Embedded Files
Gradient developed a high-efficiency, easy-to-install heat pump for residential applications.
I led the re-design of several installation components, which provide a secure and weather-tight interface between the heat pump and a user's home.
My efforts helped reduce air leakage through the window by over 30%, led to increased user confidence during installation, and enabled re-use for multiple installations.
Under my ownership, the installation kit development schedule erased a 3-month delay and caught up to the rest of the product before launch in early 2025.
Two technicians installing a Gradient All-Weather heat pump
The Gradient "All-Weather" heat pump hangs over the sill of a single- or double-hung window. The boxy outdoor and indoor units are connected by a "bridge" through which fluid and power are routed. A hinge in the bridge enables installation.
The heat pump is installed by 2 people with minimal training and no certification. The installation procedures must be safe, intuitive, and robust.
A mounting bracket is clamped to the sill of an open window. The heat pump (140 lb) is attached to the mounting bracket, and the window is closed onto the heat pump. The heat pump may be installed in a range of window sizes (26" - 40" internal width), which creates the potential for large gaps on either side of the installed product. The gaps must be filled in a manner that is water-tight, air-tight, and rodent-proof.
When I joined Gradient, the window sealing solution was in a prototype stage. The method to fill lateral window gaps relied on several irreversible, subtractive operations with high risk of product damage and low performance weather sealing:
Break two plastic "taco shells" along a thin "snap line" such that they fill the lateral window gaps (< 1/4" gap remaining)
Use 4x M2.5 screws to attach the plastic shells onto the structural installation bracket
Tear pieces of semi-rigid foam along perforated lines such that it fills the plastic shells, filling any remaining gaps entirely
Slide the heat pump onto the structural installation bracket, aligning visually between the foam-filled lateral plastic shells
I was asked to:
Improve the water- and air-sealing performance
Increase the durability of the installed gap-filling solution
Simplify the installation experience
Accelerate the development schedule to catch up to the rest of the heat pump product
The first round of decisions centered on simplifying the installation experience and improving weather sealing performance, because solutions to both affected the fundamental assembly structure and installation story. I started with some low-hanging fruit:
Installation Improvements
Removing user-adjusted degrees of freedom
Reducing the number of user-installed fasteners
Enabling re-usability
Weather Sealing Performance
Ensure all hard surfaces touch off on compliant weather-sealing materials
Use industry-standard sealing materials and methods, suited for expected contaminant types and intended loading directions
I presented several design concepts and organized conversations amongst diverse stakeholders (ID, UX, PM, V&V testing, Thermal, Mechanical, SC). I built consensus around a ground-up redesign. We agreed to move assembly joints away from high-stress areas and use industry-standard seals. We were not able to agree on a method for filling lateral window gaps.
Prototype center bar - mounts to structural installation
bracket and seals to heat pump bridge with finned pile
I prototyped the agreed upon design direction.
This two-piece center bar is a pair of identical plastic shells joined to a reinforcing sheet metal plate. Two shells were used to reduce tooling size and cost, which turned out to be a poorly informed assumption.
The upper surfaces, which interface with the heat pump, are lined with finned pile - a window industry standard for light-duty sealing around moveable panes. A rubber pad on the bottom surface seals against the window sill.
It was immediately clear that adhesive-backed pile would not be robust enough. All following iterations used mechanical retention instead.
Subtractive adjustment, with thinner lines for snapping (pile not shown)
Additive adjustment, with rigid blocks (red) added to center bar
I tried 2 methods for adjusting the rodent guard used to fill lateral window gaps.
Subtractive adjustment
One rigid plastic shell per side, integrated with the center bar, with equally-spaced thin snap lines
One foam block per side, torn to fit within the shell
Additive adjustment
Two sizes of rigid plastic blocks (coarse and fine adjustment) that a user could snap onto a central bar
Blocks would be filled with compliant foam at the factory
In testing these rodent guards, it became clear that the "subtractive" shell was very difficult to break along the intended lines, resulting in inadequate protection and dangerous sharp edges.
The additive adjustment scheme was intuitive, enabled re-use, and seemed likely viable with further improvements to interfaces between soft and hard materials.
2nd prototype of window installation kit, with heat pump bridge mock-up (pink rigid foam)
I made a second iteration to the window sealing kit and performed a test installation.
In this installation, several additional kit items have been added:
Three pieces of foam to fill the window frame's bottom and side U-channels
One adhesive bulb seal on the bottom edge of the window sash
The pink rigid foam is a mockup of the heat pump bridge cross-section.
Center bar bowed at mid-span joint due to pressure from foam below it
Unfortunately, upwards pressure from the foam placed in the window frame's bottom U-channel caused a bow in the unreinforced center bar.
Luckily, there was appetite to pay for larger injection molding tool that would allow a unified center bar. The unified design helped stiffen up the center bar and reduce BOM cost.
Test shack, preparing for heat pump install
Duct tester with flowrate at 75 Pa
I tested air- and water-sealing performance, with help from the Thermal and Testing teams. For both of these tests, a heat pump is installed into the window of a wooden shack.
Air leak testing (ASTM E783)
A duct testing fan applies a known negative pressure to the interior of the test shack (75 Pa) and measures total exhaust flowrate.
Flowrate is compared between two conditions: a fully closed window, and a window with heat pump installed
On the final production-intent prototype, we observed airflow of 13.5 CFM (cu. ft. / min). This leakage rate was 54% lower than the baseline value of 29.2 CFM observed on an initial prototype.
Several other changes were made to the internal architecture of the heat pump bridge, which also helped reduce flowrate. Due to time constraints, we did not perform enough test iterations to rigorously attribute the impacts of each change. But, if I recall correctly, the Thermal team estimated an attribution of about 2/3 to window sealing and 1/3 to bridge architecture.
Nozzles spraying water at window areas prone to leaking
2. Water leak testing (ASTM E331)
A duct testing fan applies a known negative pressure to the interior of the test shack (137 Pa)
Water is sprayed at exterior areas of the window prone to leaking area of interest, using pre-calibrated nozzles (3.4 L/min/ m^2)
After 15 minutes, the interior of the shack is inspected
To pass, no water may leak beyond the inner-most surface of the window frame
The last water leak test that I performed before departing Gradient did not yield a passing result. The team has since iterated the design:
Tape is applied to the top of the gap-filling blocks
An additional adhesive bulb seal is adhered to the heat pump and rodent guard, directly under the window sash
Gradient launched the All-Weather heat pump in January of 2025.
The video at left documents the complete installation process, including the window seal kit
Other Responsibilities:
Modified structural installation bracket design to increase strength, reduce cost, and simplify user experience (8x fewer fasteners)
Designed various plastic features on the heat pump (ratcheting mechanism for audible confirmation of engagement with structural installation bracket, filter retention snaps)
Performed various analyses and investigations (tolerance stackup, FEA to validate testing failures, RCCA for installation processes)
Ushered all of my designs through production DFM with manufacturing vendors (China), and supported initial DFA at assembly vendor (India)
Collaborated with Industrial Design to write ~6 pages in the installation manual
Outcomes:
Reduced air leak of window sealing components by > 30% (rough attribution: ~2/3 of 54% total reduction)
Accelerated product development timelines of installation kit by 3 months within my 7 month tenure
Google Sites
Report abuse