Help with PID temp on Silvia v3

[Burkeyboy]

Hi All. I've had my v3 for about 4 years now and put a PID on about a year ago. I understand that the temp on the dial doesn't really mean much, it mainly depends on how and where the temp is being read. I've made sure that there's no steam at extraction temp, but other than that I'm just guessing. My palate isn't that great so I can't pick much difference in a few degrees. If I held my thermometer under the basket and ran the water, what temp should I get?
Sorry for the rather basic newbie question!
Also, am thinking of upgrading later in the year. I like the control that the PID can give, so was thinking of an Expobar Minore 4. Will I really notice a big improvement?
Steve.

[Dimal]

G'day Steve...

Holding a thermometer of some description "under the basket" won't help you out very much I'm afraid. The only reliable way to accurately determine what temperature your brew water is being delivered at, is to use a thermocouple (bead type) coupled with a suitable Digital MultiMeter (DMM). One of the best for this purpose is CoffeeSnobs' own DMM that Andy sells.

You then load a basket with coffee, tamp etc as you would normally do, and then lock in to the Group with the t/couple bead on top of the puck, sitting about midway between the centre of the coffee puck and the basket edge. You may have to lock the Group Handle in a bit tighter than usual to ensure that water doesn't leak around the t/couple cable. This won't hurt either the Group Gasket or the t/couple cable insulation.

Start your shot as you normally do and then observe the water temperature being indicated on the DMM. The temperature will move around quite a bit at first as the water heats up the coffee puck but will settle down after a few seconds. basically, you're aiming for a temperature on top of the puck of around 93C +/-1.0C. Just adjust the setpoint on your PID Controller until you achieve this; allowing about 15 minutes between setpoint changes for the boiler temperature to settle down before taking your next measurement. There's no need to remove the Group Handle between shots as the original puck "in situ" should provide enough resistance to water flow such that the measurements you take will be fairly close.

And that, as they say, is it.

Regarding upgrading to a Minore or any other similar machine, you may not realise better quality coffee in the cup but it will be a much more convenient and easy process, given that Prosumer machines are basically miniature versions of full blown commercial machines where 'ease of use' is a priority. The quality of the coffee in the cup at the end of the day, is dictated firstly by the quality and freshness of the coffee you use and the quality of the grinder you have, to produce ground coffee of consistent size (as little as possible of fines/dust).

Hope some of this is helpful mate...

Mal.

[Burkeyboy]

Thanks for your detailed reply Mal, will look into it as you described. Regarding an upgrade, quality in the cup is my main concern, not ease of use or speed. I use a Behmor roaster and have a decent Pavoni grinder, so maybe it's best to try and fine tune my methods and read up more on roasting blends etc.
Thanks again, Steve.

[axs]

Hi Mal, thanks for your great instructions. I have PIDed Silvia with thermocouple mounted under the screw of original thermostat. Test shows 93C in coffee puck when PID set to 104C. I think it’s good enough.


Also may be you can suggest a good PID settings?
That’s what I have now:
Su = 0104
P = 01.1
I = 0060
D = 015
Souf = 0.2
ot = 002


but it gives me a big overshot.
Any suggestions, improvements?


Thanks in advance, Alex.

[Dimal]

G'day Alex...

Any chance of linking to the controller Manufacturer's Spec' Sheet, or User Manual perhaps? Or maybe just the brand and model of the PID Controller you're using?

Something that might help you out, is if the controller has an Auto-Tune mode. If so, use this as per instructions in the manual, which should get you pretty close to something that ought to work reasonably well. Fine tuning can occur after this if required...

Mal.

[axs]

Hi Mal,


Thanks for a quick reply.


I’ve done my setup with XMT7100 from DX (exactly the same as Auber pid) only single boiler controller, no steam. All cost around $30
I’m using default values for Auber from "Operating the Rancilio Silvia after PID kit modification".
Tried auto tuning but values looks a bit strange. It should work for initial start-up but not good for recovering between shots.
I think I should play a bit with default values to reduce overshot which probably will increase recovery time.


Regards,
Alex.

[koshari]

try running the auto tune again and let it run for the full 2 and a half cycles,

[axs]

Thanks, will try it.
If I understand correctly, start auto tune with first shot (when machine heat up) and stop after third one without waiting temperature stabilized?
Or start auto tune when it's still cold?

[koshari]

what the auto tune is trying to determine is the heating rate and the cooling rates (which will be close to constants) , it does this from sampling the rate of change with the element on and off, you really dont want to be pulling a shot while the autotune is happing. let the autoune heat the boiler up to the set point, overshoot, decay back below the setpoint, over shoot again and then decay, in this time it will learn the rates of changes and apply suitable PID settings.

each heat and decay phase is a cycle.

[axs]

Then you're talking about normal auto tuning:
"The auto-tuning function will use the on/offmode to heat up the system until it passes the set point. Then let it cool down.
It will repeat this about three times. ... if you stop the auto-tuning the previous PID parameters values are resumed."

Done it before and don't really like result.
I'll try it again and see what new values will be.

[Dimal]

Done it before and don't really like result.

What didn't you like?

Mal.

[axs]

I didn’t like previous auto tune values because it was oscillating a lot around set point before settle down. I think I know why, it was my fault . I’ve done auto tune without priming the boiler. It was some air inside which affect readings.

Now auto tune came up with following values:
P - 01.8
I – 0184
D – 046
Temperature rise to 110c then drops to 104c (my set value) and stays there.

Then another one starting with hot 90c
P - 01.6
I – 0288
D – 072
That one even better: rise to 107c then drops to 104c and stays there.

I think it’s perfect
Will check tomorrow from cold start.

[Dimal]

Good to hear mate...

As with all things, always best to take your time and in logical sequence. Let us know how it works out tomorrow...
There may still be a little room for improvement to control the initial overshoot.

Mal.

[axs]

Hi Mal,
Not tomorrow, today! Can’t spend a day without a shot
Unfortunately there’s no room to improvement, no overshot at all.

Ran it from cold and here’s the timetable for SV = 104c:
Temp C time m:s
26 00:00
90 04:00
97 05:00
102 06:00
104 07:00
105 07:25
106 08:00
106 12:00
105 13:00
105 16:00
104 17:00
104 30:00

Pull a beautiful double here and checked recovery:

86 00:00
104 00:35
105 01:10
104 06:30 and stayed there till next one!!!

I’m happy with results and will keep these settings.
I guess it depends on number of factors such as ambient temperature during auto tune, boiler insulation, human error (it’s me), etc.
I think it’s better to start auto tune at ~90C (where it drops after pulling a shot) and let it do its job.
Just my 2c (two cents, it keeps converting cents to cracks)

[Dimal]

Excellent...

See, there was never any reason to doubt what your controller was capable of. Happy brewing "axs"...

Mal.
P.S.
The only thing you may have to change as the seasons come and go, is the Setpoint Variable. There should be no need to touch anything else...