Announcement

Collapse
No announcement yet.

HeatSnob Temperature Reading Disparity

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

  • HeatSnob Temperature Reading Disparity

    I reside in the U.S. and I've recently purchased a HeatSnob from your American source. In my initial set-up I've converted in Preferences from Celsius to Fahrenheit, and I purportedly am now getting a Fahrenheit temperature reading from your software on my computer screen. However, even before attempting to utilize my HeartSnob in actual coffee roasting, I notice that my HeatSnob Fahrenheit room temperature reading is always about eight degrees higher than my independent room temperature readings from several other sources, including a digital probe where the sensor is laid alongside the Heatsnob probe. Is this disparity possibly attributable to the equation your software program is employing in its Celsius to Fahrenheit conversion code? I confess I'm somewhat mystified.

  • #2
    You can calibrate the probe and adjust the temp display as described in the Roast Monitor Guide which is found in this post.


    Java "Probe what?" phile
    Toys! I must have new toys!!!

    Comment


    • #3
      Originally posted by Javaphile View Post
      You can calibrate the probe and adjust the temp display as described in the Roast Monitor Guide which is found in this post.


      Java "Probe what?" phile
      Many thanks for the quick response. Although I had previously downloaded (and read) the Roast Monitor Guide -- even before my purchase -- and have subsequently just finished rereading it, I have failed to find reference to and instructions for re-calibrating the probe. Obviously I'm missing something. But what (or where)?

      Comment


      • #4
        Put the probe in boiling water (With-out the tip touching the sides/bottom of the pan.) and note the temp displayed. If it's not 212F/100C enter the appropriate offset as per page 11/12 of the Roast Monitor Guide.


        Java "Very offset!" phile
        Toys! I must have new toys!!!

        Comment


        • #5
          Originally posted by Javaphile View Post
          Put the probe in boiling water (With-out the tip touching the sides/bottom of the pan.) and note the temp displayed. If it's not 212F/100C enter the appropriate offset as per page 11/12 of the Roast Monitor Guide.


          Java "Very offset!" phile
          I regret, very much, that I find your intent to clarify and help me out still a trifle too cryptic for me. If you're referring to the following instructions which I may utilize --

          # Meter calibration factors.
          meterAdjustMults=1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
          meterAdjustAdds=0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

          I am truly at loss to understand just how to apply this (after I've measured the possible difference in reported boiling water temperature). Please don't let me flounder further.

          Comment


          • #6
            from page 12 of the Roast Monitor Guide:

            Meter calibration factors:
            You can modify the displayed values of the 9 meters by setting values in preferences.txt. This can be used either to precisely calibrate a meter or to allow a reading other than temperature (e.g. airflow) to be scaled to suit the graphs. Each meter has a multiplier (meter1AdjustMult, meter2AdjustMult. .......etc ) and an offset (meter1AdjustAdd, meter2AdjustAdd, .........etc).

            For example, if meter 2 is measuring airflow that registers 4 mA at 0% and 20 mA at 100%, you could set:
            meterAdjustMults=1.0 6.25 1.0 1.0 1.0 1.0 1.0 1.0 1.0
            meterAdjustAdds=0.0 25.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
            so that it would display as values between 0 and 100.

            In other words, for the 2nd meter the multiplier (6.25) extends the initial range of 16 to a new range of 100 (25 - 125); while the offset (-25.0) adjusts the zero position.
            In your case if in boiling water a temperature of 220F is being displayed you would modify (The modification is shown in bold to make it easier for you to see. Note the - in front of the 8.0) this section to read:

            # Meter calibration factors.
            meterAdjustMults=1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
            meterAdjustAdds=-8.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0


            Note that the temperature being displayed doesn't have to be the exact actual temperature as what is important is the repeatability of the reading. Your system will display the same reading at the same actual temperature every time you use it. i.e. If in boiling water it reads 220F it will always read 220F in boiling water. Change the offset if the displayed reading being off bothers you, as it apparently does, but you can ignore the difference with no effect on your ability to use it for roasting.


            Java "Numbers, numbers everywhere" phile
            Last edited by Javaphile; 15 December 2016, 07:08 PM.
            Toys! I must have new toys!!!

            Comment


            • #7
              Hello - i recently started with a heatsnob and noticed the same - it was about 3c degrees greater from my other 5 or so digital temp devices.
              I calibrated it to boiling water, but then found it was still out at room temp.
              Anyway- I then calibrated it to First Crack which is when you really need it (for me this was +5C to make FC at 197) - now no looking back. I am finding it very consistent and predictable for roasting - I dont care that its out at room temp.

              Comment


              • #8
                I guess I (reluctantly) have to accept (live with) what you're both telling me -- that the HeatSnob meter's inherent inaccuracy doesn't necessarily really matter. Consistency (of inaccuracy) is what's important. Whew! Many thanks for the "enlightenment." My life will go on. Thanks, and best regards to you both. Jim

                Comment


                • #9
                  The HeatSnob is actually very accurate.


                  The real problem is that thermocouples themselves are only accurate in a very narrow window and read low and high either side of that point. Most devices are set to make it accurate in the expected window of use and will be inaccurate either side.


                  The HeatSnob utilises a chipset that is "smart" enough to keep the calibration right through the range using a logarithmic scale to suit the inherent error in the thermocouple. Add to this the thickness of the probe, the heat-sink characteristics of the mounting you have chosen and you are introducing your own levels of error too, these can be "calibrated" in the software so you can see the number you expect to see.


                  The digital probe that you are reading against will have its own level of error, generally similar to the next device you sit alongside it too as these won't use a logarithmic scale to show the correct temperature. The previous CoffeeSnob digital multimeters suffered from the same problem, they probably read the same as your "digital meter" and showed the same level of error in the same part of the scale.


                  In regard to the conversion between C and F, it's a simple calculation on the fly. Your computer will do it in real time without blinking.


                  As mentioned above by Sink_cut, calibrating to 196C and first crack is a great way to ensure that the range that is most important to you is showing you the most accurate number.

                  Comment


                  • #10
                    All thermocouples/thermometers have an inbuilt inaccuracy. The smaller the inaccuracy the higher the price. With most thermocouples/thermometers being used over a small temperature range (Like your home thermostat for a furnace or AC unit.) they're calibrated at room temp (Most commonly 20C) and people never notice the inaccuracy. Only when you start running it over hundreds of degrees does that in-built error start poking it nose out and become noticeable. The good thing is the error is consistent. Use it as-is straight out of the box, calibrate it at room temp, at boiling (Remember to compensate for atmospheric pressure if you want the best accuracy!), at first crack, or any other temperature. Other than being a stickler for having things be *right* the only other reason to need a high degree of exactness is if you're sharing profiles with others who are then using your data to try to exactly copy your roasting profile. In which case everyone in the trading circle needs to use the same calibration method.

                    The easiest way to get a group of peoples HeatSnobs calibrated to the same point is to do it in boiling water and use this calculator using the barometric pressure (If you live near where there is a certified weather reporting station like at an airport or national weather service station use their data rather than that from a home barometer as your home barometer will have a much higher degree of inaccuracy.), not the altitude to determine at what temperature water boils at your location and use that figure to determine your offset. Even at elevations of only a couple hundred meters above sea level you can have a difference of multiple degrees calculating via barometric pressure vs altitude with the barometric pressure being the more accurate of the two figures.

                    If you're not Obsessive Compulsive or sharing profiles ignore the calibration and just plug the HeatSnob in and start roasting!


                    Java "Calibrate what?!? How?!?" phile
                    Toys! I must have new toys!!!

                    Comment


                    • #11
                      *snap* Andy. Sometimes I just type too slow!


                      Java "OC" phile
                      Toys! I must have new toys!!!

                      Comment


                      • #12
                        Originally posted by Javaphile View Post
                        All thermocouples/thermometers have an inbuilt inaccuracy....
                        This discussion has been variously both interesting and useful to me -- now at this later stage, however, actually more academically interesting than useful. The perceived disparity of my readings, as you've all pointed out, need not affect much the actual HeatSnob performance. Yet I still suspect that the disparity is more due to the HeatSnob itself than the particular thermocouples used with it. I base this on the fact that I've subsequently acquired (on an earlier recommendation of Andy's in another thread) a longer third-party banana-pronged probe which coincides in reading (with the probe that came with the HeatSnob) the same eight-degree disparity that my external thermometer reported -- both at room temperature and at the freezing point in ice-laden water. I had expected to have found the two different probes attached to the HeatSnob to possibly indicate some difference in reading between the two probes, if the disparity was attributable solely to attributes of the different thermocouples.

                        Comment


                        • #13
                          I took mine straight out of the box and find its 'accuracy' better than the old one, especially in 'the zone'.

                          I haven't calibrated it to anything.

                          First crack will vary with any given bean and is nominally (for the beans in my inventory) between 196ºC-198ºC for most beans,
                          and as high as 201ºC (for a particular bean) but is repeatably accurate for each bean.

                          In a multiple roast scenario where the same sequence of roasts for the blend is repeated every time and the roasts for decaf, Single Origins and secondary blends vary, the values across the whole roast profile and first crack are entirely consistent.

                          I see more variation in real time temps than the old data logger but I put this down to a higher and more accurate level of temp sensitivity.

                          I have also seen a pyrometric effect (as seen in firing ceramics) especially after first crack, which the Heatsnob is quite happy to record and is absolutely consistent with 'temp over time' dynamics.

                          What matters is repeatability, consistency and the ability of the data logger to show the impact of atmospheric and other ambient realities....
                          info that is used to correctly modify the real time roast profile against the saved standard profile.
                          Last edited by chokkidog; 23 December 2016, 09:27 PM.

                          Comment


                          • #14
                            With thermocouples, temperature in absolute terms is not their strength, but repeatability within a narrow range of temperatures is.

                            If you are really concerned about absolute temperatures, then you will need to purchase a calibrated 'analogue' thermometer (say, in the 190C to 240C range) to insert into your roaster at the same location as the HeatSnob probe. This will then provide you with a 'number' that you can then correlate against the indication provided by the HeatSnob. You can then use this to adjust an offset as described by Andy above, to give as close to an 'actual' temperature as you are going to get without getting silly about it - Maybe this is already pretty silly...

                            mal.

                            Comment


                            • #15
                              As I stated in my most recent posting, my interest in the temperature reading disparity at this stage is solely academic. I can live readily with whatever the HeatSnob is reporting to me. As has been stated, consistency is the need.

                              Comment

                              Working...
                              X