Hi there:

Scace device pressure gauges have an accuracy spec of +- 1.5% of full-scale, as delivered from Wika (the manufacturer). That's about 0.2 bar, which is good enough for coffee. It's reasonable to see pressure drop between the pressure gauge supplied on a machine compared to the Scace value because there is usually substantial head loss due to internal plumbing structure (pipes, pipe bends, gicleurs, etc).
The pressure value directly above the bed of coffee is the pressure value that we are actually interested in, because it's a boundary condition that affects brewing.

Pressure calibration - For an end user, accuracy of pressure gauges is hard to manage. I check them when Scace devices come in for service, but most people aren't in a position to do that. Simple things that you can look for to ascertain whether your mechanical gauge is performing reasonably are: Is the gauge returning to zero when no pressure is applied? If not, then your gauge needs service. Is your Scace gauge reading higher than the machine's gauge. If so, then one of the two gauges is suspect and you might wanna investigate further. I will say that as a general rule, the WIKA gauges that I use are pretty robust. If it hasn't been beaten around or frozen, then it's probably ok.

Temperature calibration - Thermocouples have one real advantage over other temperature sensors when it comes to field devices in that if they function, then they're pretty sure to be functioning correctly. Scace devices use Type T thermocouples with special wire that gives temperature uncertainty of +- 0.6 Degrees C at 93 C. That accuracy gets added to the accuracy of the readout device. Fluke 52 readouts have uncertainty of +- 0.6 degrees as well. The accuracy of the sensor and readout are "uncorrelated", so the uncertainties are combined as the square root of the sum of the squares, which gives a combined accuracy of 0.8 degrees C. Of the two components in the measurement system, I think that as long as the probe wires aren't unduly wound and unwound and the device isn't beaten around, then the probe will maintain accuracy. There are a variety of reasons why you might want to calibrate a readout and both Fluke and Thermoworks will do it for you- at a cost.

You can actually do your own calibration of the combined system of sensor and readout - AND get an accuracy of 0.06 degrees for the calibration (although you'll need to assess for yourself how long that cal is good for at that level of accuracy). Here's how to do it:

Get a very large steaming pitcher or tall sauce pan - one that can hold a couple liters of water at least. Fill it half full of distilled water (not De-ionized). Put it on a non-induction stove (induction will affect the answer) and heat on high so that it boils vigorously. Suspend the Scace device upside down over the water (filter basket with sensing element and flowmeter, but not portafilter with pressure gauge), but inside the pitcher or pan. You can support it by looping some string around the filter flowmeter and tying it to something above the pitcher or saucepan, like a screwdriver placed across the top. Snake the thermouple leads outside of the pitcher / saucepan and cover the pitcher / saucepan with aluminum foil. When the water starts boiling, read the Scace device temperature. It should be very stable. The region above the boiling water will be filled with both steam and condensing steam (fine water mist). The heat released by condensation keeps the steam hot, and makes the region from boiling water to foil cover extremely temperature stable - better than 0.06 Deg. C. You'll know that things are working right if the measured temperature is stable as a rock. If not, then you have part of the device immersed in the water or the cover isn't fitting well over the pitcher / saucepan. Finally, get the local barometric pressure by logging onto the local airport and getting their barametric pressure reading - which is used for correcting airplane barometers to local conditions. This number will be reasonably good for you to use, as long as the airport isn't on top of a mountain while you are in a valley (or vice versa). Look on the web for a boiling point conversion program such as this one: https://www.thermoworks.com/bpcalc. That should give you the local boiling point corrected for local barometric pressure. Compare the answer given by your readout device and the one given by the boiling point calculator to find out how close your Scace device is to reality.

Once you have this info, use it as a correction. For example, let's say that you read 99.5 Degrees C and the boiling point calculator says that you oughtta read 99.8. When you shove your Scace into an espresso machine, add 0.3Deg C to the measurement to get the corrected value. FWIW, if you do the calibration carefully, your Scace's accuracy will be within 0.06 Deg C for a short while after your cal. I hesitate to say how long it will be good for. That depends on how careful you are with handling the device. Wire strain from winding the wires up, banging it around, and a host of other things can eventually degrade accuracy. But if you recalibrate in a month and the offset is the same, then you can try recalibrating every 2 months, and see how that goes etc. If you plot the results, you'll get documentation of how good your device is over how long. Then you can recal as needed and have more confidence in your measurements.

Hope that answers your questions.

-Greg Scace