The 0.3 mm Hole That Grounded a Satellite: An Engineer's Q&A
Archi2 min read
Contents
The 0.3 mm Hole That Grounded a Satellite#
In late 2023, a small communications satellite stopped holding its orbital slot. The failure traced back to a propellant leak through a seal defect measured in tenths of a millimeter. I spoke with a propulsion engineer (call her Priya, at her request, since her employer has not publicized the incident) about the valve, the leak rate, and the paperwork that changed afterward. Her answers are edited for length and clarity.
The part in question#
Q: What failed?
A valve seat in one of the satellite's four 1-newton thrusters. The seat should close metal to metal. Ours had a gouge about 0.3 mm across on the sealing face, likely from handling before integration. On the bench, it held helium fine. In orbit, after a few hundred thermal cycles, it leaked hydrazine at a rate we measured at roughly 0.4 grams per day. That sounds like nothing. Over 18 months it bled the tank and the attitude control budget with it.
Q: Why didn't the leak test catch it?
We ran a pressure decay test at qualification. The acceptance threshold allowed a certain drop over a fixed window. Our leak sat just under that line at ambient temperature. The test was designed around a specification written when the valve supplier used a different seat material. Nobody had revisited the threshold when the material changed. The spec was correct for the old part and too loose for the new one.
What the telemetry showed#
Q: How did you find it?
Propellant use per maneuver crept up. Our station-keeping burns were taking about 2 percent more propellant than modeled. We assumed a calibration drift in the flow sensor. It took three months to rule that out. Then the orbit determination team noticed the drift was consistent, not noisy. That pushed us to the valve.
Q: Could you have seen it sooner?
The data was there by month two. Our anomaly review was organized around failures that stop a mission, not failures that erode margin. A 2 percent propellant trend did not trigger a review. We have since added a trend check on every thruster against its model, with a much tighter trigger.
The fix that mattered#
The fix was a test procedure. The redesigned seat came with a leak spec tied to operating temperature, not room temperature. That meant running the acceptance test inside a thermal chamber at both temperature extremes, not just once at ambient. It added about nine hours to each thruster's acceptance flow and roughly 40,000 dollars per unit in chamber time.
On small numbers#
Q: What would you tell a team building a smallsat today?
Measure the leak rate at the temperatures the part will actually see. Do not inherit a threshold from a heritage design without checking what changed. And write the trend trigger into the anomaly review, so a slow drift gets looked at before it becomes a problem. A 0.3 mm defect is invisible on a drawing. It is very visible in a propellant budget if you watch it.
More articles