It is well known that ACH50 (what a blower door measures) is a VERY inaccurate measure of what you really care about, ACH-nat (perhaps ACH-normal would be a better term). So how about using the air conditioner to measure ACH-normal? Ie, put the dog, people and plants outside and get a steady state pints/hour measurement of the moisture being removed. Compare to the indoor and outdoor dew points and calculate an ACH value.

Air handler driven infiltration would exaggerate the ACHnatural number with that approach, and I'd think it would be difficult to accurately quantify all interior moisture sources to boot.

But coming up with a more relevant way to measure air infiltration IS one of those thorny problems worth trying to figure out.

Air handler driven infiltration would exaggerate the ACHnatural number with that approach,

I consider that a significant advantage - lots of people aren't aware of or measuring this. That's why ACH-normal might be a better term.

I suppose one could tape off the toilets, but I doubt that this is significant.

I doubt an "actual" measurement would be of much use. Bear in mind that ACH50 is just a standard measurement of general tightness of a house under artificially high depressurization so as to get away from the natural factors that determine normal actual leakage. One can divide by a factor of around 18, depending on the nature of the house and its location, to get a very crude approximation to what the house MAY leak under worst conditions, when it's bitter cold and windy outside, for design purposes. When it's fairly close to room temperature outside with no wind, the house doesn't leak appreciably at all. In between it's a roll of the dice, and you never really know what the leakage rate is. So a measurement done at arbitrary outside conditions can't mean an awful lot, I would think.

So a measurement done at arbitrary outside conditions can't mean an awful lot,

Perhaps, unless you can measure the variable factors and account for them. Which in the case of humidity, wind and delta-T, we generally can. But I agree, if you only want to measure air sealing (vs infiltration, the thing that costs you energy $), ACH50 is better than a "tracer gas" type of measurement.

Blower door numbers don't tell you where the leaks are, and the location of the leaks really DO matter. If it's just a single hole the natural infiltration is two orders of magnitude lower than if they're equally distributed, and three orders of magnitude lower than if they're split equally between the attic floor and the basement walls of a 3 story house. The science on this to date is really pretty crummy, despite the best attempts to really model & measure it.

Dividing ACH/50 by 18 is no better than a WAG, with truly gia-normous error bars on what the average or peak leakage will ultimately be. But when you get the cfm/50 numbers under 1500cfm/50 it begins to lose importance, and under 250cfm/50 (a PassiveHouse type of tightness) it's pretty much a "who really cares?" kind of deal.

Picking a cold day and purposely creating significant infiltration (ie, fan blowing outward) and then looking at the interior with a infrared imager makes sense to me. Then do the reverse (blow outward and look at the exterior) to get a different view. No doubt that both of these would be best done before the final layer goes on (ie, siding, wallboard, maybe even cellulose). I haven't thought about using smoke.

I did the reverse of this, with a humidifier in the winter time. Ran the thing at 3 gal/day, and estimated how many RH points I got running versus not running, about +10%. The figure I got, ~0.3 ACHnat, was about 50% lower than a blower door scaled by the whammy factor, which is 0.47 ACHnat (predicted). S'alright. SInce my attic is well sealed, and my sill plate is not at all, maybe my stack driven ACH_nat is less than expected from ACH50?? Or maybe its just errors.

Excellent idea, especially if you account for external air humidity and any interior moisture sources. And take enough data points that other variables (like wind) average out to a normal/seasonal value.

Posted By woodgeek68 on 03 May 2013 11:31 AM
I did the reverse of this, with a humidifier in the winter time. Ran the thing at 3 gal/day, and estimated how many RH points I got running versus not running, about +15%. The figure I got, ~0.3 ACHnat, was about 50% lower than a blower door scaled by the whammy factor, which is 0.47 ACHnat (predicted). S'alright. SInce my attic is well sealed, and my sill plate is not at all, maybe my stack driven ACH_nat is less than expected from ACH50?? Or maybe its just errors.

RH is a relative number (relative to saturation at a particular air temp.) To run an infiltration calc based on the humdifier volume with any accuracy you need to track both the indoor & outdoor dew points, as well as estimating the volume of other indoor humidity sources like showers, sweaty humans, drooly dogs, etc..  The infiltration air is not bone-dry- it has substantial moisture content, and the humidifier is just supplying the "make up" that the other indoor sources didn't cover.

I hear you Dana....not like a tracer expt with some inert species, but a great/easy way to get a ballpark.

To be more precise, I picked a period in the winter when the outdoor humidity was low, when there wasn't a lot of windy days in the forecast, and tracked my baseline RH, and got something like 19±2%, over the course of several days (w/ showers as usual). I then put the hum on a setting that ran continuous, and did 3 gal/day, and the new baseline was closer to 28±2% over the course of several days. Then I shut the hum down, fell back to 19±2 over a few days, and then measured the bounce a second time. IOW, an 'ABABA' expt design.

I can solve for the psychrometrics, and work out what volume of air at 19% RH and 70°F would absorb 3 gals and only go to 28%, and I say that is my exfiltrated air volume per day. I have forced air heat, so its all well mixed.  Indoor air is thermostated, I put the hum on the lower level, and measured the RH on the other end of the house, higher level.

If my RH was jumping all around day to day I wouldn't have bothered....but in both conditions, the RH 'noise' seems quite low in the winter. And the things that do contribute: cold snap, wind storm, can be discounted if you watch the weather.

You can get fairly fine-grained outdoor dewpoint tracking from Weatherspark.com datasets. But "outdoor humidity was low" has pretty big error bars on how much infiltration dilution you have going- well into double digit percentages. Outdoor moisture levels are never zero humidity, or even close, in populated areas of N. America.

Even 10F dew point air is ~0.0013% water by weight, more than 1/4 the moisture content of healthy-comfortable 40F dew point air's 0.005%. When outdoor dew points are under -30F, the error would be in the statistical noise, but that's "cold snap in Yukon" type air.

I am just claiming that combined effect of outside air RH and water emitted within the house added up to a stable RH baseline that could be subtracted off, not simply neglected.

The error bars ARE huge, I would say ±30% or so. But I found this metric useful when I was progressively sealing my very leaky house. Same amount of humidity added 5 years earlier resulting in no detectable bounce in RH at the other end of the house, which prob means <2% increase in RH. After sealing open cavity tops, plumbing chase, masonry chimney chase and ~30 tubes of silicone in top plates and drywall joints to the attached garage, I got a repeatable 10% RH bump.

And the number I inferred is the within 40% of the ACH_nat inferred from the CFM50.

I have to come back to this issue. Sometimes you want to do a calculation of moisture transfer. So you take ACH@50 pascals, apply some factor (say 18) and get an ACH natural. But the 18 is based on a seasonal average and includes large periods when there is little stack effect or wind. I argue that peak stack effects (hot summer days and cold winter days) are hugely different than the seasonal average. Not at all sure that I did this part right, but a cold winter day might have 5x more air flow than the seasonal average and hot summer day might be 3x. On top of this, you have wind differences that might cause 10x air flow differences. So how does one take ACH@50 and for a hot, humid, windy day, get ACH_summer_peak so that peak latent load due to infiltration can be calculated? Probably a rough estimate using a virtual orifice size that is equivalent to all the leaks and then using it with more accurate pressure differentials (like 1.5 pascals per floor on a hot but windless summer day).

As Lstiburek says about blower door results for ACH, "If you don’t know both the distribution of holes and the pressure differences across them, you don’t know jack." And that makes heating/cooling load calculations subject to some large errors.

Just had a blower door test done on our cottage/eventual retirement house, which I designed and built, it's a bungalow w/walkout basement, 2hrs NE of Toronto. We wanted it to be very well insulated and sealed for leaks since we will not be living there on a permanent basis for a few years. The test results were 11sq in leakage,  207cfm@50pa with a house volume of 20867 sq ft , they were very impressed with the results. This was my first and last house build! Here is a link to my building blog http://kasshabog.blogspot.ca/

Good job! At 207cfm/50 you don't really much CARE what the distribution of those air leaks are.

That's the way to take the guesswork out of it- built it super-tight, then install the ventilation where it's needed and most appropriate!

So ~.6 ACH which is the passive house standard. Excellent, unless you do something like put a HRV supply without return into a closed door room and make the ACH@normal (for that room) the same as a 2.5 ACH@50 house.

What techniques did you use (I see lots of red tape)?

Thanks, Jonr and Dana1. I guess you reviewed my blog Jonr. My intent when I set out was to build something that would be very energy eff., to keep costs of heating low for years to come, for the last year I spent $500 on propane ($.70/L) for heating, BBQ and standby generator. I used Silver Fox ICF from bottom to top, and Fox Bucks around all window and door openings. All windows and doors were then sealed with Great Stuff Foam, and  after I installed my jambs I went with the foam again. As we stacked the Fox blocks they were all glued/foamed together, the RED TAPE (Tuck) was my idea, just thought it might seal any cracks and stop air flow, it only cost $100, and can't go back and do it after the drywall was on. The windows are triple glazed (180,clear180)  Strausberger windows they seem to be preforming very well. I have very few penetrations to the outside, 2 taps, dryer vent , HRV and furnace vents. The ceiling has R12 (which I started with just to keep the heat in while I was working) and then I blew in R50 cellulose, after the ceiling drywall was up, for a total of  R62. I used the plastic ICF elec. boxes for all outlets and foamed them in place along with foaming everywhere I removed foam to run wires.  I just read a lot on this forum and did my research before I started, and am extremely happy how it turned out.

Stuart

One benefit of the ICF construction is that, like SIPS, their design is fairly easy to air seal. How did you seal off the attic?

Congratulations on your results & enjoy the benefits!

Bob, my attic entry way is in the gable end, inside our sunroom, outside of the building envelope. Had quite a conversation with our CBO about doing this, but he could give no reason against it.