Introduction: UVC Sterilizer for COVID-19 Emergency

How to make a UVC sterilizer box. First things first. Electricity is dangerous! If you are not confident and competent do not attempt anything mentioned below. UVC (253.7nm) light is powerful, it can blind you and possibly give you skin cancer if exposed to your eyes or unprotected skin. Use a welding mask and protective equipment if you want to check the bulb function outside of the box (very strongly not recommended). If building one of these units please check with your hospital or local healthcare facility before just showing up with one.

The hospitals here, and nationwide, are running out of supplies. Our local hospital workers have to re-use their masks with NO way to effectively disinfect them. As this COVID-19 disease spreads we need to start working as a team to get this under control.

Over the past weeks I have developed a machine that can help our healthcare system cope with this. I have designed a UVC system that is able to disinfect reusable hospital equipment in minutes (n95’s, surgical masks, goggles, stethoscopes etc). This equipment is very inexpensive and easy to build. Currently our local ER's are using my systems to disinfect their PPE. These systems can effectively eliminate our n95 mask shortages. Any Engineer, Maker, STEM student, and capable handyman can make them. This will help take the stress off of hospital workers, as their disposable equipment can now be safely re-used.

How To make this:


Here is what I used to make these units:

UVC Bulb (germicidal) 2G11 (1)

2g11 base (1)

2g11 bulb holder (1)

ballast (1)

ballast options... please see end of article for updates on ballasts and bulb life.

limit switch (ac/dc) (3)

12-18 gauge wire ~5 ft available by the ft at home depot, or use the extra wire from the ballast

aluminum sheet I bought from scrap yard 3/32 inch or use aluminum tape to reflect light

aluminum tape (alternative to aluminum sheet)

plastic box (1)

mechanical timer( 1)

power socket or cable gland

circuit breaker 2A (redundant safety backup...if you are using above power socket, it has a 5A fuse)

Step 1: UVC Sterilizer

Find a way to get your AC power into the box. Use either a cable gland, or a power socket.

Step 2: Gather Up Your Components

For this you will need your circuit breaker, mechanical timer, limit switches, ballast, 2g11 socket and 2g11 UVC germicidal bulb.

(Look at pictures for actual layout and electrical schematic)

As soon as the power inters the box, connect the hot wire to the 2A circuit breaker.

Connect the neutral wire to the neutral on the ballast and connect the ground to the metal case of the ballast.

Connect the output from the circuit breaker to the input of the mechanical timer switch.

Connect the mechanical timer switch output to the normal open terminal (N.O.) of the first limit switch.

Connect a wire from the other side of the limit switch and connect it to the normally open (N.O.) terminal of the second limit switch.

Take the output from the second limit switch and connect it to the load (black wire) on the ballast.

From the ballast connect the blue wire to one side of the 2G11 base. If there are 2 blue wires, cap one of them with a wire nut. Cut the Red wire from the ballast output in half.

Connect one end of the red wire to the normally open (N.O.) terminal of the 3rd limit switch, connect the other end to the output from that switch. Attach the other red end to the 2G11 base across from the Blue wire.

Cap any other wires coming out of the ballast.

Now the fun begins.

Step 3: Arrange the Box and Add Reflective Aluminum.

Now find a way to mount all of this inside the box!

Mount the bulb in the top of the box (see Photo).

Please note pictures are of a more complex computer controlled box with extra components (dc power source, din terminals, and a solid state relay), but you get the idea.

Please see comments from jfox240 below for additional pictures!

Group all of the electrical components on one side of the box to allow for maximum sterilization area. I suggest using acrylic to isolate the electrical components from the user.

Mount the limit switches so that they only close when the box lid is tight. I used a multimeter with the "tone" conductivity alarm. Attach your meter to the switches and position the switches on the inside of the box such that it has to be completely closed to engage the switch. Mount the switches tightly with screws so they do not move during heavy use.

Now coat the interior with reflective aluminum. I used 3/32 sheet and screwed it to the inside of the plastic. just make sure your electrical is isolated completely as aluminum will conduct electricity. A cheap alternative is to coat the inside of the box with aluminum tape. I also removed the pressure relief valve and filled the hole with hot glue to allow the user to see that the bulb is lighting up.

Now make some supports for masks, goggles, stethoscopes etc. I used clear acrylic and 50 lb fishing line to create a grid 1" above the floor of the box. This allows items to not touch the bottom, thus allowing reflected UV light to sterilize both sides. I also coated the acrylic with aluminum tape to increase reflection of UVC.

Check your UVC output from the bulbs actual UVC will vary. Thanks to Dr. Bohl for this information. It appears that UVC output is ~1/3 of the rated wattage from most bulbs.

So for a 36W bulb , we would have ~12W UVC (12,000,000μW) over an area of ~1440cm2 gives 8,300 μW/cm2. Over 1 minute this doses the items with ~ 500,000μW/cm2 or 0.5J/cm2. These values also decrease exponentially with distance. These are just calculations and your intensity will vary based on temp, bulb cycles, and reflectivity of UVC among other things. Ideally these can be measured with a UVC meter if available.

JUST OUT!!!! They made a room sized version!!!!

Notes on variations:

I used instant start ballasts in this method as their strike voltage is high enough that I have never had a problem starting the UVC bulb. These ballasts do cause premature failure of the bulbs due to the sputtering of the filament material during the high frequency start. I intentionally chose this type of ballast as they worked with all bulbs that I could find. I had problems with several programmable start and rapid start ballasts as they had insufficient voltage to start the bulbs. Initially I thought this reliability tradeoff was worth the decrease in bulb life as bulbs were cheap and readily replaceable. We had our first bulb failure yesterday(unbranded bulb from amazon) it lasted ~1500 cycles and had significant discoloration near the filaments due to electrode sputtering. 1500 cycles is ok, but I think we can get better lifespan by using dedicated UVC ballasts of rapid start ballasts. I did get this rapid start model to work using two bulbs instead of one :

This type of ballast would be in improvement to the instant start as it pre-heats the filaments before starting. We should theoretically have an increased bulb life; we will find out soon. I will add a new diagram of this 2 bulb design tomorrow. Adding a bulb to the bottom of this unit will dramatically increase the UVC output, and allow better coverage of the underside of items without depending on reflection of UVC. If there are any ballast experts out there please message me so i can improve this design. I would like to get as long of a life as possible out of these bulbs as they are becoming harder to source and considerably more expensive.


If you want to make a micro-controlled version, simply wire the first 2 limit switches as inputs to the controller. send 1 signal high and 1 low then check the switches in the logic. Then use output to drive a SSR to control the ballast. I use a logic level MOSFET to drive the SSR with 12V, as this is less prone to have problems with EMI generated from the system. Also only use shielded wire for logic switches as the EMI will probably give faulty logic signals to the controller when this system is running. Always use a backup limit switch that directly controls the AC to the light if using a micro controller system as you need protection if the board malfunctions.


Great Articles on this method:

Associated smell after cleaning with UVC (smells like burnt hair) :

Please see below comments for dozens of applicable journal articles on n95's.


-Brian Crabtree

MS, BS Bio-Engineering from University of Idaho

To contact me please see my website. Please be patient as I am overwhelmed with questions currently. You can always contact the engineering and microbiology department at your local university if you have questions about this technology. I didn't invent this technology, I just designed a cheap, affordable and easy to make version.


Spokane WA

Have fun making this!