The CSV file with all of the roast data doesn’t seem to match the data shown in RoasTime. This is a sample from one roast at the 1min mark. RoasTime shows an IBTS temp of 84.2ºC and I-ROR of 51.8ºC/m, whereas the CSV shows 82.38ºC and I-ROR of 30.04ºC/m. The bean temp/ROR is also off.
Could be, although the IBTS temp is no where near as wild as the ROR and still a few degrees off. Would be nice to know how it’s calculated/smoothed if that is the case though.
RT and RW have data in sub second intervals and there is an algorithm for selecting the “right” value for that second based off where you are in the chart, whereas the CSV you are exporting gives values at a per second rate. I can take a look at this CSV generation to make sure it is functioning as expected, but that may account for some minor temperature variations (I would not expect a 2 degree difference as you mentioned though).
We do indeed smooth ROR. It has been a while since I worked on this, but could give more details.
I forgot to mention — the data from the API does actually return the full sub-second data (2 readings per second). If I take every other item starting from the first item, then the numbers all seem to match the CSV exactly. If I take every other item starting from the second item, then the numbers are slightly different, but it’s within fractions of degrees or mostly even identical. So, it seems like whichever of the 2 sub-second data points RT/RW are using, it doesn’t make much difference.
Thanks for looking into it though! Keen to hear what you discover.
The CSV exported from Roast World seems to show pretty much the same IBTS temps as RoasTime now, e.g. 84.17ºC vs. 84.2ºC (assuming there is some rounding there), but the CSV from RoasTime seems unchanged from before (so the CSVs are showing different values now depending on where you export from).
Note: The ROR values in the CSV are still very much different to what’s shown in RoasTime
