Home » Oxygen Concentrator » Our systems » Ox02: Two Column

Ox02: Two Column



  • Continuous production
  • Pressure equilibration betweens columns

Video of latest prototype


  • Column housing = Reverse Osmosis Membrane housing
  • 250g LiX in each column
  • Valves: Festo, Switching controller: Time switch
  • 100% oxygen in outlet stream (3 l/min) using pressurized air (3 bar) in input.
  • Switching cycle : 4s ON/4s OFF

Additional information

This is work in progress. Please contribute in the comments. All valuable inputs will help us define the best solution.


    • Dear Alain,

      when talking to doctors, we realized that each patient requires different amounts of oxygen, depending on his health status. But you are right, the goal is to have as much oxygen as possible at hand to be able to react to different scenarios. One simple solution regarding our prototypes would be to use as many in parallel as needed to reach the required oxygen flow.

  1. Hello again!
    I talked to a doctor today and he gave me these informations :
    Most patients need only nasal cannula oxygen help and these are usually set between 2-4 l/min (for medical grade purity). If one needs more blood oxygenation help, they would switch to mask and use a higher flow rate. There is no rule for flow rate usage. It depends on how the patient reacts and how his oxygen blood levels react to the treatment.

    Any way the doctor said that 2 l/min at 65% already helps. These are not insignificant amounts!

    Rehumidification of the air is only for convenience. The nose of the patient could dry out and start bleeding. If the patient is using the device in a room where the humidity of the air is not too low, it can be enough to count of the mix of ambient air with the dry oxygen to balance the humidity. Most of all rehumidification of the air can bring a lot of problems so the doctor said not to focus on that too much.

    I hope these inputs can help.

    • Dear Alain,

      Thanks for sharing this information, feedback from doctors is so invaluable! We will include that in our approach going forward.

      Concerning your numbers, we have received similar input, the need for oxygen is very diverse, differing in concentration and pulsation rythm. We will soon publish a next prototype with one column, around 1.5 L/min of 90% oxygen with very few parts, so we are getting there.

      Going forward, we want to setup a forum for future discussions and questions so make sure to follow it: http://oxygenator.geprojects.tech/phpbb/

      Have a nice day and stay safe!

      • Hi there,
        I’m anaesthesiologist from Canada.
        The above was certainly correct with the exception of the role humidification plays in oxygen therapy.
        Certainly this is easily corrected by bubbling the oxygen through water.

        In the OR setting, volume and concentration are important as we typically use balanced techniques using a combination of oxygen and nitrous oxide. Other gases are still used, but total intervenous anesthesia with agents like propofol are making halogenated gases less common. This is in a 1st world setting. Anesthesia in 3rd world is quite different.

  2. I have a question on the number of valves used…can’t we use a 4way directional valve for the inlet from the compressor and the nitrogen exhaust?

    • For the 2-column setup, you could indeed use a 4-way valve, one connected to the compressor, one to the exhaust and two connected to the respective column.

  3. Hey,
    just one question, what is this equilize purge and/or equlizing step in 0x02 prototypes? I don’t understand it’s usage


    • Oh, I get it what equilization step does. Now I have got another question.

      If we use exhaust nitrogen to help in re-pressurizing other column, wouldn’t it effect productivity, as we want more Oxygen in column and not more nitrogen during pressurization

      • There are two advantages of equalization step:
        1.If you look at the sketch of the process, we do not use exhaust air to equalize but production air (meaning more oxygen). This way we actually accumulate oxygen on the production site of the column.
        2. In addition, the equalization step eliminates large pressure differences while pressurizing or blowout. If there were no such step, we would each time pressurize from 1 to 4 bar or vice versa blow down from 4 to 1 bar. By doing this, we also save on mechanical energy (step wise pressurization from 1 to 2.5 and 2.5 to 4 is more efficient than 1 to 4).

        Disadvantage: Slightly more components needed and additional step to program

        • If we use the flow control valve(or needle valve) at the purge, that would slow down the depressurization and make the next pressurization phase more efficient, as pressure will be available in the column.
          This way we wouldn’t require much components. Not sure though if this would work in real world

        • I see that you guys have used needle valve in Prototype 0x02c for equalization, does that work for equalization?
          I mean if we slow the flow at needle valve too much then it wouldn’t help in pressurizing the other column(or equalization), but will help in feeding slow oxygen into the other column
          Not sure about what would be the effect if the needle valve had faster flow, I assume pressurization phase would take more time.

          • Yes you are absolutely right, with the needle valve, we do not go for total equalization but just a small flow to the other column to accumulate in oxygen on the production site. The needle valve is therefore almost closed. It is definitely a parameter that is worth playing around to optimize.

  4. Sir please send ne arduino code for 2 column o2 generator. Quantity and type of ziolite for membrane housings.

    • We are getting maximum 64% (1.5 lt/min) oxygen with two columns set using 13x zeolite.
      Is it possible to get more oxygen with 13 X zeolite?

  5. I had an idea on how to calculate the oxygen concentration, without requiring expensive sensors.
    This is just based on the fact that the atmosphere contains 21% oxygen.
    We can do a candle and glass experiment and count the time for burning of candle
    First we will take a glass with atmosphere air and then keep it inverted on burning candle and count the time for burning.
    Next we will take a glass with Oxygen rich air (produced by the o2 concentrator) and then keep it inverted on burning candle and count the time for burning.

    Here is a simple formula using cross multiplication:
    x = (Time taken to burn in oxygen rich air x 21) ÷ Time taken to burn in atmospheric air

    Where x is concentration of oxygen rich air(produced by o2 concentrator)

    Not sure, this would work, but I assume it may give approximate values

      • Hi,

        Thank you for your reply sir.

        I am trying to develop a similar system(two column system) but with non-lithium zeolite for which I have two questions if possible please help me with these:
        1 – How to calculate the amount of zeolite required to produce x amount of oxygen?
        2 – How to design the zeolite column(length, diameter)?

  6. Hello, I like the project. We are making the prototype to show it to people who can support for free.
    I have doubts about the size of the osmosis casing

    • Dear Manolo,
      We realized that the bigger the column, the more pressure you need to be able to cycle fast. And our target was using the least valves possible to make the setup simpler, while also producing at high oxygen rate. We therefore decreased the size from the first prototypes. But you are correct that each setup has its optimal case size. If you want to continue this discussion, please do so in the forum! Best, Julian

  7. Hi, i’m pietr,
    forgive my ignorance, as i wrote in the forum, I’m a newbie but I’m getting started with this project to test myself, In your project i understlood everything save the function of the valve n°7, where you wrote equalize, may i please ask you if you’d tell me overall what is its task and utility?
    Yours Pietr

    • Hi Pietr,
      Valve Nr. 7 (Correct Schematics here: http://oxygenator.geprojects.tech/index.php/2020/04/05/ox02b-getting-more-oxygen/) is open between the step of pressurizing one column while the other is being depressurized (release nitrogen). The equalization step is not absolutely necessary but serves two goals: 1. to equalize the pressure between both columns (instead of each time pressurizing from 1 to 4 bars, the pump only has to pressurize from 2.5 to 4 bars for example.) This saves on mechanical energy. 2. By letting some oxygen from the production side of one column under pressure to the other column, you accumulate oxygen on the production side.

  8. Hi, I am confused about the phase called feeding? Above it says “Switching cycle: 4.5s pressurizing/blowdown, 0.4s feeding, 1.3s equilibration”. Can you please explain what is feeding in the context of a 2 column design. Ideally if you could provide a timing diagram that would be very helpful. Regards.

    • Feeding is when we only open the valve towards the oxygen tank, you could also call it production. An alternative is to introduce a needle valve that constantly lets through some concentrated oxygen at a low rate.

  9. Good morning,

    this project is very interesting, thank you.
    I was wondering, how do you determine the switching cycle time between each phase?
    I guess depends on the beds as on the compressor, is it right?
    Any ref please?

    Thank you

    • We started off with some literature values, but we ended up using a simple 4s cycle on/off and we had decent results. It is actually much more important to regulate how much of oxygen backflow from one to the other column you have (we regulate with needle valves)

  10. We are getting maximum 64% oxygen ( @ 1.5 lit/min) with two columns set up using 13x zeolite. Is it possible get more with 13 X zeolite ?

    • We have not tested 13X with double column setups, but from our experience with single column setups and 13X, this already sounds like a good result.

      • We are using both dry pressurized air (less than 10%rel humidity) and compressed air at around 70%rel humidity. The compressed air however is dried first in silica gel columns before entering the zeolite columns

    • Are you using the lithium exchanged 13X?
      My understanding is that PVSA works better with these columns.

      • We are using LiX zeolite with PSA. We have never switched to PVSA to compare but we have also heard that people get good results with it.

    • It depends on how much flow you have. We can get to very high concentrations (>90%) but at low flow rates (1 l/min)

  11. Respected, would like that someone share the designing of zeolite container, I’m from Pakistan and work on the same project.

  12. thx much for the invitation :). I am expert of pandemic, and i can help you.
    PS: How are you? I am from France 🙂 very good forum 🙂 mixx

  13. On your schematic, are the solenoid valves properly labeled? They don’t seem to match the diagram

  14. Hello kochju,
    I tried to make 2 column 2.5″x18″ filled with Zeolite LAH Sieves dia 0.4~0.8mm, Input pressure 6kg/cm2 & Charging time 10 sec & Equalization time 1.6 sec ,But Oxygen fluctuate & reach upto 40% only. Can you suggest ,How i can improve ??

Leave a comment

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

Recent Comments