Another Gaggia PID sensor location post
-
Shuka
- Posts: 38
- Joined: 8 years ago
Hi folks-
Srry to start a new topic on this - all the others I found are closed to new comments. Just wanted to add to the body of info:
Just added a Rex C100 Pid controller with the k-type thermocouple to a 1992 Gaggia Baby. Very similar to all classics, but perhaps a slightly different boiler shape/design. This was a 25$ Amazon kit with 25A SSR (and non-conforming color coding on the thermocouple leads: red= positive, blue = negative). So a few points:
1) there was an excellent post:
http://tomblog.firstsolo.net/index.php/ ... ure-probe/
on the impact of T sensor position on the boiler. On the basis of that post,I decided to use a location lower on the boiler wall for monitoring (with a 10k NTC resistor). However, I also didn't like the idea of measuring too far away from the group head ( as large T differences could exist that varied with time especially during warmup, as I'm an impatient bird if I can get away with it). So I added the pid control sensor to the group head flange at the bottom of the boiler. Here's a photo: the boiler with a bit of insulation around it that I installed long ago. The k-type thermocouple was mounted at the circle (thermal-epoxied on), the monitoring thermistor at the x, on the side of the boiler towards the bottom, under the insulation, also thermal-epoxied. Note that this boiler has a square cross section, and there are two u-shaped heaters in the perpendicular faces to where the sensors are; the normal thermostat is mounted on one of those faces near the turn of the "u" (at the bottom).
Fine, so why bother sharing?
Good to know that: the Rex C100 is controlling at much finer precision than a degree, based on the repeatability of the thermistor resistance reading after perturbations were settled out for a given set point.
b) there appeared to be less than 1 degree shift between the thermistor reading (eye-interpolated from a resistance/temperature chart at 5 deg. C resolution) on the side of the boilder pressumably "close" to water and the flange (not far away). This doesn't speak to absolute but rather relative accuracy between measurements at the two different locations I could see changes in temperature much more easily with the thermistor than with the C100 (merely due to the resolution of the displays I'm using), but don't see a need to use the thermistor unless I am interested in sub-1degree F resolution (I'm not). This strengthens my sense that location of the control point is really not an important parameter.
3) at the flange, I was surprised that I was seeing clear steaming water at a set point of 189F (I'm at ~1.7km altitude, where water boils about 206ish F). By dropping just 2 degrees F, I removed all steaming, then I dropped another 2F, and had a very nice espresso (medium roast). I never expected to wholly do away with a temperature offset between the brew water and the control point, but didn't expect a negative offset.
Not so good to know is that, so far with limited testing, I have to do an initial cooling flush to counter the first overshoot of the controller unless I'm willing to sit around for a long time (as after steaming). Not a big deal for me, and I'm not sure that it would be a worthwhile win to slow down the initial ramp enough to avoid this as that would likely extend the initial heating more than I like (this is used at work, and I don't wanna wait around).
Have a good one -
Shuka
Srry to start a new topic on this - all the others I found are closed to new comments. Just wanted to add to the body of info:
Just added a Rex C100 Pid controller with the k-type thermocouple to a 1992 Gaggia Baby. Very similar to all classics, but perhaps a slightly different boiler shape/design. This was a 25$ Amazon kit with 25A SSR (and non-conforming color coding on the thermocouple leads: red= positive, blue = negative). So a few points:
1) there was an excellent post:
http://tomblog.firstsolo.net/index.php/ ... ure-probe/
on the impact of T sensor position on the boiler. On the basis of that post,I decided to use a location lower on the boiler wall for monitoring (with a 10k NTC resistor). However, I also didn't like the idea of measuring too far away from the group head ( as large T differences could exist that varied with time especially during warmup, as I'm an impatient bird if I can get away with it). So I added the pid control sensor to the group head flange at the bottom of the boiler. Here's a photo: the boiler with a bit of insulation around it that I installed long ago. The k-type thermocouple was mounted at the circle (thermal-epoxied on), the monitoring thermistor at the x, on the side of the boiler towards the bottom, under the insulation, also thermal-epoxied. Note that this boiler has a square cross section, and there are two u-shaped heaters in the perpendicular faces to where the sensors are; the normal thermostat is mounted on one of those faces near the turn of the "u" (at the bottom).
Fine, so why bother sharing?
Good to know that: the Rex C100 is controlling at much finer precision than a degree, based on the repeatability of the thermistor resistance reading after perturbations were settled out for a given set point.
b) there appeared to be less than 1 degree shift between the thermistor reading (eye-interpolated from a resistance/temperature chart at 5 deg. C resolution) on the side of the boilder pressumably "close" to water and the flange (not far away). This doesn't speak to absolute but rather relative accuracy between measurements at the two different locations I could see changes in temperature much more easily with the thermistor than with the C100 (merely due to the resolution of the displays I'm using), but don't see a need to use the thermistor unless I am interested in sub-1degree F resolution (I'm not). This strengthens my sense that location of the control point is really not an important parameter.
3) at the flange, I was surprised that I was seeing clear steaming water at a set point of 189F (I'm at ~1.7km altitude, where water boils about 206ish F). By dropping just 2 degrees F, I removed all steaming, then I dropped another 2F, and had a very nice espresso (medium roast). I never expected to wholly do away with a temperature offset between the brew water and the control point, but didn't expect a negative offset.
Not so good to know is that, so far with limited testing, I have to do an initial cooling flush to counter the first overshoot of the controller unless I'm willing to sit around for a long time (as after steaming). Not a big deal for me, and I'm not sure that it would be a worthwhile win to slow down the initial ramp enough to avoid this as that would likely extend the initial heating more than I like (this is used at work, and I don't wanna wait around).
Have a good one -
Shuka
Good morning, Sunshine!
