After the dust settled, our best estimate on paper is 40% rather than 25-30%.
The reason for the adjustments were roughly:
[x2] I estimated exhaust temperature at 130 degrees, but it’s more like 100 degrees if the indoor air is 70.
[x1/2] I thought that all depressurization was compensated for by increased infiltration. But probably half of depressurization is offset by reduced exfiltration instead (see here)
[x3/2] I only considered sensible heat. But actually humidity is a huge deal, because the exhaust is heated but not humidified (see here)
With 1-hose the indoor temp was 68 vs 88 outside, while with 2-hose the indoor temp was 66 vs 88 outside (using the same amount of energy).
We both agree that 10% is an underestimate for the efficiency loss (e.g. due to room insulation, other cooling in the building, and the improvised 2-hose setup).
I don’t think we have a plausible way to extract a corrected estimate.
After this comment there was a long thread about AC efficiency.
Summarizing:
I said: “In practice I think the actual efficiency loss relative to a 2-hose unit is more like 25-30%” (For cooling from 85 to 70.)
John said that this was ridiculous.
After the dust settled, our best estimate on paper is 40% rather than 25-30%.
The reason for the adjustments were roughly:
[x2] I estimated exhaust temperature at 130 degrees, but it’s more like 100 degrees if the indoor air is 70.
[x1/2] I thought that all depressurization was compensated for by increased infiltration. But probably half of depressurization is offset by reduced exfiltration instead (see here)
[x3/2] I only considered sensible heat. But actually humidity is a huge deal, because the exhaust is heated but not humidified (see here)
John also attempted to measure the loss empirically, but I’d summarize as “too hard to measure”:
With 1-hose the indoor temp was 68 vs 88 outside, while with 2-hose the indoor temp was 66 vs 88 outside (using the same amount of energy).
We both agree that 10% is an underestimate for the efficiency loss (e.g. due to room insulation, other cooling in the building, and the improvised 2-hose setup).
I don’t think we have a plausible way to extract a corrected estimate.
I endorse this summary.