Hi friends,

Just got some PID stuff to put on my Gaggia Twin - I believe it'll be the first PID'ed twin, since I didn't find on the web any reports from such setup.

I saw on the web something about being possible to manually adjust P, I and D parameters, in a manner that resistance would rapidly kick in during extraction, thus ensuring minimal intrashot temperature drop.

Question is if it's possible and how to accomplish this.

Thanks a lot.

Leonardo

I can't find it but there is a post here with a triple PID'd machine with additional controls set up to adjust every paramater of the shot. With the proper PID spec you can control the PID from a computer via RS232 or some such monitor/control setup. Just depends on how far you want to go.
KDM

Basically, the PID can only turn the heating element on or off. The ability to change the temperature during the shot is limited by the response time of the heating element, distance between boiler and coffee, amount of heat loss or gain from the group, etc. Since an extraction only lasts around 25 to 35 seconds normally, that is not much time for the boiler to respond.

Now... If you are feeding the boiler pre-heated water (fed by a pipe passing through the steam boiler so that the output temperature is constant and above the desired brew temperature), then you could use a mixing solenoid the blend that water with cooler water (like the warmed water, pre-brew boiler from the feed pipe) and control the mixing valve by a PID. ..simple.

I agree completely with Randy's assessment of the problem with overcoming thermal lag. The Gaggia has a relatively low mass boiler and high wattage heaters, though, so I think it is possible you'd see the effects of the heater near the end of the shot.

Generally, to get the fastest response from your PID when a temperature drop is sensed, you will want a high proportional gain and a high derivative gain. To increase derivative gain normally means increasing the "D" or "rA" rate parameter (units usually seconds or minutes).

Increasing proportional gain usually means lowering "Pb" or "band" (units are degrees), or raising "P" if it is a pure gain (unitless). Sorry for making it so complicated, but different manufacturers define the P, I, and D parameters differently.

Integral, or reset, action does not have much effect in this situation.

Jim

On my PID controller, I believe there is a variable (I think it's referred to as dwell, but I could be wrong), which allows the controller to recover more quickly as far as moving back to the set temperature. I will have to look back through my notes, but I believe Auber mentioned it once when I was asking them questions via e-mail. My understanding is that when the value is larger, the PID controller is able to respond more quickly to changes.

If your PID controller has this same value, then increasing it may allow the controller to respond more quickly (i.e. intrashot) to drops in temperature at the boiler on account of pulling a shot.

Ray

duke-one wrote:I can't find it but there is a post here with a triple PID'd machine with additional controls set up to adjust every paramater of the shot. With the proper PID spec you can control the PID from a computer via RS232 or some such monitor/control setup. Just depends on how far you want to go.
KDM

Well, I'm pretty sure that I won't go as far.

Randy G. wrote:Basically, the PID can only turn the heating element on or off. The ability to change the temperature during the shot is limited by the response time of the heating element, distance between boiler and coffee, amount of heat loss or gain from the group, etc. Since an extraction only lasts around 25 to 35 seconds normally, that is not much time for the boiler to respond.

Now... If you are feeding the boiler pre-heated water (fed by a pipe passing through the steam boiler so that the output temperature is constant and above the desired brew temperature), then you could use a mixing solenoid the blend that water with cooler water (like the warmed water, pre-brew boiler from the feed pipe) and control the mixing valve by a PID. ..simple.

Yes...VEEEEEEEERY simple. Huh??

Hi all,

First, thanks for the attention.

Second, I'm really starting to agree that maybe it won't give the expected result. The extraction time is very short and depending on the termocouple location it can take several seconds to the PID to realize the temp drop.

Anyway I'll try to adjust the parameters as suggested and compare it with the autotune adjusted values, just to see if it makes some difference. At least we will have a response.

[]'s

Leo

leopm wrote:Anyway I'll try to adjust the parameters as suggested and compare it with the autotune adjusted values, just to see if it makes some difference. At least we will have a response.

[]'s

Leo

If you take the time to properly tune your PID for your machine, it will always be much better than autotune which really doesn't work for our uses. The problem with autotune is the PID is only reading a stable boiler while just sitting. It isn't measuring the variations during a shot. The best chance of setting up your PID is either find someone with the same machine who's done the work and start with their findings or rent (or buy) a scace device and meter. There's a great article about setting PIDs written by Greg Scace somewhere on the web.

Thanks Mitch,

Do you think is possible to use a meter in line with a termocouple, performing measurements in the shower screen? I could pass the termocouple by inside a spouted PF and keep it very close to the screen...

Leo