Gas Booster

Home Up


 Warning - Boosting O2 and other high pressure gasses is extremely hazardous. Do not attempt unless you have a complete understanding of the risks and are willing to accept them (O2 fires, stored energy, etc.) !!!

I designed and constructed this device to boost (pump) O2 and other gases from a supply cylinder at a lower pressure to a fill cylinder at a higher pressure. Shop air was selected as the motive force to eliminate the source of ignition motors and electronics provide.

Shop air at 120psi supplies the red "drive" cylinder with its motive force of over 1,500lbs. A two to one lever multiplies that force providing over 3,000lbs of drive force to the pump. The hand-crafted 3400psi piston pump uses brass in all areas which come in contact with the supply gas, Sirvon seals in the piston, and two soft seat check valves in the head.

The Up-Stroke speed is controlled by the flow of air into the drive cylinder. Once at the top of the stroke a pneumatic switch is tripped allowing the drive air to bleed off in a controlled manor. The head pressure on the pump presses down on the drive cylinder until it reaches the bottom of the stroke where it throws the pneumatic switch the other direction. The drive cylinder starts to fill again and the cycle repeats. 


O2 Supply Gauge.jpg (59321 bytes) Fill Whip He-O2 Adaptor.jpg (50113 bytes) Air Supply Gauge.jpg (45097 bytes)

The supply gas can come from various sources. Here are the supply side fittings I'm using. O2 tank fitting, supply gauge, and needle valve (left). Helium adaptor (center).  Air/Nitrox yoke and gauge (right).


Fill Gauge.jpg (46557 bytes)

To provide accurate fills it's important to know the precise amount of gas added to the tank being filled. To accomplish this I'm using a giant fill gauge with 10psi divisions.


Control Panel.jpg (38312 bytes)

Control Panel - Shop air enters on the left, goes through a moisture trap, and on to a regulator (which limits the maximum fill pressure). The air then goes through a needle valve (to limit the speed of the up-stroke) then on to the spool valve and drive cylinder.


Front Shot.jpg (51010 bytes)  Rear Quarter Shot.jpg (47906 bytes)

Front and rear shots show red drive cylinder with hand-made brass pump.


Spool Down.jpg (37632 bytes) Spool Up.jpg (43253 bytes)

Spool valve...


The pump rod failed after a few hours use. Close examination of the 1/2" stainless steel shaft revealed mushrooming of the shaft and excessive wear where it went through the guide block. With these two observations in mind I took the following steps. First I increased the diameter of the pump shaft to 5/8" which almost doubled it's strength. Second I modified the guide block and installed a bronze bushing. This now provides the pump shaft with a much larger area of support and reduces friction.

Bent Pump 1.jpg (67668 bytes) Bent Pump 3.jpg (61344 bytes) Shaft-Guide Original.jpg (66010 bytes)

Bent pump shaft and original guide block.

Shaft-Guide Modified.jpg (44570 bytes)

New Guide Block with bronze bushing.

Modified Pump Extended.jpg (55550 bytes) Modified Pump Compressed.jpg (40770 bytes)

Completed "Revision 2" pump.


After a few tanks of gas were pumped the pump started moving in a herkey-jerkey kind of way (i.e. didn't move smoothly). My diagnosis was the lube on the piston seals (EDPM o-rings) was wearing off and an o-ring failure was immanent. Tearing down the pump confirmed my suspicions.

Bad O-Ring 1.jpg (48842 bytes)

This o-ring is showing signs of severe extrusion.

I had an exchange of e-mail with Steve at Airspeed Press and he suggested using either Balseals or Chicago Rawhide's "Sirvon Rings" as a piston seal. A Sirvion Ring is a Teflon seal designed to run dry. It can also be fitted with o-ring to provide a tighter seal for handling gas. After a lot of phone calls I finally determined the part number and my local bearing house obtained it for me. The main drawback to the Sirvon Ring is it requires a two piece piston for sizes this small (.875") therefore I had to build a new piston (a bit of a challenge).

New Piston and Parts.jpg (50510 bytes)

 The Sirvon ring is blue with it's o-ring sitting against it. The white rings are Teflon backing rings I used as guide rings (keeps the piston from rubbing the bore and reduces friction further).

 New vs Old Piston 4.jpg (49364 bytes) New vs Old Piston 3.jpg (47103 bytes)

The new piston is on the left with the old one on the right. While the new piston was a lot more complicated to build it is working great!


My drive cylinder seized up due to lack of lubrication (not unexpectedly). This happened due to the piston seals being designed to run with hydraulic fluid and not the shop air I'm using. The solution was to  install Teflon seals from McMaster-Carr. These seals are very similar to the CR Sirvon rings and can be found as 9420K series "Teflon Seals" (but come in limited sizes). I had to get creative with the backing o-ring size to make everything fit properly but once they were installed the cylinder worked great!


I may do more work on the spool valve as it occasionally fails to function and the booster stops. No safety hazard but an inconvenience... 


I receive a lot of e-mails regarding the booster so I've added the following items to answer the more common questions.

bulletTo find specs on the Sirvon Rings I had to e-mail and call CR headquarters as my local bearing house couldn't find it in their books (although CR swore it was there on pages 52A-53A of their main catalog). I then had to press the bearing house to special order what I needed.
bulletMcMaster-Carr is stocking rings which are very similar to the CR Sirvon rings. They can be found as 9420K series "Teflon Seals" but come in limited sizes :-( I just replaced the o-rings in my drive cylinder with one and they work great (Although I had to get creative with the backing o-ring size to make everything fit properly).
bulletThe brass pump is made from a piece of 2.5" round brass stock with a 7/8" bore and 8" stroke. This gave me a minmum wall thickness of .75" (way overkill but I figured I had enough other things to worry about).
bulletIt is extremely important to pay attention to surface finish inside the pump. On the first iteration of the pump I just reamed the hole and quickly hit it with some ScotchBrite. During the rebuild and switch to Servon Rings I did a proper job honing the bore and finishing with finer and finer emery paper.
bulletThe drive cylinder has a 3.5" bore. With this combination and my 2:1 lever I can just pump 3200psi with 125psi shop air (this is after I resolved all my seal problems using a Servon ring in the pump and a Teflon ring in the drive cylinder).
bulletMy first check valves were the all brass hard-seat ones. Unfortunately they didn't seat well enough for this purpose and had to be replaced before the pump would function at all. I ended up using the following (and excepting a calculated risk due to the Buna-N seats). McMaster-Carr - 46105K29 - Compact Ball-Check - Valve Brass, Soft Seat (Buna-N) - 1/8" NPT Male
bulletWhen designing a pump it's critical to keep the headspace volume to an absolute minimum. Remember, headspace encompasses all the volume between check-valves when the piston is at the top of the stroke.
bulletUtilize the longest pump stroke you can to obtain the most efficiency.
bulletThe booster efficiency is directly related to the input pressure vs the output pressure.
bulletDon't trivialize the complexities involved in designing and constructing the spool-valve. It's a delicate part of the machine which in my case is still in need of refinement.
bulletWhen designing your booster make sure to consider all the frictional components or you will no be able to achieve your maximum boost pressure.
bulletWithout access to your own lathe a booster project could get expensive as your design will probably go through many iterations.
If boosting high FO2 gas isn't a requirement I'd look for one of the surplus boosters found in military surplus stores. These would be fine for most uses and scavenging mixed EANx. You can even top off EANx tanks by mixing in an empty tank then boosting into the almost full tank (although with the cheep price of O2 it is hardly worth the effort). Unfortunately these surplus boosters have aluminum parts which make them unsuitable for O2 and high FO2 mixes.
bulletA booster is not a compressor and will not work with the input at atmospheric pressure. If you disagree with this statement please go and do more research before returning to this web site.
bulletI'm not sure I'd tackle this project again even though I had most of the materials on hand. It's a great tool when I need it (i.e. boosting O2 or high FO2 into deco bottles) but it was, and sometimes still is, an ongoing challenge.

Remember - Not only is there a fire hazard utilizing O2 there is a lot of stored energy in the various parts of a machine like this so take all appropriate precautions and proceed at your own risk.


Questions and comments                Copyright 1997,2006 Brian D. Basura                This site was last updated 04/02/06