This is a hydraulic design study considering before and after conditions of HVAC ducts in
the crawl space overhaul reported here:
The home is a well-built farm house, 1590 sf, completed in 1973.
It was necessary to remove and dispose all crawl space HVAC ducts, including some of the elbows approaching floor boots. This permitted cinching heavy R25 cotton batts into needed contact with the floor sheathing.
Believe that the ducts demolition was needed for access to tie up the batts.
Know that once again I have done a ducts rebuild by my own commonsense rules, without first characterizing the performance of the found ducts. And, what does performance characterization mean? Characterization includes Diagnostic Static Pressure Measurements and a measurable sense of the equality of bulk air velocity at each floor register. I'm thinking the register equality is sensed by comparison of noise level at each register.
Removal of all found HVAC ducts was a needed first step. I found then that all supply distribution was with stupid leaky hacks of a single D-box directly under the furnace.
Five flow paths out of this D-box are visible in this photo. On the RHS (front), two 9" ducts. On the side, two 6" ducts and on the back, one 6" duct.
Find this Medusa distribution of conditioned air to be stupid and offensive, never to be taught to anyone. Air in motion must constantly be guided in least-resistance paths; else it must generate the net motion through gradients in wasteful and noisy, unstable turbulence.
The three 6" duct outlets each served one nearby 6" register. One 9" outlet served four 6" registers of southern rooms. One 9" outlet served an 8" register and three 6" registers in northern rooms. Assuning that the D-box outlets delivered equal bulk velocities, calculate register flows thus:
Assume equal pressure drops at the five paths in the supply D-box. This happens with equal path bulk velocity in the paths if they have identical geometric loss, say k = 1.8 in each path, an imperfect assumption. A direction having a majority of flow, to the East at 9" starters. will have higher drive pressure.
The sum of D-box exit areas a bit less than .7854 * ( ( 3 * 36 ) + (2 * 81) ) = 212 sq in
The three 6" registers each have total area 84.8 sq in.
The South-serving four 6" registers have area 113.1 sq in
The North-serving three 6" registers and one 8" register have total area 135.1 sq in.
The total area of pipes leading to registers is 333 sq in. Where pressure drops in various paths are in proportion to square of velocity, Q/A, the majority of fluid energy is dissipated at the D-box walls, mostly-regulating the flow distribution. The square of (212/333) is 0.4. The North-serving registers were favored by almost 135.1/113.1, 20%.
Accurate math here is impossible and futile.
Study further the sizing of the improved HVAC supply ducts:
The trunk area is 14 in x 19 in , 266 sq in.
The flow constriction through electric resistance heaters is to area area about 120 sq in, with extreme high velocity and unfortunate high pressure drop equally reducing all path flows.We should wish some day to have a ground-source heat pump, eliminating the supplemental heaters.
The above diagram of this crawl space warm air duct system is a 1:120 scale drawing, with overlay of structure details upon a Google satellite photo. Please note similarity to planning graphics and involved ducting elements for similar achievement in an attic.
The really curious may examine too, a full photo album A Deep Energy Retrofit, Fall of 2013 including found conditions of the failed UltraTouch batt installation by "twining" method, and of impassable and leaking conditions in the found system of steel warm air ducts. A separate blog post will address the cotton batt insulation repair. The involved home in Hillsboro, Oregon is also the site of a publicly-shared full photo album documenting attic floor weatherization with UltraTouch cotton batts: Real Sealing Fixes HPwES Sealing In An Oregon Attic. Full pdf photo albums always tell a story best, but please know they are poorly rendered when viewed in GoogleDocs. If you want to examine such albums, please bother to download them.
Elements of a new warm air duct system are in two nearly-equal branches, called North and South. The branches include a well-considered set of plenum components below the furnace. A 14x19 turning plenum replaces a crude D-box.
The furnace plena have a 1" insulation liner and end in a triangular head with two 12" starter collars serving a system of flexible ducts.
Two 6" registers near the furnace are served by the first of four lateral take-offs, circuit label T12x6. These and all steel couplings of flexible ducts, are hung from the overhead via eye bolts. These two, hanging low, are pulled up by wire loops. Never support steel couplings by strapping upon the flexible ducts. Don't rely on screws and tape for suspension via metal couplings.
Here note the 1" insulation lining, and turning vanes. Getting prepared with such well-conceived and well-built elements is a small step of installer professionalism. Expect you may need a couple of weeks lead time for fabrication by a superior sheet metal fabricator. My fabricator in Portland, Oregon is Vinje & Son.
Don't accept dissing of flexible ducts, that they have extraordinary friction. Where each duct section is fully supported by straps and by steel couplings, draw small tension in the liner, taking up coils as needed, to align the duct and coupling. High-resistance problems are always the result of careless alignment, with tripping over steel edges. Plan excess of insulation jackets, to draw them fully-overlapping, over couplings.
This is coupling component R8x6, used two places at branch ends, where 6" ducts to registers are of length not more than six feet. I wish for simpler one-piece reducers, but want ample starter lengths for liner take-up and secure taping. Here and in the previous photo, see that the eye bolt hanger of a steel fitting passes through a hole in plywood overhead. A wood dowel cut from nimble tree branch passes through the bolt eye to releasably hold everything up. The release can allow better fit of insulation jacket overlap.
Here is the completed end of the North branch at a 6" register. The hanger of the R8x6 is out if sight above the jacketed water pipe. See only tough black liner, strapping and some Nashua 557 tape. Leave no opening for mice. Wherever possible, draw ducting tight against plywood guides, as barrier to mice.
This is the new heart of the duct system under the furnace, with a turning plenum.
This is the well-strapped head of the North branch 12" duct, tightly in contact with overhead 1/2" plywood strips. One can readily crawl anywhere in this crawl space, sliding under the smooth plastic covering if necessary.
Please see that residential HVAC should be naturally balanced with good flow dynamics in orientation and more, of plena, take-offs and reducer or wye fittings. Air flow in thoughtless ducting can be extremely chaotic, and not amenable to tweaking. Do accounting math as in this example for the construction planning. With intelligent design, achieve equal register discharge velocity and noise level at the eleven floor registers.
The practical concerns in duct design are to allow overall free flow. Never think that a damper hidden and forgotten somewhere would allow residents to fix unbalanced, bad duct design. If someone finds too little heat in a favored room, just tackle the duct path to that room. Find an obstruction, perhaps a crushed or pinched duct, or misalignment in a coupling. Failing that, increase the size of ducting for that room, perhaps just in a more-efficient register. Look for missing insulation as the cause of discomfort.
Know that I am unaware of any other person sharing my practice of novel insights. I observe that taught common practice, persisting in new homes, is absurd and wasteful. Google HVAC Duct Design ACCA Manual D and easily find instruction in the wrong practice I have undone in this home. I engage in this practice as a rare person experienced in the federally-subsidized design of commercial nuclear power reactors, foolishly thinking to find value in my service in the engineering operation of nuclear submarines as a Regular Line Officer, then obtaining my first Masters Degree, in Niclear Engineering, at UC Berkeley in 1972. At now-all-gone Combustion Engineering in Windsor, CT, I had a senior PhD-level assignment as a sole Principal Investigator engaged in controlling the even distribution of water cooling nuclear power reactors, to enable higher power output. Not really a very helpful career and short-lived.
Here are example illustrations found in a wasted hour of looking at results of that Google Manual D search:
Do silly Manual D Worksheet entries upon duct friction lossesof energy, when geometric effects dominate. Pooh pooh the virtue of fitting chamfers; accept needless abrupt change of velocity.
Medusa foolishness!
I have a genuine need for a bedside reading light. But for the living room with its fancy ceiling, each room is lit with one or more Glimpse LEDs, mounted in new RACO175 junction boxes of course. An overhead light is crummy for seeking sleep, badly aimed at a book, and not amenable to drowsy shutoff.
