Too Much Coffee

I had a crack at the pid kettle, it took one trip to the hardware store and about two hours of my time - it was easier than it sounded! Photos: www.flickr.com/jasonscheltus

Instead of ripping up the kettle I just used the mains plug and made sure the kettle stayed on. It works okay, I just have to play with the pid values.

I added a photo from the top to show the wiring basically. Brown=active Green/Yellow=Earth/Misc! Blue=Neutral.

Hi Jason,

told you it was easy, you only have to wire the PID to the SSR (2 wires) and install the SSR on the power supply line to the kettle. Oh and add a temperature sensor (usually 2 wires).

I should have my PID unit next week and I already have the TC and the SSR, so it should go smoothly from there. The only thing missing at the moment is an IP54/57 rated box.

Regards,
dsc.

i'm sure my box was IP54/57 rated... okay to be honest i banged it out of three-ply and I don't know what that rating means..

Hi Jason,

I've seen the photos:) looks a bit like a PID controller drawer:)

IP5# is a waterproof rating, I believe IP54 means that the device can be splashed with water and IP57 means it can be submerged.

Regards,
dsc.

dsc - I think you should make a few of these and send me one for testing on my kettle.

Hi sys,

there is a chance I will be making more than one, so there is a chance I'll need someone to beta-test it:)

Regards,
dsc.

dsc wrote:Hi Jason,

I've seen the photos:) looks a bit like a PID controller drawer:)

IP5# is a waterproof rating, I believe IP54 means that the device can be splashed with water and IP57 means it can be submerged.

Regards,
dsc.

yes it's a good drawer.

programming the pid is a pain in the ass. I have my P,I,D values at 0.3, 70, 20 for the k-thermocouple, but it's waaaay too slow to come up to temp, holds the temperature really well, but it's too slow to respond to a change in the system i.e. a glass of cold water in the top of the kettle as well

any suggestions?

Hi Jason,

I'd say make the P bigger as it's affects the response time directly (being the multiplier for the error, so the bigger the P and bigger the control signal change to make the output stable). The I looks a bit big which might lead to overshoot (the I is a multiplier for the sum of errors) but I guess you've got a pretty big D as well which should take care of that (D being the multiplier for the difference between the error in this moment and the previous one).

So to sum it up, yeah try making the P a bit bigger and see what happens.

Regards,
dsc.

dsc wrote: so there is a chance I'll need someone to beta-test it:)

sweet!

Jasonscheltus wrote:...t it's waaaay too slow to come up to temp, holds the temperature really well, but it's too slow to respond to a change in the system i.e. a glass of cold water in the top of the kettle as well

any suggestions?

It may be obvious, but no one has mentioned the fact that the response time is directly proportional to the efficiency of the heating element. The PID can respond perfectly, but for the heater to bring a glass of cold water up to boiling temperature may take minutes. The element will need to be very large and the contact area massive in order to reduce this lag time. Or am I missing out on something here?

bruceb wrote: mentioned the fact that the response time is directly proportional to the efficiency of the heating element. The PID can respond perfectly, but for the heater to bring a glass of cold water up to boiling temperature may take minutes. The element will need to be very large and the contact area massive in order to reduce this lag time. Or am I missing out on something here?

No you're spot on, but i did mean it took ages compared to just regularly boiling water - especially to boil the water initially.

Initially the pid would be stopping and starting the kettle even up to 70 degrees when the set-point was 90.

It's getting closer now and overshooting less.

Tom, I'm up for one of those PID kettle jobs if you're doing a batch run, can forward you some moolah for parts if you need it. Just got a Brita filter kettle which is good but opening the lid only exposes the pre filter reservoir so cooling down to brew temp takes an age.

Jason how much overshoot are you getting, and does it settle down after a while? Don't some PID controllers have some sort of error correction to get over this?

Hi Steve,

I should have a PID controller with me soon (I hope) and I'll let you know when the thing is up and running. If it works properly I'll be able to put one for you as well.

Regards,
dsc.

Sweet!

steve, after a few more tries at programming it, i found a "SV+2" variable that I had missed somehow. No more overshoot and quite quick to come up to temperature.

If I dump a lot of cold water in my kettle while it's stable at 90c to bring it down to 60c it takes 90 seconds to come back up to temp without overshoot. I don't think that;s too bad for an open boiler.