the return work is done. Hazard: running system with panels on, probes through the ports only.

When the site does not match

Filter grille in a ceiling with no rack depth: the only media that fits is what the grille takes, so the budget calculation still runs and the answer may be a second return rather than a filter. Customer already bought high MERV panels: measure first anyway, and if the budget will not carry them, say so with the two static numbers in hand rather than as an opinion. No safe port location on the supply side: report the assessment as incomplete rather than issuing a partial figure, because a supply reading upstream of the coil is not a smaller measurement, it is a different one.

The record this produces

The assessment sheet carries: nameplate maximum external static with the model it came from; port locations; total external static with the existing filter; the existing filter's size, depth, MERV and measured drop; measured filterless total external static, marked as measured rather than derived; the computed budget and the permitted clean drop; the media selected with the table drop and the face velocity it was read at; and the post-change static, filter drop and cfm per ton.

Two people read it. The tech at the next maintenance visit, who has a change-out trigger in inches of water column instead of a guess about months. And the estimator, if the budget came out negative, because a return-air proposal built on two measured numbers survives a customer asking why.

Worked pass: three-ton air handler, customer asking for MERV 13

Existing filter is a 20 by 25 by 1 in MERV 8 in a return grille, last changed at an unknown date.

Step 1. Nameplate maximum external static reads 0.50 in w.c. Recorded with the model number.

Step 2. Ports set between the filter and the blower, and downstream of the coil. Power proved dead before drilling, blower isolated, cavity checked behind the panel.

Step 3. Total external static with the existing filter in place: 0.86 in w.c.

Step 4. Drop across the filter alone: 0.31 in w.c. Filter recorded as 20 by 25 by 1, MERV 8, heavily loaded.

Step 5. Filter pulled and bagged at the grille. Measured filterless total external static: 0.62 in w.c. Note that 0.86 minus 0.31 would have given 0.55, and the real reading is 0.62, because airflow rose when the filter came out. The measured number is the one used.

Step 6. Budget is 0.50 minus 0.62, which is negative 0.12 in w.c. Step 6 fails and takes its stop rule. No filter of any rating fits this system as it stands. No MERV upgrade quoted. Finding written as return-air restriction, with the two static figures beside it.

Return work performed on a later visit: one additional return grille and drop added. Re-measured filterless total external static: 0.34 in w.c. Budget is now 0.50 minus 0.34, or 0.16 in w.c., and the permitted clean drop is half of that, 0.08 in w.c. The change-out trigger is the budget top, 0.16 in w.c. measured across the filter.

Step 7. Combined filter face area is now a 20 by 25 plus a 16 by 25, which is 500 plus 400, or 900 square inches; 900 divided by 144 is 6.25 square feet. Measured airflow 1,150 cfm; 1,150 divided by 6.25 is 184 fpm face velocity. The manufacturer's table for a 4 in deep MERV 13 media reads 0.07 in w.c. clean at that face velocity, which is at or below the 0.08 permitted. Selected.

Step 8. Media fitted. Measured total external static 0.42 in w.c., under the 0.50 nameplate. Measured filter drop 0.075 in w.c. against the table's 0.07, a difference of 0.005, inside the 0.02 gate. Airflow re-measured at 1,120 cfm; 1,120 divided by 3 tons is 373 cfm per ton, inside the 350 to 400 band. Pass, and 0.16 in w.c. goes into the maintenance record as the change-out trigger.

References

• Equipment nameplate and the manufacturer's blower performance data for the rated maximum external static and the tap-to-cfm relationship
• ASHRAE 52.2, in the edition your media supplier rates against, for what a MERV class means and how it is measured
• Filter manufacturer's pressure drop table for the specific media and depth, read at your measured face velocity rather than at a nominal one
• 29 CFR 1910.333(b)(2) with NFPA 70E-2021, 120.5 for live-dead-live before drilling; 29 CFR 1910.147 for the blower; 29 CFR 1910.134 for respiratory protection under a written program when pulling loaded media
• See related: the UV and air cleaner installation SOP for powered devices in the airstream, and the annual maintenance SOP for like-for-like changes