Thanks Bob for the heads up, and my apologies for not entering the fray sooner. A certain (*ahem*) buyer's guide impeded my work for an extra week or so. But there are some initial measurements and observations I would like to share from this past weekend.

Followers of the Elektra A3 thread may recall that I took it to Pheasant Creek Coffee for a friendly shootout against Geoff's La Marzocco FB-70. Of course I wouldn't pass up the chance to measure his pride and joy using the thermofilter. Below is the resulting graph:


Setpoint was 203.5F and this was the third or fourth "warm up" shot

Don't be fooled by what appears to be a hump towards the end. The incremental increase in temperature is quite small, so the shift from 202.1F at the 34 second mark to 203F at the end appears larger than what the graph suggests. Without the benefit of the WBC protocol, I'll assume that the number they are looking for is the uppermost, in this case 203F, or about 0.5F less than Geoff's set temperature. How did I create this chart? Ah yes...

Please pause for a reminder from our friends at espressoparts.com...

The device plugs into a digital thermometer, but this is not included in the price. The device will work with any digital thermometer capable of reading a type T thermocouple, however the best available Digital thermometers are built by Fluke and you can reference and purchase the different models available at graingers.com.

Indeed, the thermofilter is only part of the equipment you'll need to take measurements, you'll need a thermometer. Shopping for one was a frustrating experiences as I compared specifications and functions. In the end, I decided on the Fluke 54-II:


Datalogging digital thermometer

In addition to my own research, others recommended Fluke's equipment. One feature that adds quite a bit to the price is datalogging. This particular model can log up to 500 datapoints at one per second. It's controls are intuitive and I'm satisfied with the consistency of the results. My only complaint about the unit is that you must purchase special software and an infrared reader to download data for an additional $150. I opt to key them in myself since the data I'm interested in is only 25-30 entries.

HB wrote: A certain (*ahem*) buyer's guide impeded my work for an extra week or so.

LOL, scuse me, you talkin to me? Because I thought you were. I'll have you know sir that I have a few "ahems" of my own on the topic.

gscace wrote:Andy's comment and Terry's as well hit the nail on the head. It's about coffee taste. The thermofilter's role is in quantifying the temperature profile that produces the best possible taste, whatever that profile happens to be. Once you've got it nailed, you can make the measurements that help insure that you can nail it again, from technique training to machine setup to quality control programs etc. It's important to keep an open mind about profile, but I'll say that whatever the profile happens to be that produces mind blowing espresso, the ability to hit that profile time after time is paramount.

-Greg

Although my previous post of Geoff's La Marzocco profile was a quick look at the boiler temperature calibration aspect of the thermofilter, after reading Greg's comments, my interest took another tack. One of the most important goals of the buyer's guides and how-tos on this site is to provide steps on getting the most out of a given machine under the home barista's real world constraints. The biggest constraint is time. My weekday mornings are probably like other espresso lovers -- we want the best possible shot in short order. There's little if any time for retries.

Temperature profiling helps tremendously in this respect, not only in showing how the brew temperature development affects the espresso's characteristics, but also in getting the magical "first shot, best shot" experience. I think this point is hard for professional baristas to appreciate; they toss out two or three "garbage shots" just to warm up the group. Ignoring the question of cost, I strive to develop a routine that optimizes the first few shots of the day as a practical necessity of getting to work on time.

To create these profiles in the past, I used the same Fluke 54-II thermometer, but instead of the thermofilter, I ran a thin 36-gauge type T thermocouple wire over the basket's edge to measure temperature, laying the wire straight across the surface of the puck and locking the portafilter in. You can get an idea of what the setup looks like in this video of a cooling flush. While over the basket approach does work reasonably well, I have to agree with Greg's comment that it's a "pain in the yass". And as others have noted, probe placement can make a significant difference. You can tell immediately if the probe is off because the profile shape will be all wrong... and you must start again with a new puck of coffee. It is very tedious, time consuming work.

I wondered if the thermofilter might shave some time off my HX fine tuning, especially for those precious first shots that demand extra attention. So in the last week leading up to our friendly shootout at Geoff's cafe, I used the thermofilter to map out the initial flush requirements for the A3. Again, my goal was to get as close to fifth shot quality right out of the gate. These heavy group machines need an extra kickstart to get the system to the same temperature, which pro baristas deal with by just running extra shots... so how close could I get to that with flushing alone?

Each time I started with a "very idle" machine, i.e., at least 30 minutes. Elektra is a fine looking lady, but as you can see in the photo below, she doesn't like to be woken up:


A real fire breathing dragon at 6am

My first approach prior to the arrival of the thermofilter was a variation of the technique I applied for rousing the Cimbali Junior out of bed quicker: Flush... wait... flush... wait... flush... pull shot. It helped the first shot, but the second and third were always better. Let's try it another way, and this time I won't have to waste a lot of coffee. My theory was that I could nail the temperature more accurately for the first shot by overcompensating slightly for the first flush. In other words, instead of counting off Chris Tacy's "one Mississippi, two Mississippi... six Mississippi", add a couple extras to warm the group up (just a tad more) and cool the heat exchanger (just a tad more).

I measured the resulting brew temperature for each of the three first "shots" into the thermofilter, adjusting the first flush by small increments. I'm still experimenting with using the thermofilter as a training tool. So far I've learned:

• Flushing for a few extra seconds the first time to compensate definitely helps the first shot. For this machine, add three Mississippis if it is "really really" idle
• Pay attention to excessive "rinse" flushes. I have the habit of running water through the group to clear the dispersion screen. If the next shot is pulled soon thereafter, it affects the next shot's flush time. I know, I know... duh . Now if I'm pulling successive shots, I hold my OCD tendencies at bay and let the cooling flush handle the rinsing.

These small changes have increased my temperature consistency, so much so that Geoff remarked that it was a non-issue in our shootout. That's high praise considering the capabilities of his gear.

Dan:

Using the thermofilter to learn about specific machine behavior and how to get the best out of it is one of the real benefits. You've pretty much hit the nail on the head here. As a training tool it's pretty cool because you can get a ton of feedback right away, as you found out.

-Greg

Greg,

It would be nice if I could count on the device to be a real shot simulation. Is there anything you can do to bring it closer to measuring a real shot? I assume the killer is designing the simulated coffee cake. But if you are able to do that, the Thermofilter will earn its rightful place as an important instrument in improving our espresso experience.

Abe Carmeli wrote:It would be nice if I could count on the device to be a real shot simulation. Is there anything you can do to bring it closer to measuring a real shot?

what do you mean by "measuring a real shot"? the thermofilter provides a single data point at the upper surface of a substitute puck. what do you want to measure? if you want to map the temperature distribution and shifts within a shot, within the puck, then you're looking at a far more complex and costly device, with numerous very very small sensors imbedded throughout a simulated puck. i'm not saying it's not possible, but you'll likely be looking at a $1000+ plus device, and i question the utility of such an instrument beyond research or curiosity. mathematical modeling might be a better tool.

--barry "too many variables"

barry wrote:what do you mean by "measuring a real shot"?

The measurements are actually different, since in one case, one is measuring purely water temperature (Scace Thermofilter), and in the case of the probe snaked over the pf using real coffee, one may be measuring the water temperature attenuated by the coffee. (Greg's words)

In practice, the two methods produce different results. Temp variation within a shot is different, and also average brew temperature. Just as an example, in a few comparisons I did, thermofilter reported temp variation from 10th second on which was twice as large as one measured with coffee. Highest temp reached in the shot on dual boiler was also higher by ~ 1 f.

Perhaps Greg can come up with a formula that allows a user to translate one to the other, both in temp variation within the shot and in average brew temp.

Abe Carmeli wrote:Just as an example, in a few comparisons I did, thermofilter reported temp variation from 10th second on which was twice as large as one measured with coffee. Highest temp reached in the shot on dual boiler was also higher by ~ 1 f.

I will check myself, but if it's consistent, does it really matter? If the primary purpose of the thermofilter is calibration, then real coffee would only serve to add an unnecessary variable, would it not? Moreover, I don't trust the over the basket technique with coffee to the same degree of precision since placing the probe even offset by a 1/4" can affect the results (however the profile shape will generally be correct).

This might be heresy, but I rarely bother verifying the actual measured temperature my thermometer reports in boiling (or ice) water. That's because I'm only concerned with consistency. If the temperature is X fifty times in a row and one time it is X+Y, I want to know the reason for Y. In terms of absolute temperatures, I trust taste more than instruments.

HB wrote:I will check myself, but if it's consistent, does it really matter? If the primary purpose of the thermofilter is calibration, then real coffee would only serve to add an unnecessary variable, would it not? Moreover, I don't trust the over the basket technique with coffee to the same degree of precision since placing the probe even offset by a 1/4" can affect the results (however the profile shape will generally be correct).

If you are using your taste buds to really calibrate, why do you need the thermofilter in the first place? It is a precision instrument, that with a little more thought, perhaps will allow us to do away with the T/C over the lip altogether. I'm hoping to get it there.

As to the reliability of the over the lip T/C, I can't help but chuckle. Man, it sounds like I'm talking out of your mouth HB. When exactly did I start possessing your body? I didn't get the memo, but I think we may have switched roles here . You are right, the T/C over the lip can be highly unreliable because the placement of the T/C in the basket can affect the reading. There are other problems with it. But when you do 20 measurements and they are all consistently different from the Thermofilter, there is enough there to tell you that the difference is real.

Abe Carmeli wrote:If you are using your taste buds to really calibrate, why do you need the thermofilter in the first place?

Because the back of my head would pop off if I spent the morning trying to refine things like HX flushing techniques by taste alone.

I only consider the over the lip technique accurate for characterizing the temperature profile shape. In terms of absolute temperature, I have never trusted it. Based on Greg's and your comments, the thermofilter is approaching from the other angle, i.e., optimized to determine calibration-worthy temperature. It would be a welcome bonus if a coffee puck simulation accurately modelled a thermodynamic reality, but I would still taste the coffee before I'd believe it.

But when you do 20 measurements and they are all consistently different from the Thermofilter, there is enough there to tell you that the difference is real.

The key word is "consistently" is it not?