Introduction: Nitrogen Gas Generator

Picture of Nitrogen Gas Generator

This instructable will walk you through the basics of making your own nitrogen gas generator. Using pressure swing adsorption with carbon molecular sieve you can make an endless supply from the air without using any consumables. You can use this for filling your tires (nitrogen stays in tires much longer than oxygen, reducing the time you need to refill them), having a non-combustible gas or, in my case, to feed into a liquid nitrogen generator.

I won't go into any theory here. I'll just go over the basic construction. A unit of the size I will describe can go between $6000 to $8000. You can build this for a fraction of the price. If you are interested in this subject in more detail you can go to my nitrogen generator web tutorial. There you will find more pictures, animations and other information. There is also a video going over the nitrogen generator. You can also view my liquid nitrogen generator tutorial.

If you are ready to build this for your shop just click next.

Step 1: Basic Theory

Picture of Basic Theory

This is a big project and will most likely need to stay in your shop or garage. The generator runs off of compressed air. It has two tower beds, each filled with CMS (carbon molecular sieve). Pneumatic valves control the airflow into one at a time. Under pressure, oxygen is preferentially trapped, while nitrogen passes through and out the bottom. The controller opens the next bed for filling, isolating it from the first one as it opens it to the atmosphere. During decompression, the first bed releases the oxygen back into the atmosphere, regenerating the CMS.

The picture above shows the CMS, which looks like chocolate sprinkles. They are approximately 0.5 x 2mm in size. Let's go over a basic materials list.

Step 2: Supplies

Picture of Supplies

The unit I will describe will produce 98.5 - 99% pure nitrogen gas at a flow of 1 SCFM.

Air compressor: 6HP, 6-9 SCFM depending on the pressure.

Air compressor tank: 30 Gallons

Two 8" steel tubes x 33" tall

CMS-180, 200 or H: 20kg per tower. You can get this from molecular sieve dealers. Refer to web tutorial about this material.

Two pairs of 150 psi flanges (one flange blank/one 8" flange): you can get these at Lincoln supply or McMaster-Carr.

1/4 - 3/8" stainless steel (304 or 316) plumbing. I use 1/4" as this keeps the cost lower. You will need enough pieces to make the connections.

Prefilter: you need a 5um particle filter/water trap and a 0.01um coalescing filter. I use Wilkerson filters. Again, refer to my web tutorial for all the part numbers.

Versa pneumatic valves: you need two of these.

Check valves: you need two

200 psi pressure gauge

Shut-off valve

Rotometer or other flow control device

Pneumatic controller: schematics to follow

You will need access to a welder

You will need to build a cradle to hold the towers, which will weigh over 150lbs each. Mine has wheels on it so I can move it around. You're not going to be carrying this up and down your stairs, though.

Step 3: Basic Construction

Picture of Basic Construction

Before making this nitrogen generator you should know a few things. The unit is at least 300lbs when complete. You will be dealing with pressures close to 150 psi. You will need access to welding equipment and possibly steel plate rollers.

The PSA consists of two tower beds, solenoid pneumatic valves and a controller. I sized my unit to deliver about 30 L/min of 98% or better purity N2. The bed is made from an 8" ID schedule 40 steel pipe at 33" long. The top and bottom has a welded low-pressure (150 psi) steel flange plate. They are an 8" 150# raised-face slip-on flange and a 8" 150# raised-face blind flange. A gasket helps seal this plate with a solid blank flange plate on top. A hole is drilled in the center to accept a 1/4" steel nipple, which you weld to the outer, removable plate. You can go to 3/8", but this did not seem necessary for the flows I'm using, and this increases the costs of the other hardware and pipes.

Each of the two towers holds 20kg of CMS-200 or CMS-H. You need to prevent these grains from emptying through the plumbing. You accomplish this with an insertible steel screen. You need to fashion a ring of steel that will just fit the inner diameter of the tower. Weld some steel screening onto this ring using whatever method works best for you. The screening needs to have holes about 1mm in size so the CMS grains do not pass through. This still ensures good airflow. As an added precaution you will use 10-12 MERV air-conditioning filter paper to trap dust from degrading the CMS.

Insert the bottom screen component. I welded two loops of steel on opposite sides so I could drop and retrieve the screen with two long poles. The ring holding the screen fits snuggly, so you will not be able to reach down with your hands and grap it out. Next, drop the filter paper and then fill the tower almost to the top with the CMS. Now, place another layer of the filter paper, followed by the screen. There should be now way for the grains to drop through the bottom or get discharged out from the top. Screw down the top plate.

As mentioned above, these towers are very heavy. I have mine sitting on a rack with large wheels on the bottom so I can roll it around. You will not be carrying this up and down flights of stairs, so pick a good resting place to keep it.

Above are pictures showing the bed with the screen. Again, this screening is duplicated on the bottom, except there are steel loops on the insertable ring to allow for extraction.

Step 4: Valves and Plumbing

Picture of Valves and Plumbing

Water vapor and microscopic particles can fowl up the CMS. One needs to make sure the fresh gas is clean. The system uses two filters: a pre-filter and a coalescing filter. The first traps water vapor and 5 um particles. The second filter particles down to 0.01 um. These are 1/4" port Wilkerson filters. The part numbers are F16-02-000 and M16-02-000, respectively.

The PSA has high pressure tubing coming off of the check-valve. The tubing from the two beds join together on a TEE. The output then goes to a high pressure valve that can shut off all flow. From here, the output goes to a Yokogawa rotometer, so one can control and measure the output flow.

Step 5: Pneumatic Valves

Picture of Pneumatic Valves

The bed is partially pressurized by the previously charged bed and from the pressurized reserve tank of air. When one bed receives fresh gas, the other's valves isolates it from the fresh gas and vents its tank to the atmosphere. The PSA system uses a three-way Versa valve. My system uses a normally-closed valve, so it needs power to open it to allow the fresh gas through to the bed. The model number for the 3-way, 1/4 brass, normally closed, 120vac Versa valve is VSG-3321-120. There may be an additional letter at the end.

Fresh gas enters from the top and flows into the bed when there is 120vac on the valve. When the power is off, the valve closes to the fresh gasflow and opens to the atmosphere. Exhaust gas flows from the bed out the hole on top to the left of the inflow port. Since there are two beds there are two valves. A controller handles powering and de-powering the valves.

The enriched gas leaves the bed at the bottom of the tank. Remembering that only one tank is under pressure, we need a means to prevent the pressurized, enriched gas from entering the other tank from the bottom. The system uses a check-valve that only lets gas flow out.

I got the check valves for a few dollars on Ebay. There is one under each bed.

Step 6: Arduino Pneumatic Valve Controller

Picture of Arduino Pneumatic Valve Controller

A simple arduino-based controller manages two solid state relays. The arduino runs a cycling program and manages the gates on two triacs. These allow the controller to energize and de-energize the Versa valves with 120vac.

Above is the timing diagram and the controller schematic. The c-code is simple enough to write.

Step 7: Final Product

Picture of Final Product

So, there you have it. You can now take regular air and pull out 99% pure nitrogen gas. If you want purer gas you will need a second stage, and unless you are a chemist making industrial grade compounds you won't need this.

Again, you can read more detail about this here or see a video of it here.

I have other high-detail web tutorials.

You can read about my liquid nitrogen generator here.

You can read about how to make a 3 or 12kw induction heater here.

You can see one of many video of my induction heaters on Youtube. This is the 3kw unit here.


Anthonymart33 (author)2017-06-24

how much nitrogen can we expect out of your design per hour of running

Phoelan B (author)Anthonymart332017-11-18

You could convert SCFM to LPM.

GabrielP11 made it! (author)2017-10-24

I used 13X zeolite molecular sieves to make my PSA system produce pure oxygen. Just trying to get those three way versa valves to work. X3

imsmooth (author)GabrielP112017-10-25

No, 40 psi works fine. Make sure you are connected to the correct port. They should just work. I would call the company.

GabrielP11 (author)imsmooth2017-10-25

The correct port is connected. The pipe is in the IN and the waste gas on OUT. I guess it just needs to be cleaned and oiled.

imsmooth (author)GabrielP112017-10-25

No oil when using concentrated O2. You should be able to find some lithium or silicon grease. Many SCUBA compressors use something like this. I've had one valve stick once and I fixed it by cleaning the solenoid valve. I probably have a reference on my Liquid N2 site at www home made liquid nitrogen dot com

GabrielP11 (author)2017-10-24

@imsmooth The solenoids in my VSG-3321-120 work when I apply power as I can hear the solenoid and metal pin click. However, despite the solenoids are working, gas will not vent to the atmosphere when power is off. I opened one of the ends of the solenoid to see a wide brass pin held tightly down by a spring. Is this pin supposed to slide back and forth when pressure is applied or not? Thanks.

imsmooth (author)GabrielP112017-10-24

The valve opens when under pressure. The gas goes one of the two on top and through the single vent on the bottom, pressurizing the tank. When the power goes off the pressure pops the valve, opening the other vent. If the valve is not brand new there may be dirt that is making the solenoid stick and you will need to carefully disassemble and clean it. You can also call Versa. If you do not like your valves I have two for sale that work perfectly.

GabrielP11 (author)imsmooth2017-10-24

Or maybe it's because it takes 100 psi or higher to pop the pin into a different position so the waste gas exits out the top. I only need 40 psi to purify oxygen and adsorb nitrogen instead so that's the pressure I use. I guess it's not enough?

vwrabbittrk (author)2017-01-16

imsmooth, Great build! do you have any advice or links for designing the Arduino control circuit side of your PSA system? I'm building a PSA system like yours but for generating oxygen for my torches. Thanks!!

imsmooth (author)vwrabbittrk2017-01-16

I think if you go to my website tutorial the code is there.

vwrabbittrk (author)imsmooth2017-01-16

I see the schematic and cycle timing but no code. Maybe I'm missing somthing..

imsmooth (author)vwrabbittrk2017-01-17

Send me your email and I'll reply back with the arduino file you need for the build.

supergilligan (author)imsmooth2017-08-29

Think you can edit the Instructable to add the code instead? Some of us aren't the most familiar with programming Arduino code.

Alim29 (author)imsmooth2017-02-05 please send me the mail

as my current experiance there need of ceramic balls layer and the activated aluminum to obtain 99.995% pure.

imsmooth (author)2017-05-03

The project will probably run about $1000

surfncdog (author)2017-05-01

Can anyone give a ballpark price estimate for the cost of materials to build this?

blakesteranderson (author)2017-03-26

great tutorial, I seem to be missing one thing. How do the tanks acquire pressure and release the nitrogen. It seems like the gas flows freely from the bottom of the tank and wouldnt build up to 100-ish psi.

Visit for more information.

Each tank is isolated with a solenoid gas valve. Both tanks join at the bottom and feed to a flow valve that controls the amount of gas that exists. This allows the pressure to build since more gas is going in than coming out. See the diagrams at the above link.

ahhh, that makes much more sense. would that mean that while the tank builds pressure the gas coming out would have a higher oxygen concentration, or does it build quickly enough that most of the oxygen is trapped anyways? Thanks again

Look at the videos on the link. O2 diffuses into the porous CMS due to the high pressure and N2 rich gas exists.

dauhen (author)2017-03-18

Can you review the throughput equations? It seems that 650g/l for CMS vs 2250g/l density of carbon puts your carbon "fill" at 28.9%. 40kg of CMS fills a volume of 61 liters. That means there is 43.7 liters of space for the gas. At 130psi this is 370 standard liters or 13.1 SCF. The cycle is one minute so you are processing 13.1 SCFM air. Air is 78% N2 so your input rate is 10.2 SCFM N2. The output of 1.05 SCFM N2 seems low in comparison to the input rate. Is there a theoretical maximum separation efficiency?

GabrielP11 (author)2017-03-11

imsmooth,where did you get your carbon molecular sieves?

imsmooth (author)GabrielP112017-03-12

I think I left the contact information on my website tutorial for the N2 generator.

imsmooth (author)2017-01-28

I currently have a matching pair of two brass Versa 1/4" 3-way valves that I used for this project. If you are interested in them you can email me privately. They are $130 a piece or $240 for the pair. They are practically brand new and in perfect working condition.

pietradelaar (author)2017-01-23

Dear imsmooth,

Can you make it with one pipe and on a smaller scale?

Can you leave the versa pneumatic valves out if you don't want a continuous proces?

BTW: Nice design.

Sun Gear (author)2017-01-11

Where can one obtain CMS-200, I cant seem to find a US supplier and the shipping fees for shipping this out of china is a lot 9free sample of 500g casts $67 to ship)

imsmooth (author)Sun Gear2017-01-11

I got mine from Interra Global. See if Bill Wallace is there. He helped me.

Tell them Crystal Clear Solutions referred you. I don't get anything from it but I'd like them to keep track of this. Let me know if they can help.

Sun Gear (author)2017-01-11

can this be done on a smaller scale? perhaps with pvc and at lower pressures?

imsmooth (author)Sun Gear2017-01-11

Smaller scale - yes.
With PVC? Not a good idea. With the given diameter of the tube and a pressure over 100psi you risk bursting the tube and getting shrapnel in your body.

imsmooth (author)imsmooth2017-01-11

You can get a little more info at

Sun Gear (author)imsmooth2017-01-11

well, I'd obviously be working with pipe that could handle the pressure, and some of the smaller pvc pipes out there can handle 120psi if i recall correctly.

How do you figure out the timing or is it just guess work?

imsmooth (author)Sun Gear2017-01-11

The higher the pressure, the purer the outgas. I get over 98% pure with a pressure of 130psi. You can do 100psi and get about 95-93% with the same amount of material. If you use narrower tubing you will need very long pieces to hold enough material to get a reasonably concentrated gas.

Sun Gear (author)imsmooth2017-01-11

actually taking a look at your site and finding that info there.

I assume you can get away with one tank, just with the disadvantage of having downtime

imsmooth (author)Sun Gear2017-01-11

Yes, you can use one tank if you don't mind a non-continuous supply. You still need a means to cycle the pressure/depressurization.

Sun Gear (author)imsmooth2017-01-11

having trouble locating a supplier (still looking), got any leads?

rolonrene (author)2016-09-29


Nice work! You said that to get a purer gas you need a second stage. I have read some papers and many of them said that the purity of the nitrogen, in a nitrogen generator by the psa method, is inversely proportional to the flow rate. Have you try that?

abpeak2016 (author)2016-05-19
rimar2000 (author)2014-05-27

WOW, impressive professional work.

DavidL459 (author)rimar20002016-04-18

would a mebrane be sutable like this?

stuffdone (author)2016-03-30

Thought. I just lucked into a very nice used oxygen concentrator for $20 along with a huge bag of hoses etc. My thought was to use to supply O2 for a small propane torch set instead of buying solid-ox tablets. Why not also capture the exhausted nitrogen gas at the same time?

The O2 concentrater is exactly this same principle and has the two sieves, compressor etc.

Before I spend a lot of time...has anyone here already done this? If so what is best way to collect and store the gas? ( both the Nitrogen and the O2 )

Geekrealms (author)2016-03-24

Hypothetically, could this work on Mars? Just wondering...

Geekrealms (author)2016-03-24

Hypothetically, could this work on Mars? Just wondering...

Geekrealms (author)2016-03-24

Hypothetically, could this work on Mars? Just wondering...

Empguy (author)2016-01-19

Can I increase the purity of the N2 if I just make the tower larger and increase the amount of CMS

Empguy (author)2016-01-19

Were could I buy some CMS-H there are so many sellers don't know which one I can trust

leetard made it! (author)2015-12-15

This took a while but I built it and it works great. Keep in mind the CMS will be the most costly and difficult to source part of the build. I also built mine 2x as big as the one inthe instructions due to nitrogen requirements for my application. Thanks for all the help.

imsmooth (author)leetard2015-12-15

Very nice. I built my analyzer. However, I got the O2 electrode from an anesthesia machine. You will have to find that part. The rest just involved building a circuit that could sense the voltage generated by the electrode and convert that to a value between 0 and 100.

tap-tap5 (author)2015-12-07

i have been doing so much research since first seeing your video and your pressure swing absorption machine. great video very inspiring, especially all of your failures explained. i was initialy inspired to learn about the production of the gases. my concept before i started anywhere was to produce oxygen, to mix with fuel gas, eg propane, acetylene, gasoline, etc. to be able to build a cost effective welding ,brazing system. looked at lots of different zeolite products even to the point of how are they produced. not sure if i have missed something , or if someone has messed with your info . in your video it shows 13x whitey coloured balls with odd blue ones in,. however in another tutorial it states walnut or coconut shells processed, i am curious to which one is the most effective. how can i purchase it in small quantity 20 to 40 kilograms. the pressure swing machine you have built is well smart. but it also appears you have a rocket shape system where air flows through one upright zeolite sieve. you make it look fun and exciting, great choice of music on the vid.. ps . i am a carpenter by trade, done lots of creative building work. would appreciate your input to build, pressure swing absorption machine or similar.

TheMageKing (author)2015-12-01

I would like to know if there is any way to take the nitrogen from the generator and make liquid nitrogen. Can you create an instructable on that? I saw one already, but I am uncertan about it.

About This Instructable




More by imsmooth:Nitrogen gas generatorHomemade liquid nitrogen generatorInduction Heater 12 KW
Add instructable to: