Orphan Part Angel

0
The location, shape and orientation (inset) of the port are important details in ensuring accurate pressure sensing of the static system on the Glasair Sportsman. (Photo courtesy of Glasair.)

The effect that COVID had on the homebuilt industry affected almost everyone. Parts, if you could get them, were often backordered for months—or years—and costs ballooned terribly. Glasair Sportsman and GlaStar builders were especially hit hard. While the shutdown of the company’s exalted Two Weeks to Taxi program garnered the headlines, DIY builders were orphaned indefinitely while the company regrouped for better days.

As easy as one, two, three! A good drawing makes all the difference. (Drawing courtesy of Gus Gustavson.)

Sometime mid to late 2023, a thread emerged on the Glasair Owners’ group from builders seeking options for obtaining 521-1501-003 static ports. While most static ports are pretty basic—a flange with a small hole, or in the case of RVs, where a pop rivet (sans mandrel) replaces a bucked rivet—the Glasair design has a beveled segment that bisects the centerline of the port. This adds more than a minor complication to the machining process. According to a post by Arlo Reeves, “The shape was arrived at…after flight testing.” And so it was that this small and relatively unassuming little part was holding up progress for about two dozen builders.

Suggestions popped up for various solutions and one builder, Jan Detlefsen, even had a port 3D printed in aluminum. Eventually, builder Brian Hoffman from Michigan stepped in and offered to machine a batch based on a drawing that was derived from a factory part by another builder, Gus Gustavson.

In between all this, KITPLANES’ web editor, Omar Filipović, asked if I would be interested in looking at the part for a Home Shop Machinist column and (hint, hint) possibly make a few extras. I was happy to tackle the idea of how to make it in the home shop, but I am not set up for production. If it came down to it, I might have been persuaded to make a few for the most urgent in-need owners. But as it turned out, Brian offered to make as many as the community needed. Brian is what I call an “orphan part angel.” As those who got static ports from him know, the price he charged was a fraction of what the lowest-priced job shops wanted. So, basically, he did it for free (at least in terms of labor). If nothing else, it shows the value of having a solid owners group to support whatever kit you decide to build.

Using the Gustavson drawing as a guide, I modeled the part in Solid Edge in order to derive additional reference dimensions for the machining process (left). After some contemplation, I came up with a fixture block to accurately machine the bevel (right). The drawing is available here.

The challenge to making the static port was, as mentioned previously, machining the bevel per the dimensions in the Gustavson sketch. When you’re working from a sketch, where no tolerance limits are provided, it leaves no room for error! Machinists prefer to have engineering drawings over sketches. In other words, a drawing with not only the necessary dimensions, but general tolerances (except where noted to be higher or lower than the general tolerances), as well as specifications for edge treatments, surface roughness and finishing. So, when making a part or parts from a basic sketch like this, you need to take a few moments to think, Do I need to make this exactly like the sketch? Are the dimensions nominal (see sidebar below)?  In the case of the Glasair pitot static port, the bevel feature is obviously critical. Why else would it be there?

The OD of the stem was made by removing 0.050 inch per pass (left). Once you get close, allow the part to cool down to take a measurement (right).
Taking a light pass (left) to the final dimension (right).
Drilling to depth with the tap drill (left) followed by through-drilling with a 0.063-inch (#52) drill for the port opening (right).
With the lathe off, the threads were hand-tapped (left) and repeatedly checked with an exemplar fitting (right) until sufficient thread engagement was achieved. The American Society of Mechanical Engineers (ASME) chart for NPT fittings says that a ⅛-inch NPT fitting should achieve “hand tight” engagement between four and five turns, plus three additional turns to be “wrench tight.”

To get the bevel right, I spent some time making up a fixture block to clamp the part for milling the face. This solved two problems: how to hold a round part in a vise for milling and how to keep everything square.

A precision bevel protractor with Vernier scale was used to set the 6° 37′ angle on the fixture block (left). Note that while the Vernier scale graduations are in 5′ (1/12°) increments, it is possible to get a reasonably accurate 37′ setting by nudging between the 35′ and 40′ scribe lines (right). You could also set the angle with a sine vise. Having the tools to set the angle is “the catch” that comes with trying to make this part!
Clamped in and ready to go (left). Before any machining can start, you have to locate the center of the port (middle) and edge (right).
Touching off using the paper method—standard 20-pound office paper is 0.004-inch thick (left). Milling crosswise in increments (middle) until the centerline is reached. It is important to go by the machine dials. Do not try to hit the centerline by eye (right)!
Flip the fixture, touch off the edge again and start milling the bevel (left). If everything was set correctly, you will end up with a thin sliver of material as you hit the centerline position on the machine dial (middle). Advance the mill an additional 0.001 to 0.002 inch to clean away the sliver. Checking the reference dimension (right).
Drilling and countersinking the mounting holes wrap up the machining steps (left). The final steps are removing the tool marks by either tumbling or carefully hand-burnishing the flat faces with a Scotch-Brite pad (right).

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.