This report describes the development and initial testing of a small vacuum-chamber test facility. This facility was developing by Citizens in Space under contract to SpaceGAMBIT, with funding from the Defense Advanced Research Projects Agency (DARPA).

The goal was to demonstrate the construction of a low-cost vacuum-chamber facility for small payloads, which is suitable for use in a HackerSpace or similar setting. The facility will also be used to test citizen-science payloads that are to fly with Citizens in Space.


Citizens in Space has issued a Call for Experiments to fly on the XCOR Lynx suborbital spacecraft. We anticipate that most of the experiments will fly inside the pressurized Lynx cockpit, using the Lynx Cub Payload Carrier developed by Citizens in Space with support from the Space Engineering Research Center and Texas A&M University. A small number of experiments may fly in one of two aft-cowling ports, which are open to the space environment. The vacuum-chamber facility is intended to support the aft-cowling experiments.

The aft-cowling port is a cylinder 15 cm in diameter and 20 cm long, capable of accommodating a 2U CubeSat-sized payload. This set a requirement for the vacuum chamber size.

Aft-cowling payloads will experience a short exposure (approximately 4-5 minutes) to vacuum or near-vacuum. The test facility is intended to simulate this exposure to verify the proper functioning of electronics and other components.

Electronics may fail in vacuum due to overheating. Off-the-shelf electronics often depend on air flow (sometimes assisted by fans) for cooling. Low-power boards such as the Arduino, BeagleBone, etc. are less likely to overheat due to their low-power consumption. Nevertheless, they still need to be tested.

Electronic components may also be damaged by exposure to vacuum. For example, electrolytic capacitors contain semi-liquid dielectrics, which can outgas. Outgassing does not happen instantaneously, however, and may not be a concern on a brief suborbital flight.

Ambient temperature is a concern as well. The payload will be exposed to ambient temperatures during both ascent and descent. While in space, the cowling-port hatch may be opened. In that case, the payload will experience radiative heating and cooling, which will vary depending on the vehicle’s pointing angle. (The payload might be facing the space, the Earth, or the Sun.) Ambient cooling (“cold soak”) is a major concern for high-altitude balloon payloads, due to the duration of flight. Since Lynx flights are of shorter duration, cold soak is much less of a concern. We will need to do temperature-compatibility testing at some point, but we do not have sufficient data on the thermal profile at the present time. As a result, no provisions for thermal testing were built into the facility.

Vacuum Chamber Assembly


The major components of the vacuum-chamber facility are the chamber itself, the vacuum pump, and a vacuum-pressure gauge. We chose the following components:

We chose the Nalgene Vacuum Chamber based on safety and utility. Traditionally, vacuum-chamber bell jars are usually made of glass. Glass jars can fail, especially if the glass is scratched or damaged. A polycarbonate jar is much safer. The optical properties of polycarbonate are not quite as good as glass but adequate for our purposes. The Nalgene 5305-0609 jar is large enough to accommodate a 2U CubeSat-sized experiment (4″ x 4″ x 8″), which is the largest piece of hardware we contemplate testing.

Metal chambers were another option. We considered a number of metal chambers, with clear polycarbonate lids, which are offered for sale on Ebay. The design of these chambers limits the viewing se chambers offers limited viewing angles on the interior. In addition, we were concerned that the metal chamber might obstruct radio communication with devices inside the chamber. Price and weight was also a consideration.

The Robinair 15600 was selected based on product specs and recommendations from a vacuum-chamber website, The two-stage vacuum pump allows us to obtain low pressures (down 20 milliTorr, according to manufacturer’s specs). The pumping rate of 6 cubic feet per minute allows us to rapidly pump the chamber down to operating pressure. We could have used a lower-speed pump, if we had been pressed for money.

The INNOVA 3620 vacuum gauge was selected primarily on the basis of price. The gauge appeared adequate for initial testing, and we can easily replace it in the future if we decide it is not accurate enough.

To connect the components, we required a small length of vacuum line and an adapter:

Finding the proper adapter proved to be the hardest part of the project. The Robinair vacuum pump is designed primarily for servicing HVAC systems. It has three vacuum inlets, which use SAE Flare fittings. These will not fit any connector you will find at Home Depot. We tried a couple connectors that were recommended by various websites. They didn’t work. Finally, we found the correct connector.

We listed two possible connectors. Either one will work. They fit different inlets on the vacuum pump, but you can use either one. We haven’t noticed any difference in performance.

We’ve placed an asterisk next to the prices because Lesman has a $25 minimum order. To make up the minimum order, we ordered four of each type, so we now have spares. The fittings were not in stock and had to be special-ordered from the factory, which took a bit of time. So, you might want to order these first, if you’re putting together a vacuum chamber in a hurry.

We also ordered the following supplies:

The Robinair 15600 vacuum pump comes with 16 ounces of oil, which is enough to fill the pump. So, in theory, we did not need to order additional oil. We spilled some oil, however, so having the extra gallon came in handy. We still have nearly a gallon on hand for future oil changes.

Finally, we purchased some safety equipment. As noted above, the polycarbonate vacuum chamber is much less likely than a glass chamber. Nevertheless, the safety instructions that ship with the Nalgene chamber recommend using safety glasses or goggles and an acrylic safety shield. We purchased the following items at Lowe’s:

  • (2 pair) Safety goggles. $7-8 each.
  • (1) “Optix”-brand acrylic sheet, .22-inch thickness, 24×48 inches. Approximately $50.

Our total expenditure came to slightly under $700. If you’re short of cash, you could reduce that somewhat by buying a used pump and vacuum chamber on Ebay.


Once we had all the necessary components, assembly was straight forward. We completed the following steps:

  1. Filling the Robinair 15600 vacuum pump with 15 ounces of oil, then running the pump briefly according to the instructions in the Operating Manual.
  2. Installing a Hose Barb to Female Flare adapter to to one of the pump’s vacuum inlets. All you need to do is remove the intake cap (screw off) and screw the adapter on. Robinair 15600 has three inlets: two 1/2″ Flare and one 1/4″ Flare. Your choice of inlet will depend on which of the two adapters you are using. The 146HBLFSV-4-4 fits the 1/4″ Flare inlet while the 146HBLFSV-4-6 fits the 1/2″ Flare.
  3. Cutting the vacuum-line into two pieces. This was done with a pair of household scissors.
  4. Connecting the two pieces of vacuum line to the pressure gauge. The INNOVA 3620 gauge comes with a small plastic T-connector. The T-connector has three barbs which fit 1/4″ inside-diameter vacuum line. We connected the gauge to the center barb of the T-connector and one piece of  vacuum line to the barb on each end.
  5. Connecting the free end of one vacuum line to the Hose Barb to Female Flare adapter that we installed on the pump.
  6. Connecting the free end of the other vacuum line to the barb connector on the Nalgene vacuum chamber.
  7. Smearing a small amount of grease over each connection to help prevent vacuum leaks

First Test Series

When all components were assembled, we ran our first series of tests. The primary objective of these tests was to verify proper operation of the chamber.  The secondary objective was to demonstrate our ability to capture images of test subjects inside the chamber.

We originally planned to capture images by shooting through the transparent sides of the chamber, using a camera located outside. Shortly before the tests, we realized that our Hero 3 GoPro camera was small enough to fit inside the chamber. The GoPro’s waterproof case is designed to withstand dives down to 197 feet, which corresponds a pressure of 7 atmospheres. Therefore, we reasoned, it ought to easily withstand a one-atmosphere pressure difference in the vacuum chamber.

Our desire was to mount the camera, inside its watertight case, to the “ceiling” of the vacuum chamber, looking down at the base. This raised the question of how to mount the camera inside the vacuum chamber. We could not use any sort of screw fastener without damaging the chamber. Suction cups were obviously out of the question. The obvious solution was to use tape. First, we tried Duck-brand low-residue tape (i.e., painter’s tape). Unfortunately, the low-residue tape was not quite strong enough to suspend the camera. We switched to the old reliable silver Duck tape, which worked well but did leave a messy residue on the camera case.

During the tests, we turned the camera on and off and changed video modes by means of a WiFi connection, using the GoPro Remote app running on an Apple iPhone.

During this test series, we conducted two test test runs. For the first test, we used a small plastic toy balloon as the test subject for the camera to focus on. The following video shows the balloon swelling up as the pressure inside the chamber falls.


Watching the video, we realized there was a problem with the test. As the balloon swelled up, it obstructed the vacuum inlet at the bottom of the chamber. It is possible that the balloon’s obstruction affected the pressure we were able to achieve in the chamber.

For the second test, we replaced the balloon with a small plastic dish containing warm water. In the following video, you can see the water starting to boil in the vacuum chamber as the pressure falls. (The flashing red dot that appears in the water is a reflection of the camera’s flashing red operating light.)


During the second test, the gauge showed a pressure differential of better than 29.5 inches of mercury (749 mm Hg). Both tests were fairly short, lasting under two minutes.

We ran into a problem after the tests were complete. We were unable to open the GoPro camera’s watertight case. After some momentary puzzlement, we realized our mistake. Obviously, the seal on the GoPro watertight case is designed to keep pressure out, not to keep pressure in. We now had a vacuum inside the case, holding it shut. We set the camera and case aside for a while, hoping air would leak in. It didn’t. As we said, the case is designed to keep pressure out. Next, we tried slipping a blade into the case (being careful not to damage the seal), hoping we could “burp” some air into the case. That didn’t work, either. Finally, we resorted to brute force: pulling the case open, using a pair of pliers for grip. This worked, although the case suffered some moderate damage as a result.

This accident proved two things: 1) The GoPro watertight case is not vacuum-tight. 2) The GoPro camera can survive (and continue to function in) at least a partial vacuum.

If you’re thinking of exposing a GoPro camera to vacuum, please note: we did not have any means of measuring the vacuum within the GoPro case. We don’t know what level of vacuum the camera was actually exposed to. To really prove the camera’s vacuum hardness, we would need to expose it directly to vacuum, without the protection of the watertight case. That’s not an experiment we plan to conduct at this time. Our GoPro camera is the $399 Hero 3 Black Edition, which we would rather not sacrifice. If we decide to do this experiment at some point in the future, we would like use the $199 White Edition.

Second Test Series

Having satisfied ourselves that the chamber and pump were operating correctly, we began testing electronic components. During this series, we performed four test runs.

During the first test run, we exposed a 9V battery to vacuum for five minutes. The battery would be used to power circuit boards for later tests. Before we started adding other components, we wanted to verify that the battery would survive. If the battery was going to rupture or explode inside the vacuum chamber, we wanted to find that out without risking damage to a board.

Our Safeway 9V battery survived the test with no apparent harm. During the test, the vacuum gauge registered a full 30 inches of mercury (762 mm Hg). This represents a full vacuum, within the limits of gauge accuracy.

After verifying that the battery would survive vacuum, we prepared an Arduino Uno board for vacuum test. We loaded the board with a small program that would blink the onboard LED in a predetermined pattern, which we could monitor during the test.

Arduino Uno vacuum test

The Arduino Uno (shown above) survived vacuum exposure for a full five minutes — a time chosen based on the expected duration of a suborbital mission.

Next, we repeated the test using an Arduino clone, the OSEPP Uno (shown below).

OSEPP Uno Arduino clone vacuum test

Once again, the board survived and continued functioning for the full five minutes. Finally, we repeated the test with another Arduino clone, the RedBoard from Sparkfun (shown below). Once again, the board survived vacuum exposure for a full five minutes.

Sparkfun RedBoard Arduino clone vacuum test


This work was funded by SpaceGAMBIT, which is sponsored by the Defense Advanced Research Projects Agency (DARPA) of the US government. Views expressed this page are those of the author and do not reflect the official policy or position of the Department of Defense or the US government. No official endorsement should be inferred.

Written by Astro1 on February 20th, 2014 , Citizen Science (General)

Leave a Reply

Your email address will not be published. Required fields are marked *