The Scace Device is great for its intended purpose, but that purpose is not cost accessible to me. Therefore, I do what I did and if you revisit the data that I mentioned in the edited post above, you will understand why I am very pleased that the claims of a steady infusion temperature by the Londinium crew are suggested by my first determination. After a 17 second ramp up in the temperature of the infusion stream from 122°F to 200°F, the effluent stream stayed steady, ±1°F for the rest of the infusion*. I am satisfied on this run with the assumption that the temperature stability of the late infusion suggests a similar stability to the temperature that reaches the puck for the entire infusion. My confidence in this assumption will increase if additional runs replicate this tracing. Others may require a temperature assay above the puck to reify such temperature stability; to me it is not worth drilling into a basket. The additional assumption that the effluent stream reaches thermal equilibrium with the dosing inflow once the heat content of the coffee and PF have also reached equilibrium (that is, 200-201°F is what the puck is 'seeing' during the entire infusion) is a less defensible assumption to make, but I am also comfortable with this in lieu of drilling and gluing. I am so far reassured that the performance of this machine corresponds to the claims made for it by the company. I have no problems with the quality of the shots that cannot be explained by my own inexperience that would require a more precise temperature investigation. As long as subsequent runs do not show unexpected aberrations, I look forward to retiring this piece of apparatus. Oh yes, I hope the plastic taste has been flushed away, and after a few more runs at .9 bar, I will return it to .85 bar, where erics had it, then I will make a few more runs with the housing replaced, which might necessitate going down to .8 bar, or part way.

*This range is anecdotal variation and is not meant to be taken as formal standard deviation or variance.

Also see:

Olympia Cremina Temperature Study Part 3
Cracking into the Extraction: when and where the espresso puck brews.

for significant temperature differences depending on where you measure. Dan's data is so different from JimG's that it needs more investigation or explanation. One can ignore this stuff, but at the risk of "why does my 200F (measured above with a Scace, same coffee, same grind) taste so different than yours?".

Marshall wrote:Right. It's an approximation. The point was to create a product that would be easy to use and provide useful results. Its adoption all over the world and use as the temperature measurement tool to qualify espresso machines for the World Barista Championship suggests Greg was successful.

The point of the Scace is that it's a standard, which is no surprise given Greg is an NIST engineer. Is it the "best" way to measure espresso machine temperature profiles? Maybe, maybe not. There are numerous ways to do this with different types of probes, probe locations, shot protocols, using actual coffee vs. simulated flow, etc. The Scace device is designed with a specific measurement in mind, and they are all built exactly the same and hence measure exactly the same thing, no matter what machine you're using. Two people can test and calibrate completely different machines with separate Scace devices following the standard WBC protocol and get results that are directly comparable. There is no way to do this with all the various DIY solutions. That's what makes the Scace device so useful.

Bob_McBob wrote:Two people can test and calibrate completely different machines with separate Scace devices following the standard WBC protocol and get results that are directly comparable.

Comparable up to the point where these two people start with grounds at different temperatures and then get not very comparable tastes. The Scace also has a slow response time that can mask short-term temp differences.

I have no argument with a Scace being a useful standard (with some limitations).

there are not many limitations to the Scace device itself. The Scace device is a temperature sensor and needs an external measurement system. These external devices range in accuracy, and thermal time constant more then you might think. As long as two people are the using the same system (and process) the results are comparable. The Temperature sensor used in the Scace is pretty fast, about 0.25 sec. Your system may vary since the Scace requires other components.

The most import results come from measuring temp without coffee. Coffee is too large of a variable for the information to be used for other than your own use.

The Scace is designed to measure the performance of a machine and not how the machine interacts or performs with coffee. There have been quite a few measuring devices over the years with most made by machine manufactures and not for sale to the general public. The Scace was designed (and a testing procedure) for measuring and comparing water temperature performance for the machine qualification of the WBC sponsorship. The measurement procedure/process is as important as the device so if you are not using both it is hard to compare results (unless you agree on a process). Most using the Scace want to use it to say my machine is set to "XXX". Once you start the path down accurate temperature measurement you will find out that it is not that easy to do....... with precision and accuracy. From my experience, most who say they have an accuracy to within 0.5 degrees do not have the equipment necessary to measure to this precision. They have a meter that has a decimal point but a system/process that does not support this amount of precision.

Temperature is very very time dependent and you must pick "a time" to take a measurement.

BillC

My main concern is what the water temp is when it first hits the top of the grounds so I measure with my bead on top. I can only control the water temp in the machine is injecting into the grounds. How much the pucks and portafilter cool the water once it starts passing through is pretty much out of my control. I can make sure the portafilter is locked in and hot when not in use, and do the little preshot routine before making a shot, but as soon as you take it out, wipe it out, add the grinds and do all the little magic acts many do, the longer it's out the more it cools. Same with many how fill a cold basket and then insert it in the portafilter. All of these are temp variables that affect the final extraction temp, so it is what it is, knowing it, other trying to figure out how to make changes to your prep routine that might increase ofrdecrease the temp, I see no need in going through a whole lot of trouble to measure it.

Billc wrote:The Temperature sensor used in the Scace is pretty fast, about 0.25 sec.

This sounds like the time constant. How long to reach .5 degree accuracy after a 100F step change? Maybe even in steam (a case I ran into).

> it is not that easy to do....... with precision and accuracy

I agree.

> The Scace is designed to measure the performance of a machine and not how the machine interacts or performs with coffee.

Ie, if you are using it for other things (like optimizing taste), at least consider other devices or the inaccuracies that can occur when coffee is involved.

BenKeith> I see no need in going through a whole lot of trouble to measure it.

It's hard to optimize taste when changes are occurring that you don't know about.

One method that we used for years in the 90's was to use this similar approach using a blind basket. Frill a hole in the basket, thread a bead probe through the hole, bend it over and begin using. There was enough leakage through the hole to simulate the correct flow rate.

Jon,
The thermal time constant is usually defined as the time it takes to reach 62.3% of the instantaneous temperature change. It can also be measured in standard deviations away from the actual temp but a bit more confusing. Generally all probes are rated in a similar manner so it is a good estimation about he speed of the sensor. Just as an example, a fluke bead probe is about 1-2 seconds, a standard meat thermometer (at least the ones I have measured) are about 10-20 seconds. In my opinion you must have at least 2-3 seconds to accurately measure temp on an espresso machine.



BillC

Say 11 or 12 time constants to get to .5 degree accuracy after a 100F step. So if a ~.25 second TC is correct, the Scace can't produce a 2 second response time. Not sure about water vapor/steam, but air is ~6 times slower than water. So a steam bump might be completely missed.

I think BuckleyT has the superior design for taste optimization.

I was inspired and tried this out with 12 gauge nylon:


Two observations after measuring the temp with / without coffee:
1. They are not the same. (Sorry, its obvious...)

• - Without coffee was ~93-94 deg C. Issue is that the flow is obviously unimpeded (no pressure) so boiler temp changes as new water is introduced e.g. 5-10 sec later
  - With coffee the temperature was ~88 deg C

2. Due to above, I find this of limited use. Having said that, it was good to see how stable the brew temp was (with coffee, so with actual flow). I suppose I could try to calculate a correction factor.

I'm using fast response k-type probe.

I like the idea of using a blind to generate pressure. Might try to fashion one from some plastic or rubber...