No announcement yet.

Civilizing the i-Roast2 (Part 1)

  • Filter
  • Time
  • Show
Clear All
new posts

  • Civilizing the i-Roast2 (Part 1)

    I got an iRoast2 last October (from Les). Id already done some reading
    about it, so wasnt surprised to find that, like many others, it ran hotter
    than programmed. I figured I could live with that, and just make
    corrections programming it.

    Reading around, it quickly became obvious that performance is related to
    voltage, and that one of the reasons they run too hot out of the box is
    that theyre designed for Europe, 230V. There are many reports, from Aus
    and the UK, of them running too hot on 240V supplies (e.g. at home,
    typically over 242V). So I found some profiles used in such conditions,
    and got quite good results.

    I also found mention, on TMC (, of a modification that
    some people were making to fix this problem. This involved changing a
    resistor in the base of the unit. The resistor is attached to a circuit
    board, inserted into a white nylon connector for reasonably easy removal.
    It can be accessed by removing the inner three screws in the base, which
    need a U-shaped screwdriver (such as in Dick Smith cat number T0098, mobile
    phone toolkit). I got some information from Gouezeri (thanks!) at TMC
    about how to do the replacement. For anyone with electronics experience,
    its a trivial operation. For those less competent (like me), its not
    hard, but there are a couple of gothchas worth looking out for. If you
    want to try it, send me a PM and Ill pass on the information I have,
    including what Ive figured out in the process of doing it.

    I resisted the temptation to tinker for quite some time. The thing worked,
    after all. But then came Harrar. I did a lousy roast, twice. Then I
    discovered (on CS, of course) that it likes a nice long roast. So I tried to
    make my profile longer, and couldnt. Which got me annoyed enough to try
    out the mod, and do some measurements in the process. In particular, to
    try to correlate changes with results in the cup, to be convinced that the
    mod made an improvement, not just a difference

    In short, Ive concluded that swapping the installed 1000 ohm resistor for
    an 1800 ohm resistor, in my machine, effects an enormous improvement. If
    thats all you want to know, skip to the end (next post) and look at the roast results.

    What Ill do now is summarize some measurements with the original
    resistor, a first try with a 2200 ohm resistor, and then various tests with
    the 1k8 resistor.

    I will show temps in this form:
    So, 221 225 235 means a thermocouple measurement of 221C, the iR2
    display showing 225C, and programmed at 235C.


    This can be made to work reasonably well with the profile I mentioned in
    This gives a quite good roast, with reasonable predictability of second
    crack, and quite consistent results.

    But sometimes you want to build up more slowly, such as with Harrar.

    I tried to do this by stretching the 160 component of the above to 9 mins.
    I tried
    160 / 9 mins; 220 / 1 min ; / 230 1 min
    trying to get a longer buildup before boosting to get SC.

    It didnt work -- second crack came at about 7:30. Clearly too much heat
    going in.

    I verified this by running the extended profile without any beans (ie with
    max airflow, which should tend to reduce heat buildup and keep the chamber
    cooler). This showed, at the end of the 9 min 160C period,
    250 209 160
    At the end, at 11 mins:
    268 228 230

    CONCLUSION: The iR2, as it is supplied in Australia and the UK, runs too
    hot out of the box. It can be used to give very good roasts, but in a
    manner reminiscent of the earlier, fixed-profile models.

    This makes the provison of ten profiles pointless. I know that a lot of
    people do manage to use different profiles, and get roasts near its max of
    15 mins, so I set out to find out how I can get this flexibility on mine. I
    decided that the best way to get some sense out of this machine was to try
    this resistor change.

    In the UK, the most commonly used values were in the range 2 - 2.7K.
    So I started with 2200 ohms (2k2).

    Here is a summary of measurements taken at 2k2.


    As above (160C/9 mins): after 6 minutes, 195 209 160
    With 210C/7mins: after 6 minutes 216 225 210

    So this is looking promising, at low and medium temps. Clearly less
    heat being generated.

    Next test: can we make it run really hot, and get a dark roast?

    Tried with 250C/15mins:
    after 12 mins (no beans), 215 245 250
    which suggests its running out of puff i.e. maybe we have cut back the heat flow
    too much.

    So tried with 150g beans: the temp ramps up quite nicely at 6 mins to
    244 245 250
    but then the temp actually starts to drop, and at 9 mins,
    224 245 250
    (the roast did actually struggle through to SC at 10 mins, and was quite
    nice, but this isnt whats expected from a sustained max temp blast!)

    Conclusion: 2k2 improves the situation, but goes too far.

    Next step: try a lower value resistor. A bit of considered thought
    suggests 1800 ohms (1k8).

    [well I exceeded max post length to be continued ... ]

  • #2
    Civilizing the i-Roast2 (Part 2)

    Next step: try a lower value resistor. A bit of considered thought
    suggests 1800 ohms (1k8).


    Ive done a lot of testing at this value. Here are a few results.

    . test with no beans at 160C
    160C/9 mins: after 9 minutes, 209 209 160
    (compare with above sets of 160C figures -- now we have matched actual
    and reported temps, but its clear that the iR2 aims higher when
    programmed ats lowest setting of 160)

    . tests of actual roasts

    Ive now done quite a few of these over the last three weeks, with
    encouraging results. Here are some examples:

    PNG "A", using preset1 (195/3mins; 220/4mins; 235/2mins)
    after 3 mins: 195 221 195
    after 7 mins: 222 228 220
    after 9 mins: 233 235 235
    which gave a nice lightish roast with just a snap or two of SC.

    ZAMBIA AA, using 210/6 mins; 220/2 mins; 230/3 mins; 250/ 4 mins
    after 6 mins: 212 225 210
    after 8 mins: 217 228 220
    after 11 mins: 230 235 230
    at 11:30: 235 248 250 (stopped here, about 10 secs into SC)

    COSTA RICA TARRAZU 160/5 mins; 200/4 mins; 220/3 mins; 250/3 mins
    (the idea here is to get a slow buildup before increasing heat)
    after 5 mins: 200 209 160
    after 9 mins: 216 221 200
    after 12 mins: 225 228 220
    after 13 mins: 236 245 250 (stopped here, about 15 secs into SC)

    So we are getting much better correlation between actual, displayed and
    programmed temperatures.

    Of course the real test is in the cup. Ive been looking for signs that
    these longer roast times are baking rather than roasting, but so far we
    have been very impressed with the results. Harrar, for example, was
    greatly improved over previous efforts, with much more depth of flavour.

    . test of "really dark roast" 250C
    (just to make sure heat supply carries through to max setting)

    CATIMORE 250C / 12 mins
    after 6 mins: 250 245 250
    remained stable at TC readings of 245-250 from here onwards

    First crack was 5:10 to 6:00, then went through a vigorous rolling second crack
    from 9:10 to 10:00, when the beans were dark and oily. Stopped at 10:40.
    A nice, usable dark roast.

    Observations suggest that, for this configuration, a typical TC reading for
    first crack is 215-220C, and for SC, around 235C. This is with the TC near
    the bottom of the chamber, about 2/3rds towards the centre, and using the
    slower profiles.


    The iR2 is now giving the sort of flexibility I was hoping for when I
    bought the machine. I can create profiles with genuinely different
    behaviours. PRESET1 now fairly reliably produces a light roast, finishing
    near second crack -- for anyone so inclined, this (or a slight variant)
    could be used by just setting it going and coming back later.

    It is a roaster capable of producing a spectrum of roasts, with some
    predictabilty of first and second crack points. In particular, I can now
    program a slower build-up for beans such as Harrar.

    Its not perfect, but Im now very happy with it. I think it represents
    very good value for money.


    • #3
      Re: Civilizing the i-Roast2 (Part 1)

      Sounds like some good investigation and a good result in the cup(which is all that matters )


      • #4
        Re: Civilizing the i-Roast2 (Part 1)

        Hi Hazbean,
        Ive been trying various resistor changes and am getting better results as well.
        The best result so far is with 2.5k.
        I did a roast with Zambia AA to compare with your figures.
        My roast settings were: 204/6 mins; 220/2 mins; 230/3 mins; 240/4 mins.
        Using your reporting sequence: TC Displayed Programmed.
        After  6mins  215  219  204
        After  8mins  222  228  220
        After 11mins 231  235  230
        Same as yours SC kicked in just before 11mins so I hit COOL at 11:15.

        Still has trouble with lower settings but now seems good higher up.
        Ill try again with a lower temp setting first and see how that goes.
        Or maybe a higher temp 1st since setting at 220 was fine.Now Im confused.
        Then Ill try a higher resistance or a single resistor rather than the pot and 1k in series.
        At least nothing has gone KABOOM yet so thats good.



        • #5
          Re: Civilizing the i-Roast2 (Part 1)

          Looks good!

          IME lower temps usually seem to have somewhat anomalous behaviour.
          Im happy to see a profile that has all three meaasurements in close
          agreement after a reasonable time. You seem to pretty close to
          that. It will be interesting to see how you go with a higher value R --
          I wonder if yours will end up not putting out enough heat, the way
          mine did.


          • #6
            Re: Civilizing the i-Roast2 (Part 1)

            Still no Kaboom.

            I tried 2.7k and your settings for the Zambia AA again.
            210/6min; 220/2min; 220/3min; 240/4min.
            After  2min  187  221  210
            After  4min  212  225  210
            After  6min  219  225  210
            After  8min  221  228  220
            After 10min 230  232  230
            After 12min 242  245  240
            Hit cooling at 13mins.
            FC started 6:15 to 6:30 mins.
            SC started 12:30 mins.
            Getting better at higher settings still high for lower settings.
            Ill try again with 196c as first setting as that only got up to 204c with a previous roast.
            If that works then Ill be very happy.



            • #7
              Re: Civilizing the i-Roast2 (Part 1)

              Interesting that youre up around the 2.7K mark for your resistor -- that
              value has also been reported by people in the UK with similar high line
              voltages to yours (which is around 245VAC IIRC). That suggests a pretty
              high sensitivity to line voltage as Im using 1.8K at 240-242V.

              I think youll get a somewhat lower result by setting temp to 160C eg here
              is another recent profile I did earlier this week:

              YIRGACHEFFE 160/5 mins; 200/4 mins; 220/3 mins; 250/3 mins
              (same "slow" profile as earlier)
              after 5 mins: 200 211 160 cf 200 209 160 on same profile above
              after 9 mins: 216 223 200 cf 216 221 200
              after 12 mins: 224 228 220 cf 225 228 220
              i.e. essentially the same as before.

              Keep it going ...


              • #8
                Re: Civilizing the i-Roast2 (Part 1)

                Hello HZ,
                Oh boy its a fussy little beast isnt it.Yep here the voltages are consistantly around 244/245.
                No wonder a lot of users get so frustrated.
                Ill try 160c tomorrow and if I get about 200c to 210c then Ill be very satified.
                  The only trouble is all this roasted coffee I now have.Oh well I can give some to a mate who
                has a low end fully auto.He tried some of an earlier roast and is still raving.
                So either it was really good or he and his wife are easily pleased.Im going with it was
                really good.
                 Can you suggest a similar bean from the list I pmd you as Ive run out of Yirgee.
                Also just as a rough guide when did FC and SC occur?

                Many thanks,


                • #9
                  Re: Civilizing the i-Roast2 (Part 1)

                  FC was 6:00 to 7:10, SC underway at 12:30, I let it go for 10/15 secs, which as Im
                  beginning to try it, I think was too long. What I might try next time is
                  to stop at 12:05 (as the profile ramps up at 12:00, this should inject enough
                  heat to make sure its on the onset of SC, maybe a few snaps into it).

                  IMVHO Yirg is Yirg and if I had run out Id go out and buy more
                  From your list, Id try Red Sea blend (but I havent roasted any of
                  mine yet, I have only 1kg and still plucking up the courage).


                  • #10
                    Re: Civilizing the i-Roast2 (Part 1)

                    Originally posted by hazbean link=1176777753/0#8 date=1177764018
                    IMVHO Yirg is Yirg and if I had run out Id go out and buy more
                    I have to agree.Ill be hunting some down early in the week.
                    Yep the Red Sea Blend will be kept safe until Im happy with the temp settings.
                    Much too good to throw on the funeral pyre.



                    • #11
                      Re: Civilizing the i-Roast2 (Part 1)

                      Hi Hazbean,
                      Back again and very happy now.
                      I tried again with Costa Rica Tarazu and 160/6 mins; 220/2 mins; 230/3 mins; 240/4 mins.
                      after 3 mins: 197 209 160
                      after 4 mins: 202 209 160
                      after 5 mins: 204 211 160
                      after 6 mins: 205 211 160
                                                              FC approx started 7:45 in just after temp increase to 220C
                      after 7 mins: 219 225 220
                      after 8 mins: 222 228 220

                      after 9 mins: 228 235 230
                      after 10 mins: 231 235 230
                      after 11 mins: 231 235 230

                      after 12 mins: 240 245 240
                      stopped at about 12:30 about 15 to 20 secs into SC

                      Now this I can live with.You were right about setting to 160c.It really aims much higher.
                      Ill try 170 and 180 first off to see if its the same.Ill use the beans I bought locall as I dont
                      have much emotional attachment to them

                      Getting nearer....


                      • #12
                        Re: Civilizing the i-Roast2 (Part 1)

                        Its me again,
                           I tried 3 roasts in the Iroast yesterday with identical beans and profile.The only variable was the value of the resistor.
                          The profile I used: 160/5min; 200/4min; 220/3min; 250/3min.
                        Same order as before    TC   Display   Prog
                        With 2.4K..After  5mins  208     211      160
                                       After  9mins  220     219      200
                                       After 11mins 230     228      220

                        With 2.5K..After  5mins  206     211      160
                                       After  9mins  221     221      200
                                       After 11mins  226    225      220

                        With 2.6K..After  5mins   199    221      160
                                       After  9mins   204    221      200
                                       After 12mins  218    229      220

                        Im still getting better results with the higher settings.I didnt try 2.7K as it was getting late.Ill try
                        again tomorrow and see how that goes I may try 2.8K.Hell I have plenty of beans to offer up t othe great God of coffee.



                        • #13
                          Re: Civilizing the i-Roast2 (Part 1)

                          I see youre still at it Ken!

                          Just for interest here are results for a roast I did just an hour
                          ago, using exactly the profile you used (and similar TC posn I think),
                          with my 1.8K resistor at 242VAC:

                          5 mins 190 209 160
                          9 mins 211 219 200
                          11 mins 219 225 220
                          12 mins 220 225 220
                          14 mins 233 248 250 (stopped here)

                          Beans were Yemen Ismaili. These must be tough so-and-sos -- Ive never had
                          to take a roast to 14 mins before, and it wasnt far into SC then. Unfortunate
                          thing about these is that they need to be screened a few at a time while putting
                          them away, looking for rocks -- I ended up eating about 30 in the process,
                          really nice flavours.


                          • #14
                            Re: Civilizing the i-Roast2 (Part 1)

                            Ken, would be interested to know what resistor value you settled on
                            in the end.

                            Ive been keeping an eye on profiles as the weather gets cooler, wondering
                            if there would be much difference. So far, very little difference with ambient down to
                            12C cf low to mid 20s for most earlier roasts.

                            Also just did a test using 100g instead of the normal 150. Wouldnt normally
                            do that, but a friend has asked me to do some 100g batches for him to compare
                            flavour profile with the 100g lots he does in his popper. The temp profile shifted
                            down slightly, but still worked well. For example, after 10 mins TC read 227C,
                            cf 230C for a previous 150g batch using the same profile. Reached SC at 12:15,
                            cf 11:40 for the 150g batch.


                            • #15
                              Re: Civilizing the i-Roast2 (Part 1)

                              Hi Hazbean,
                               I did 4 roasts recently using same beans and 4 different resistor settings.
                              2.7K still gives me the closest results in terms of TC temp and Prog temp.
                              I guess in the long run the TC indicator is best and if I use that to plan the
                              roasts all will be OK.
                               The best thing as you found out has been lowering the temps so you can achieve
                              a roast of reasonable times.Not the "quickie" 4 or 5 min total from before.
                              The temp hasnt dropped much yet so I havent noticed any changes.Ill keep
                              the pot connected so I can still fiddle and adjust if needed.
                               Im now confident to give the red Sea blend and Altitude Premium blend a go.
                              At least now I dont think Ill kill them.