Announcement

Collapse
No announcement yet.

temperature monitor - design project.

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • temperature monitor - design project.

    Hi, Im in the process of designing a data logging temperature monitor for my friends coffee roaster as my main project for my Bachelor of Engineering Technology at CPIT in Christchurch. So I thought Id share what Im doing.

    As part of the program we have a large design project which is designed to take the year to complete. My project is to design, build and test an embedded data logging system to monitor the temperature of my friends coffee roaster (a little home built 2kg roaster). This will transfer temperature data to a PC in real time so that a temperature graph is displayed while roasting. I also am working on a method of detecting when first and second cracks occur so these can be automatically be displayed on the graph too.

    Im very aware now that there are already a few options out there which perform a similar function to what I want to do but I chose the project before I found out about them.

    Im going to use this thread to post (hopefully) regular updates and if you have any questions about what Im doing feel free to fire away.

  • #2
    Re: temperature monitor - design project.

    This sounds very exciting! I cant wait to hear more. Its projects like these that make me wish Id done Engineering.

    Comment


    • #3
      Re: temperature monitor - design project.

      Hiya

      Originally posted by 767D71796F7D7278791C0 link=1337561181/0#0 date=1337561181
      My project is to design, build and test an embedded data logging system to monitor the temperature of my friends coffee roaster (a little home built 2kg roaster). This will transfer temperature data to a PC in real time so that a temperature graph is displayed while roasting. I also am working on a method of detecting when first and second cracks occur so these can be automatically be displayed on the graph too.
      Well the temp monitoring is a common thing today with many transducers that take thermocouple or thermistor output and convert it into serial data streams for PCs. Not much opportunity for innovativeness unless your going to be designing the actual A to D circuit. There is quite a lot of amplification & signal conditioning to do in the circuitry.

      However picking up 1st and 2nd crack automatically is not common, if done at all, and would be really useful. For my GeneCafe I thought a few years ago that a stethoscope on the side of the roaster and a DSP doing real-time signal processing might be the way to go. The cracks are quite sharp and fast compared to background noise. Just record them with a mic and look at their spectrum in the freq domain - you will see what I mean.

      The other thing would be to make a tiny bean from teflon and implant into it a RF antenna out of a metal that has a high thermal expansion. From the outside send it a RF pulse. Pick up its response. As it expands its tuned RF signal will vary with the temp of the telfon bean hence being an indication of the real bean mass temperature. The real work here is in RF antenna design :-) [This last para is a free idea for you, if you patent it and make $ give me a cut :-) ]

      Best wishes
      Mike

      Comment


      • #4
        Re: temperature monitor - design project.

        Originally posted by 7477626B62686A6E6C62070 link=1337561181/2#2 date=1337597870
        As it expands its tuned RF signal will vary with the temp of the telfon bean hence being an indication of the real bean mass temperature
        And, for bonus points, make it all fit inside a Behmor - thatd be cool. I suppose RF through the drum might be problematic?

        Comment


        • #5
          Re: temperature monitor - design project.

          Hi

          Originally posted by 20303221362A530 link=1337561181/3#3 date=1337635150
          And, for bonus points, make it all fit inside a Behmor - thatd be cool. I suppose RF through the drum might be problematic?
          Could be a problem, the Behmor has some glass area though, does it not? Certaily with the Gene there would be no probs in getting it to respond to an RF signal though the large glass area surrounding the beans.

          Mike

          Comment


          • #6
            Re: temperature monitor - design project.

            Darn, I had this whole message typed out with links and pictures in the text but since Im a noob, the forum wont let me post the links.

            Thanks for the interest and input everyone. Heres a picture of the roaster:
            chirny.com/wp-content/uploads/2012/02/IMG_20120226_190921.jpg

            Originally posted by 3033262F262C2E2A2826430 link=1337561181/2#2 date=1337597870
            However picking up 1st and 2nd crack automatically is not common, if done at all, and would be really useful. For my GeneCafe I thought a few years ago that a stethoscope on the side of the roaster and a DSP doing real-time signal processing might be the way to go. The cracks are quite sharp and fast compared to background noise. Just record them with a mic and look at their spectrum in the freq domain - you will see what I mean.
            Yea, unfortunately my friends roaster (he built it himself) is pretty noisy and on an audio recording of the roast I couldnt distinguish between roaster noise and the cracks. I managed to track down a company here in Christchurch (stonka.co.nz) who manufacture coffee roasters. They have a video on youtube which makes it very easy to hear the cracks. From that video I did some basic filtering then frequency analysis on two seconds of the crack, resulting in the following:
            flickr.com/photos/63111829@N02/7253905354
            (frequency is on the vertical axis in 1kHz increments, time is on the horizontal axis in 1 second increments, the difference in colour indicates the intensity of the sound).

            What Im wanting to do now is use a microphone, then a band pass filter (only allowing frequencies between about 4kHz and 10kHz), then pass the signal through a comparator, thatll mean that when the signal exceeds a threshold it will be indicating that the cracking is occuring. Im hoping to knock together a circuit to test this out for when I go to Stonka next week.

            Originally posted by 3033262F262C2E2A2826430 link=1337561181/2#2 date=1337597870
            The other thing would be to make a tiny bean from teflon and implant into it a RF antenna out of a metal that has a high thermal expansion. From the outside send it a RF pulse. Pick up its response. As it expands its tuned RF signal will vary with the temp of the telfon bean hence being an indication of the real bean mass temperature. The real work here is in RF antenna design :-) [This last para is a free idea for you, if you patent it and make $ give me a cut :-) ] 
            Sounds interesting, but I dont have any experience with designing RF antennae, I also think that it would be difficult to detect exactly what frequency is being transmitted since it will change as the antenna length varies. But you did give me another thought - maybe I could look into using an internal temp sensor powered by a thermoelectric generator (like this one: tegpower.com/index.html) and then transmit the temperature to the microcontroller.

            Thanks for throwing your ideas out there, its helping me get a different perspective on some of these things.

            Comment


            • #7
              Re: temperature monitor - design project.

              Hi James

              Great to see that you like the audio processing idea. Much more challenging than a simple thermocouple and off-the-shelf data acquisition :-)  There are quite a few people that find it hard to pick up the cracks in roasting. This would be really good for them. Once cracks can be reliably detected one can flash a LED each time a crack occurs, once rolling first crack occurs the LED would be on fully, until first crack ends.

              The thermoelectric generator idea to power a temp sensor inside a bean or even inside the roast drum wont work. You need a temperature **differential** for it to work. The inside of the drum is hot and the outside of the drum is cool so it might work if glued to the surface of the drum.

              I can see the roaster pic shows its all metal encased so no RF through - oh well. You could power the bean/temp tranmitter with a super-capacitor if you could find one that withstands a high temperature.   

              Mike

              Comment


              • #8
                Re: temperature monitor - design project.

                Sounds like a great project!

                Originally posted by 3D363A322436393332570 link=1337561181/5#5 date=1337769978
                I managed to track down a company here in Christchurch (stonka.co.nz) who manufacture coffee roasters. They have a video on youtube which makes it very easy to hear the cracks. From that video I did some basic filtering then frequency analysis on two seconds of the crack, resulting in the following:
                Keep in mind that YouTube videos are very poor in audio quality, and the original audio itself will have been recorded on a shitty video camera mic anyway. It will serve as an ok rule of thumb, but I would recommend a proper live audio recording of the cracks at some stage.

                If a roaster in Brisbane wants to participate in this project, Im more than happy to help out with capturing a good recording for your reference.

                All the best with your project.

                Comment

                Working...
                X