The Space That Felt Cold at the Right Temperature

Why this matters

Three visits, three techs, three tickets that say the equipment is operating correctly, and a customer who is still cold. Nobody was careless and nobody was wrong. The complaint was about a position in a room and every record ever made was about the room, so each visit re-validated the equipment and none of them ever measured where the person actually sat. This is a reconstruction of that job from the paperwork, and the gaps in the paperwork turn out to be the most useful part of it.

First, the space heater under the desk

Before any diagnosis, deal with what the occupant has already done about it, because that is the live hazard on this call.

A portable electric heater was running under the desk, plugged into a power strip, with its cord run under a chair mat and its discharge about six inches from a fabric-covered partition panel. Three separate problems, each with its own action.

Unplug it and take it out of service before you do anything else. Portable heaters draw near the limit of a general-purpose branch circuit and belong plugged directly into a permanent receptacle, never into a power strip, extension cord or multi-outlet adapter, which are not built for a continuous load of that size and overheat at the connection.

A cord run under a mat or carpet is a cord that cannot shed heat and cannot be inspected. Route it in the open or do not use the appliance.

Clearance to combustibles on a portable heater is a listed value in its own instructions, and a fabric panel six inches off the discharge is a combustible in the airflow path. Tell the occupant, in plain terms, what specifically was wrong with the arrangement rather than just taking it away, or an identical one appears next week.

Then check the circuit it was on. A heater run for weeks on an undersized path leaves evidence at the receptacle: discoloration, a warm faceplate, a loose connection. Looking inside that receptacle means opening an enclosure with live parts, which is energized electrical work under 29 CFR 1910.333(b)(2), so establish an electrically safe working condition and prove the tester live-dead-live per NFPA 70E-2021, 120.5 before touching anything inside.

What three tickets actually recorded

Pulled from the history, stripped to what each one contains.

Every number in that table is defensible. The equipment made supply air in the mid to high 90s degrees F on all three visits, the thermostat held within a degree of setpoint every time, and the airflow was checked once and adjusted once. A fourth visit doing the same things would produce a fourth row that looks the same.

What the tickets did not record

Four fields, none of them exotic, and their absence is what let the same visit happen three times.

Which position in the suite. "Customer states cold" is written about the suite. The suite has a wall of glass on one side and interior partitions on the other, and those are not the same thermal environment. Nobody wrote down which desk.

The time of day. Recorded on the dispatch record but never carried into the technician's findings, and never compared across visits.

The outdoor condition. Not the forecast, the actual outdoor temperature and sky condition at the time of the visit. A glass-heavy space behaves completely differently on a clear cold morning than on an overcast mild one.

Any measurement at the occupant's position. All three visits measured at the thermostat and at the equipment. That is the correct place to validate equipment and the wrong place to evaluate a comfort complaint.

The pattern in those four is worth naming: each one is a fact about where and when the complaint lives. The ticket template captured facts about the machine. So the record could confirm the machine indefinitely and could never confirm or deny anything the customer said.

What the dispatch times gave up

The dispatch timestamps were still in the system even though nothing about them made it into the findings. All three original complaint calls were logged between 8:00 and 10:30 in the morning, between November and January.

The second visit was dispatched for a 2:00 pm arrival. The third was late morning. So on two of three visits the tech arrived outside the window the complaint lives in, and on the third the condition had had two hours of occupied heating to soften.

That single reconstruction, from data nobody had thought of as diagnostic, reframed the whole thing: this is a cold-morning, heating-season, one-side-of-the-building complaint, and it had been investigated on afternoons.

The measurements nobody had taken

Return visit, 8:15 am, outdoor 22 degrees F, clear sky. Taken at the complaining desk, which is 3 ft from the glass.

• Air temperature at seated height: 71 degrees F. Same as every prior visit.
• Interior glass surface: 44 degrees F, about 40 sq ft of it in view.
• Exterior wall around the glass: 58 degrees F, about 90 sq ft.
• All remaining surfaces in view, interior partitions, ceiling, floor: about 70 degrees F, roughly 430 sq ft.
• Air speed at ankle height: 55 ft per minute, against 15 ft per minute at the interior desk.

The arithmetic at the position

Using the area-weighted mean radiant temperature method and the operative-temperature relationship, both of which are derived in the sibling article on air versus radiant temperature:

MRT at the glass desk = (40 x 44 + 90 x 58 + 430 x 70) / 560 = (1,760 + 5,220 + 30,100) / 560 = 37,080 / 560 = 66.2 degrees F.

Operative = (71 + 66.2) / 2 = 68.6 degrees F.

At the interior desk, every surface in view is near 70 degrees F, so MRT is about 70 and operative is about (71 + 70) / 2 = 70.5 degrees F.

Two desks, one air temperature, and about 1.9 degrees F of difference in what a person experiences. That is real and it is worth having, but on its own it is not a dramatic number, and this is where the honest part of the diagnosis starts.

The 55 ft per minute at ankle height is the other half. Air against 44 degrees F glass cools, gets denser, falls, and spreads across the floor toward the desk. That downdraft is a convective effect that the operative-temperature average does not carry, and it lands on ankles and lower legs. Alongside it sits the asymmetry: a large cold surface on one side of a person and neutral surfaces on the other is a discomfort factor in its own right, which ASHRAE Standard 55 treats separately from operative temperature with its own limits depending on which surface is cold.

So the finding is a stack of three, all originating at the same 40 sq ft of glass: a 1.9 degree F operative deficit against the other desk, a measurable floor-level downdraft, and radiant asymmetry across the occupant's body.

Why raising the setpoint had been failing

Visit three raised the setpoint from 71 to 73 degrees F. Follow it through the same arithmetic, treating the surrounding interior surfaces as roughly following the air for a first pass and noting that the glass barely moves at all, because its inside surface temperature is set mostly by the outdoor condition and the glazing.

At the glass desk with air at 73: operative is about (73 + 66.2) / 2 = 69.6 degrees F, up 1.0 from 68.6.

At the interior desk with air at 73: operative is about (73 + 70) / 2 = 71.5 degrees F, up 1.0 from 70.5.

Both desks gained about the same amount, so the gap between them did not close, the person by the glass got about half of what the thermostat number suggests they got, and the person at the interior desk got warmer than they wanted. And nothing about the downdraft changed, because the glass surface is still at 44.

That is the mechanical reason three visits of doing more of the same could not work: every lever anybody pulled acts on the air term, and the deficit is in the other term.

Weighing the options against the numbers

Four candidates, each sized against the measurements rather than against intuition.

Move the workstation off the glass. Puts the occupant in the interior desk's environment, about 70.5 degrees F operative, out of the downdraft path and out of the asymmetry. Costs an afternoon of somebody's time and no equipment work at all. It is the cheapest correct answer and it is the one nobody had proposed because it is not a repair.

Treat the glass. An interior insulating panel or film that brings the inside surface from 44 to about 58 degrees F changes MRT to (40 x 58 + 90 x 58 + 430 x 70) / 560 = 37,640 / 560 = 67.2, and operative to (71 + 67.2) / 2 = 69.1, a gain of just 0.5 degrees F over the original 68.6. Both figures come from the same surface set and the same weighting, so they compare directly. Judged only on operative temperature this looks like a poor return, and that judgment would be wrong: raising the glass surface by 14 degrees F is what removes the downdraft and most of the asymmetry, which is the part the occupant actually reports.

A local radiant source aimed at the position. Adds a warm surface into the occupant's view, working directly on the term that is deficient. Solves it for one position and does nothing for anyone else, which is fine if the complaint is one position.

Raise the setpoint. Already tried, already quantified above, and it moves the wrong term while overheating the other side of the suite.

The recommendation went out as the first two together, with the honest framing that the glass treatment buys comfort mostly through the downdraft and the asymmetry rather than through the half degree of operative temperature it adds.

The confirmation visit, and the field they added

Return on a comparable morning: outdoor within a few degrees of 22 degrees F, clear, same 8:15 arrival, same instrument, same seated position. Comparable conditions are not a nicety here; the entire complaint is condition-dependent, and a mild overcast morning would have shown improvement whether or not anything was fixed.

Three things to confirm, each predicted in advance so the visit can fail honestly: glass inside surface at or near 58 degrees F, ankle-height air speed down near the interior desk's 15 ft per minute, and the occupant's own report at the position. If the glass hits 58 and the air speed does not fall, the downdraft has another source and the diagnosis was incomplete.

Then the change that keeps this from recurring on the next building: four fields added to the comfort-complaint ticket, which are exactly the four the old record was missing. Position in the space. Time of day the complaint occurs, taken from the customer, not from when the truck arrived. Outdoor temperature and sky condition at the time of the reading. And at least one measurement set taken at the occupant's position rather than at the thermostat.

The first tech on this job had every skill needed to solve it in one visit. What they did not have was a ticket that asked where the customer was sitting.

References

• ASHRAE Standard 55, Thermal Environmental Conditions for Human Occupancy, for operative temperature and radiant asymmetry limits
• Manufacturer instructions and equipment labeling for portable heater clearance to combustibles and direct-receptacle connection requirements
• 29 CFR 1910.333(b)(2) for energized electrical work; NFPA 70E-2021, 120.5 for live-dead-live proving
• See related: How Air Temperature and Radiant Temperature Differ; The Stratification That Fools a Single Reading