Results 1 to 10 of 10
  1. #1
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,757

    Default CO gas sensor monitor

    You can buy a CO gas sensor monitor on eBay for about $10 so this project was not about saving money but rather about learning / relearning how sensors and microprocessors combine and programming them to do what I want, rather than accepting the usual single number readouts the commercial units provide.

    I had an industrial strength CO monitor in my shed for my forge/furnace that I purchased on eBay from the UK. It cost about $70 and died after about 4 years. I have since learned that the sensor will last longer if it is turned off when not in use which for me is basically all over summer.When turned on they take some time stabilise and I just needed to remember that.

    Anyway I looked up the replacement cost of the actual sensor used in the old monitor and it was going to cost around $50 just to replace the sensor. I looked around on eBay and saw here were lots of cheaper gas sensors that interfaced with Arduino processor boards. I also needed to learn more about Arduino because I in the middle of interfacing a new generation dust particle sensor to an Arduino and it seemed like putting together a CO sensor would be good practice.

    Hooking a CO sensor up to a basic Arduino board and getting a number on a computer screen is pretty easy. There are heaps of Instructables/Youtube clips/web pages that show how t do this and also provide the Arduino code. I reckon a primary school student could do this with not much effort.

    However it's one thing to hook these up on a prototype bread board its another to package these components, plus other bits and pieces into something portable/robust and reasonable looking.

    Anyway this is what my gismo looks like.

    As well as CO sensor and an Arduino the box contains a Temp/Humidity sensor, a 2 x 16 character LCD display, a reset switch, and a toggle switch to send different output data to the LCD. All the parts, including the box, LCD, sensors, switches, arduino board cost ~$25.

    It runs off a Li-ion USB battery pack, or a USB wall wart, and a USB cable.
    Here the LCD is displaying the Humidity and Temperature updating the values every second or so.
    The red button resets the measuring system if it gets confused but after running it for 48 hours its still churning away nicely.
    IMG_2322.jpg
    The LCD toggle switch on the RHS, switches the output of the LCD from Temp/humidty to show CO gas concentration.

    IMG_2323.jpg
    The ∆ - +58% shows the change in CO readings was +58% over last 10 readings so you get an idea of how quickly its changing.
    These sort of features are no available on a $10 monitor.

    I wanted to also package a flammable gas sensor into the package but there was not enough room in the box. I will make another one that incorporates the two gas sensors and this time use a bigger box and I also want to incorporate data storage via an SD card into the box..

    The programming options are almost infinite and to do that I hook the unit up to a computer via a USB cable.
    This also powers the device while allowing for programs to be edited and updated.

    Like I said, its not rocket science, it helps if you have some basic programming skills - the code is written in a cut down version of C++ - especially if you want to get into making more use of the data.

    Overall I was quite happy with the outcome.

    Have since played with a raft of other sensors including an accelerometer, tacho/position sensor, and a dust particle counter and have reported on the latter in the dust forum.

    Apart from building anther gas sensor and getting a portable particle detector going I also plant to have a crack at automating the data collection of torque and RPM on my Horsepower Rig.

  2. # ADS
    Google Adsense Advertisement
    Join Date
    Always
    Location
    Advertising world
    Posts
    Many





     
  3. #2
    Join Date
    Jan 2014
    Location
    Sydney Upper North Shore
    Posts
    4,464

    Default

    Can you program so the CO sensor reading displays % rather than ppm?
    Cheers

  4. #3
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,757

    Default

    Quote Originally Posted by Lappa View Post
    Can you program so the CO sensor reading displays % rather than ppm?
    Cheers
    Sure, that would take about 30 seconds to modify the program to do that.
    This would make the reading above be 0.0081 % which is quite high.

    The recommended levels are
    9 ppm CO Max prolonged exposure
    35 ppm CO Max exposure for 8 hour work day (OSHA)
    800 ppm CO Death within 2 to 3 hours
    12,800 ppm CO Death within 1 to 3 minutes
    I have been doing a bit of reading about CO and CO sensors.

    These CO sensors only read reasonably accurately once they have been running for ~48 hours and have a lower limit of reading of about 10 ppm and an upper limit of 1000 ppm.
    Normal levels in the atmosphere are well below the 10 ppm limit of detection so these sensors are not suitable for general atmospheric monitoring.
    What they are really designed for is to warn operators of equipment that generate CO what sort of exposure they are experiencing and perhaps trigger alarms.

    There are ways to sort of calibrate them but I have not played with this yet. There is a bias/sensitivity resistor on the sensor that can be tweaked to get a more representative zero reading.

    When first turned on they read very high, around 300 ppm and then quickly comes down to less than 50 ppm in about a minute and then continues to drop slowly before it levels out.
    The sensor draws about 150mA at 5W and it does get noticeably warm.

    One reason this sensor is not super accurate is it also picks up to some other gasses (e.g. if you breath on it this changes the reading by a few PPM, apparently its not the CO2 but the methane) so it gives higher that usual readings if these are present.
    So it can give false positives - better that way than the other I suppose.

    Placing it behind car exhaust sees it really rocketing up quite quickly, I stopped it when it got to 950 ppm.

    When using my forge with one extractor fan the CO level could go over 200 ppm after only about 30 minutes of using it, now with 2 extractor fans it stays at less than 35 ppm but after about 2hours of running the forge the whole shed gets too hot so two hours is about the max I can run it anyway. If needed for extra security I could run the DC system as well.

    Am currently building the portable box and display for the dust particle sensor. Its fiddly stuff, 2.5 and 3mm nuts and bolts, small switches eyc. I'm not used to dealing with such small components and find it a bit frustrating walking back and forth to the shed to get a "this and that". Should have it finished tomorrow so I can take the particle detector down to the shed for proper testing.

    The particle detector program is coming along nicely. It generates a set of data about every 3 seconds and I have it automatically accumulating about 20 of these so that I can compare it directly to the particle counter from work.
    I've ran the particle detector overnight on the bench in my study which uncovered a couple basic programming bugs like integer overflow and timing error.

  5. #4
    Join Date
    Jan 2014
    Location
    Sydney Upper North Shore
    Posts
    4,464

    Default

    In regards to your comment on calibration - I was reading articles on these sensors and most said they needed to be calibrated after running continuously for 48 hours. We calibrate our machines with a known gas mix so I was wondering whether this would be viable with your unit and whether it would be required?
    Cheers
    Last edited by Lappa; 27th June 2017 at 08:17 PM. Reason: Damn autocorrect!

  6. #5
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,757

    Default

    Quote Originally Posted by Lappa View Post
    In regards to your comment on calibration - I was reading articles on these senders and most said they needed to be calibrated after running continuously for 48 hours. We calibrate our machines with a known gas mix so I was wondering whether this would be viable with your unit and whether it would be required?
    Cheers
    A known gas mix with accurately known water content is indeed the best/industry/scientific way of calibrating.

    I think we may be talking chalk and cheese here.

    CO sensors used on vehicle exhaust analysers have to operate at much higher levels than the sensors I am using which are similar to those used in the consumer.household monitoring devices that run continuously for years without calibration. It's probably why they won't claim anything above 1000 ppm 0.1%.

    Motor vehicle analyser sensors are measuring well above this and I wonder if they are used for long periods they become saturated with all the other crap that comes of of motor is why the need recalibrating?

    The consumer level units may be +/- 10, 20 or even 50% out but like car speedos they would be deliberately made to read too high.

    If I needed accurate measurements, I would have to calibrate properly but there are a ways to get a crude calibration so that one can stay safe at low levels.

    A low end reference point is relatively easy to get by visiting a local air quality station site and watching their atmospheric CO readings on line and calibrating a zero to that data.
    The other way would be to go to the beach when the sea breeze was blowing and the CO should be less than 1ppm.
    In Perth unless right next to a road its rare to see it go high anyway, it's relatively windy and we have very low CO air coming at us from all sides.

    The high end (1000 ppm) is trickier and is nominally achieved at 4.0 +/- 0.1V analog output in the sensor I'm using. Bearing in mind I'm working at < 50 ppm it doesn't matter if the high end is +/- 50% out because at <50 ppm the uncertainty is dominated by the zero end reference point which I will set at +10ppm for added safety.

    I will be doing a bit of testing on the calibration and will report back. Its a pity I wasn't a bit closer and could compare it to your engine analysers.

  7. #6
    Join Date
    Jan 2014
    Location
    Sydney Upper North Shore
    Posts
    4,464

    Default

    We get ours calibrated once a year just to ensure they are accurate. Not a biggie, just part of a normal service routine. The machines self test on turn on. They test low end and high end readings. There are 3 filters and a water trap. Filters are changed regularly and they have purge functions if they sense abnormal levels when not testing.
    Only real servicing is O2 sensor cells and filters and a cal. check.
    I'm looking forward to more results from your unit. it would be interesting to run one in a auto. workshop to measure CO in the actual air.
    Cheers

  8. #7
    Join Date
    Aug 2007
    Location
    Saskatoon, SK, Canada.
    Posts
    1,436

    Default

    Bob I doubt I'll ever make a CO detector for myself but I am curious what kind of readings you would get from around the barbecue. Maybe the red meat isn't doing us in, just the cooking of it.

    Pete

  9. #8
    Join Date
    Feb 2003
    Location
    back in Alberta for a while
    Age
    68
    Posts
    12,006

    Default

    Quote Originally Posted by BobL View Post
    The recommended levels [of CO] are
    9 ppm CO Max prolonged exposure
    35 ppm CO Max exposure for 8 hour work day (OSHA)
    800 ppm CO Death within 2 to 3 hours
    12,800 ppm CO Death within 1 to 3 minutes

    These [low cost] CO sensors only read reasonably accurately once they have been running for ~48 hours and have a lower limit of reading of about 10 ppm and an upper limit of 1000 ppm.

    What they are really designed for is to warn operators of equipment that generate CO what sort of exposure they are experiencing and perhaps trigger alarms.

    When using my forge with one extractor fan the CO level could go over 200 ppm after only about 30 minutes of using it, now with 2 extractor fans it stays at less than 35 ppm but after about 2 hours of running the forge the whole shed gets too hot so two hours is about the max I can run it anyway.
    now that is interesting, and potentially very concerning for home metal workers
    regards from Alberta, Canada

    ian

  10. #9
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,757

    Default

    Quote Originally Posted by ian View Post
    now that is interesting, and potentially very concerning for home metal workers
    I should add my forge is a forced air, mains gas powered forge. The oxidation level of the flame is controlled by controlling the air/gas ratio with too much air producing a higher amount of metal scale so less air is provided which means incomplete burning of the gas and high amounts of CO being produced. It's not like I use it every day, maybe once a month and then for just a couple of hours. Never the less the CO levels have to be managed.

    In a study undertaken by the Australian Govt in 2004 ( see http://www.environment.gov.au/system...es/report9.pdf) the levels of a number of gases were monitored in >100 Australian homes using unflued heaters. The CO levels in 7 of those homes exceeded 9 ppm with one home reaching 40 ppm! This why unflued heaters should not be used in homes and range hoods should be used with mains gas stoves. LPG heating and cooking is considered safe indoors but that assumes the appliance is always working correctly, if air inlets become blocked all sorts of problems can arise.

    We had one of those unflued heaters for about 5-6 years in the early 90s coincident with SWMBO suffering from one mystery illnesses after another during winter. She used to watch TV sitting almost on top of it. When we got rid of it she slowly recovered.

    The usual CO levels of indoor and outdoor air averaged over a year is usually <1 ppm is shown below.
    This assumes no unflued heater.
    There are 5 plot lines covering the range from lowest 10% to 90% of homes (highest trace)
    Note the small peaks in morning and evening concentrations associated with human activity - mainly cooking and heating
    (from https://www.environment.gov.au/syste...-dwellings.pdf)

    The CO sensor I am using is the same as those used in domestic type detectors and only nominally starts reading at ~10 ppm! so if these sensors registers anything you know there may be a problem.
    COindoors.jpg

    RE: Readings you would get from around the barbecue.
    I assume this is outside and being hot the gas would rise rapidly and I doubt my setup would be capable of detecting the emissions. Next I go to a BBQ I will bring my detector along.

  11. #10
    Join Date
    Feb 2006
    Location
    Perth
    Posts
    27,757

    Default

    I have been using the old CO monitor in the shed and have uncovered a few issues.

    The main one is the display is too small and the alarm buzzer was too quiet so have made a new one that has the following features.

    - Power supplied to Arduino by a 12V adapter (previous one ran off 5V using a standard USB power adapter) with extra wiring/switch to turn to it on/off at the box rather than just the GPO.
    - 90 dB alarm that runs separately off the 12V supply - makes it a bit louder than running the buzzer from the Arduino board's 5V supply.
    - To switch the 12V on/off to the buzzer I used 5V relay triggered by an Arduino digital OP, protected using a diode/tranny combo. Next time for about the same money I will just buy an optoisolator switch.
    - Larger brighter LED (as compared to an LCD) display.
    - Better spec HT sensor - why ? just cos!
    - New combo CO and flammable gas sensor.
    - Ability to turn alarm off manually and set the alarm level from the box using a switch/Pot combo, rather than by reprogramming.

    I tried to get it into a smaller box but the effort goes up exponentially trying to get it into a smaller package.
    It would tidier if I used an Arduino Shield (or a dedicated circuit board) like I did for the particle counter.

    CO-mark2.jpg

    CO-mark2a.jpg

    The spare room inside was going to be for a battery but I've given up on that.

Similar Threads

  1. Is there a stud sensor that works?
    By AlexS in forum HAND TOOLS - POWERED
    Replies: 15
    Last Post: 28th February 2015, 03:23 PM
  2. Sensor to tell you when a DE bags full ?
    By JDarvall in forum WOODWORK - GENERAL
    Replies: 6
    Last Post: 14th July 2008, 12:48 AM
  3. Movement sensor software
    By Burnsy in forum COMPUTERS
    Replies: 1
    Last Post: 7th February 2008, 07:47 PM
  4. Motion sensor malfunction
    By Biko in forum PLUMBING, ELECTRICAL, HEATING, COOLING, etc
    Replies: 7
    Last Post: 28th September 2007, 09:27 AM
  5. Sensor Light
    By Garves in forum GENERAL ODDS N SODS
    Replies: 12
    Last Post: 9th June 2006, 10:42 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •