Announcement

Collapse
No announcement yet.

Project PID Silvia -Melbourne

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

  • #31
    Re: Project PID Silvia -Melbourne

    Thanks Mal. I cant recall why, but I changed that setting.
    I write down each P-I-D setting Ive been experminting with to no avail, and so far have 3 pages of A4 sheets full fo them!

    Ive autotuned a few times, and each time it has set different parameters.

    P around 120 to 400, I 390-1000, and D always around the 60 mark.

    The other problem is discrepancy or accuracy. The instructions say the controller is calibrated and should not need re-calibrating, but it is telling me the boiling point of water is 103 C at sea-level.

    Ive changed the t/c offset so it reads 100, the same as my DMM. But when I reattach the thermocouples (from electric jug water to Silvia boiler) the controller reads about 3 degrees hotter than the DMM.

    I dont know if I trust either, now. Water seems very hot, and Silvia also seems very hot.

    --Robusto




    Comment


    • #32
      Re: Project PID Silvia -Melbourne

      Hi, robusto -

      I threw in the towel today on my Auber. Because interference was suspected, I switched to a stick-on bead type t/c that is electrically isolated from brass boiler. This made the interference go away and allowed me to kill the filtering, but the temp control did not get any better. I got tired of twiddling with it and decided not to risk sending it to a customer. I will leave this particular PID to someone who is wiser than I.

      Regarding comparisons between PID and DMM boiler temp readings: I have found it almost impossible to measure same temp at any 2 spots on Silvias boiler. Depending on where the controllers t/c is located, 3C differential does not sound unusual. (IME, hottest spot is under left brew tstat screw.)

      I just finished a special order kit for a customer using a Watlow 965A. It worked very well on my bench machine, ATd itself perfectly, and has 2-line display. These can be had (used) periodically on eBay for under $50USD. If you decide to switch controllers, maybe you will want to give this model a try (specs say 240VAC mains OK).

      -- JGG

      Comment


      • #33
        Re: Project PID Silvia -Melbourne

        Thanks JGG.

        I wont quite give up yet.

        If it came to that, I would probably see what the consistent steady state overshoot is, and then make the set point lower by the same number of degrees.

        I shielded the t/c by wrapping it in kitchen aluminium foil. In any case, temporarily, its only adjacent to wiring inside the PID enclosure --- the rest of the way it comes nowhere near it.

        The Auber t/c is located under the left brew thermostat screw.
        The DMM under the right brew thermostat screw only half an inch away.

        What irks is I adjust the controller to read boiling water at 100 degrees, in sync with the DMM reading. They both happily do that, but when the thermocouples are put back on the coffee machine, they go their own way again.

        Robusto

        Comment


        • #34
          Re: Project PID Silvia -Melbourne

          I wouldnt necessarily assume that one of the boiler readings is wrong. The temperature variation on the top of the boiler is suprisingly large. My bench setup has 2 t/cs also, one on the left brew screw, the other on the central screw. Readings vary by approximately 5-6F (left screw is warmer). Only time they agree is with a stone cold Silvia.

          Hope you get this tuning figured out!

          -- JGG

          Comment


          • #35
            Re: Project PID Silvia -Melbourne

            I emailed Suyi at Auber Instruments about the problem today and Suyi responded very quickly.

            First he said to reduce I and try increasing P if that didnt work.

            Then he contacted the engineer who suggested just increasing P. But he wanted the model number before offering more advice.

            The instructions say that the Auber definitions of P-I-D are different from normal controllers, which may explain that.

            My (limited) understanding of PID, in a very glib nutshell, is that: P handles the speed of heating, D stops overshooting and I holds to the setpoint. Of course, they are all inter-related and changing one parameter changes the lot.

            I would have thought that after an hour or so,
            with no water entering or leaving the boiler, the temperature would be pretty even on top. But I could be wrong.

            Tonight I might put both thermocouples together under the
            one left hand screw.

            Assuming that fixes the discrepancy in readings ---the problem then becomes: what set point to use to get 95°C at shower screen?

            --Robusto

            Comment


            • #36
              Re: Project PID Silvia -Melbourne

              Originally posted by jggall01 link=1144669688/30#31 date=1146366277

              Regarding  comparisons between PID and DMM boiler temp readings:  I have found it almost impossible to measure same temp at any 2 spots on Silvias boiler.  Depending on where the controllers t/c is located, 3C differential does not sound unusual.  (IME, hottest spot is under left brew tstat screw.)

              -- JGG
              Ive now attached both PID and DMM t/cs under the same left-hand brew thermostat screw.  It works: Both readings are now about the same.

              Hard to believe a 13 mm change of address on Silvias boiler top could make the difference of about 3°C.

              Still twidling with PID parameters, though, in bid to eliminate idling overshoot.

              --Robusto

              Comment


              • #37
                Re: Project PID Silvia -Melbourne

                Hi Again Robusto,

                Making progress..... thats good .

                Now that you have both the controller and digital thermometer reading the same temps, you could probably give some thought to removing the thermo & t/c..... The scaling, resolution and repeatability of the controller display should be at least as good as the thermo and probably better so no need to double-up with additional redundant complexity? Dont forget the KISS principle ,

                Mal.

                Comment


                • #38
                  Re: Project PID Silvia -Melbourne

                  Hi everyone, sorry to be incommunicato, but Ive  had to rush out of town, out of state, at short notice, and Im still away, and have just got access to the internet briefly.
                  Havent had a chance to read all the posts, should be back next week (but only briefly) before dashing off again.

                  So, I  miss my coffee machine and look forward to fine tuning the Auber PID.

                  Regards, Robusto


                  Comment


                  • #39
                    Re: Project PID Silvia -Melbourne

                    Hi, Mal -

                    Against my better judgement, I dragged my Auber out the other night and fiddled with it some more. Went through a completely manual tuning procedure - crashed and burned again ;-{.

                    But I did 2 things during the manual tuning attempt that I had not done before: 1) I momentarily set P=0 which resets an offset variable (not explained terrifically in the manual); and 2) I dialed in even more digital filtering (was 4, now set at 6).

                    Then last night, since it was still hooked up, I started an autotune from dead cold, and charted it using a 2nd sensor hooked to a digital thermometer.

                    To my amazement, it tuned this time. After some oscillations immediately following the AT routine, it settled down to the 108C setpoint and stayed there within 1C.

                    Feeling very brave now, I pulled a couple of blank shots. As the attached plot shows, I got considerable overshoot (5C-8C), but then it settled right back to SP again. The shot recovery time was approaching 4 minutes, which is not great, but tolerable. (I can usually tune the Watlows to restabilize at SP no more than 2 minutes after hitting the low point from a shot).

                    I have no idea what I did differently this time from the many, many other AT attempts I made, other than what I described above with resetting P=0 and increasing the digital filtering.

                    FYI.

                    -- JGG

                    Comment


                    • #40
                      Re: Project PID Silvia -Melbourne

                      Gday JGG,

                      Hmmm, I think it comes down to a matter of having all relevant documentation on hand so that you can gain some understanding of how Auber have setup the FL PID arguments. Without that info, it really is like trying to shoot black targets in the dark, damn difficult. Maybe you can prepare an argument to the Tech Rep(s) at Auber to try and convince them to provide you with this level of documentation..... after all, you are trying to promote their product (in a way) by including them in the kits you package together :-?.

                      In a nutshell, I have no idea of whats going on either but I think you have proved that the usual interpretation of the PID parameters in most other controllers does not necessarily apply to the Auber units. Very confusing to say the least. The o/shoot you quote as achieved by the Watlow controller is in the same ball-park as my Shimaden unit and what one would expect to achieve in this application. In the end, I dont know what else you can do except pressure Auber for more and better documentation, otherwise Id not bother with them anymore.....

                      Cheers,
                      Mal.

                      Comment


                      • #41
                        Re: Project PID Silvia -Melbourne

                        Just got back home --- so much to catch up on, including finishing off installatiion of the Auber PID on the Silvia.

                        It is all attached, but the wires have to be re-routed through the black flexible pipe and the unit, which now sits alongside Silvia on the kitchen bench, has to be attached to its side.

                        Trying to tune the Auber has been a major headache, as JGG says.

                        Forget autotune.

                        Forget the advice offered by the sellers in email support (even if in good faith).

                        The advice was to INCREASE the P setting....

                        Well, I must have tried every numerical permutation and combination, spending hours and hours of trial and error, recording the setting and then going up or down for P, I and D.

                        Finally stumbled across my own settings, which appear to work well.

                        Settles to within a fraction of a degree of set point during steady-state, and overshooting is also minimal.

                        I do not have the settings with me, but will check them and post them when I get home tonight, hopefully.

                        The totally inadequate instructions say dont toy with the factory settings --- well, ignore that, too. You have to offset the thermocouple error --make that AUBER error-- by a few degrees to start with.

                        I must say, it is novel now fronting up to the Silvia at anytime, grinding coffee and hitting the brew button without temperature or time surfing.

                        Incidently, I wrapped the thermocouple wire in aluminium foil to prevent the induction of electromagnetic interference by the AC supply lines. Dont know whether that is totally necessary, but what the heck.

                        --Robusto

                        Comment


                        • #42
                          Re: Project PID Silvia -Melbourne

                          And here are the figures for the Auber PID settings which give a good steady-state performance as well as low overshoot:

                          P=55
                          I=1251
                          D=36

                          Im sure that with even more experimentation, the overshoot of 2 or so degrees can be corrected, but for now I am happy with things as they are.

                          --Robusto

                          Comment


                          • #43
                            Re: Project PID Silvia -Melbourne

                            Hi Robusto,

                            I am considering doing the exact same mod you are doing. For around $100 I think its a bargain. What are your feelings. I see kits on the internet for $100s, yet, purchasing all the parts separately seems is a lot cheaper.

                            Did you think the PID was worthwhile?

                            If all is sucessful with you, I will be ordering all the parts from auber soon.

                            Comment


                            • #44
                              Re: Project PID Silvia -Melbourne

                              For the price, or maybe at any price, I think the Auber and solid state relay works very well.
                              It certainly is a bargain and does away with waiting and surfing.

                              I like the two-line displays showing both set temperature and actual temperature in contrasting green and red digits.

                              There are also two alarm functions , but cant think of any use for them.

                              Good luck,
                              Robusto

                              Comment


                              • #45
                                Re: Project PID Silvia -Melbourne

                                Hi, Wushoes -

                                I have a little different take.

                                My experience with the Auber ultimately did result in getting it to work (sorta), but took much more effort than it should have. Not sure of your level of experience with things electronic, but tuning the Auber was a daunting task for me.

                                As folks have patiently endured hearing from me many times recently, I much prefer the Watlow line of controllers for this application. For the about the same cost as a new Auber, you can buy, on eBay, a new or used Watlow 965A or 93. My experience with both of those controllers has been better than with the Auber, particularly in regards to tuning and stability.

                                If you go this route, make sure you pick up a controller with at least one C output (DC pulse) and make sure you dont get one of the (less common) DC-powered units.

                                Best of luck. Either way, once you are up and running, you will like the PID.

                                -- JGG

                                Comment

                                Working...
                                X