Do ultra precise brew temperatures really matter?

[AndyS]

...split from Synesso Cyncra Single Group - A Home Barista's Perspective by moderator...

Synesso has made a couple of changes. The PID & temp sensor have been updated to give 0.1F degree resolution. I didn't bother making this upgrade to my machine- I think that degree of control is probably not as big a deal as some have claimed.

Presumably some blends (like David Schomer's for example) are more temperature sensitive than others. But no one (to my knowledge) has published a legitimate blind tasting experiment showing that 1F differences are detectable, much less 0.1F differences.

[Ken Fox]

Presumably some blends (like David Schomer's for example) are more temperature sensitive than others. But no one (to my knowledge) has published a legitimate blind tasting experiment showing that 1F differences are detectable, much less 0.1F differences.

For me, that disgusting monsooned malabar would overwhelm most any temperature change between, say, 0 and 100 celsius.

ken

[JimWright]

Presumably some blends (like David Schomer's for example) are more temperature sensitive than others. But no one (to my knowledge) has published a legitimate blind tasting experiment showing that 1F differences are detectable, much less 0.1F differences.

In all seriousness, this sounds like a worthwhile experiment. A lot of energy is spent here and on other sites discussing the temperature stability and reproducibility of various espresso machines, and various baristas say they can taste the difference when a certain blend is at a degree or two higher or lower - but can they, really, and consistently, in a blind test?

After all, if readers are considering spending up to $7500 on a GS/3 or Cyncra, or even $2k on a Vivaldi, in material part in order to get precise and stable temps, maybe we need to test the premise that even if achieved, this would really positively affect the resulting shots on a consistent basis!

[erics]

I agree with Jim - perhaps something like a mini-EspressoFest in 2008?

[Ken Fox]

In all seriousness, this sounds like a worthwhile experiment. A lot of energy is spent here and on other sites discussing the temperature stability and reproducibility of various espresso machines, and various baristas say they can taste the difference when a certain blend is at a degree or two higher or lower - but can they, really, and consistently, in a blind test?

After all, if readers are considering spending up to $7500 on a GS/3 or Cyncra, or even $2k on a Vivaldi, in material part in order to get precise and stable temps, maybe we need to test the premise that even if achieved, this would really positively affect the resulting shots on a consistent basis!

I have a better solution, one I have adopted already. I have decided to never knowingly drink any more espresso blends containing MM, including those of Mr. Schomer. This should eliminate most of the temperature specific issues, plus leave my palate in better overall condition.

As to actually performing this sort of an experiment, I would rate your odds at successfully pulling it off as about 1 in thirty thousand (which may be overly generous), by the time you consider all of the inherent difficulties in executing such a study in a way that would be believable and statistically valid.

ken

[JimWright]

For my own part, I would be interested in the results of even an imperfect study (e.g., too few tasters) that were able to show several different tasters consistently preferring certain blends at particular temps with intervals of 1 degree higher or lower.

[luca]

Presumably some blends (like David Schomer's for example) are more temperature sensitive than others. But no one (to my knowledge) has published a legitimate blind tasting experiment showing that 1F differences are detectable, much less 0.1F differences.

Great point, Andy. For all of the mucking around that I'm sure that everyone has done, I can't remember any blind tasting experiments off the top of my head. Doubtless this is because we're all usually too busy adjusting the temperature to make the shot taste better!

I might be able to get some people, machines and coffee together at some stage. If I, or anyone else, could, what do you think that it would actually take to make a blind tasting experiment legitimate? Obviously there's only so much tasting that is practically going to happen ...

Cheers,

Luca

[cannonfodder]

The difficulty would not be lining up equipment, tasters, coffee, and location but getting enough exact duplicate shot to form a valid test sample. A tester is only going to be able to sample 4 or 5 shots in an hour. Even at that leisure pace after a couple of sample sets palate fatigue will set in and the taster's ability to detect the subtle changes will be shot. You cannot rotate tasters since taster x will have a different palate and perception than taster y. You would have to do all of the sampling in one day and within a relatively tight time frame. If you try to break it into two or more days then your reference coffee will have shifted in flavor from the ageing/humidity/temperature change and you would have to start all over.

Then there is the strict quality control you would have to have when producing shots and the inherent variance from shot to shot would make the task even more difficult. I think it would be a good study if you could put enough resources into it to have a valid sample to draw a conclusion. What that number is I do not know, I hated statistics but there are others that could provide a minimum sample set recommendation. I would guess that the number of shots required for each sample set to form a valid study would make the study itself nearly impossible. I think it would be a good study and any data is better than no data but very difficult to pull off. As to my personal opinion, I think I can taste a temperature shift up or down by a degree or two but I might simply be fooling myself into perceiving that as well.

[Ken Fox]

Great point, Andy. For all of the mucking around that I'm sure that everyone has done, I can't remember any blind tasting experiments off the top of my head. Doubtless this is because we're all usually too busy adjusting the temperature to make the shot taste better!

I might be able to get some people, machines and coffee together at some stage. If I, or anyone else, could, what do you think that it would actually take to make a blind tasting experiment legitimate? Obviously there's only so much tasting that is practically going to happen ...

Cheers,

Luca

As one of the only two people I know of (myself and Jim Schulman) who has actually pulled off simultaneous paired blind tasting espresso experiments such as this, I can tell you that the effort involved would be considerable. I'll try to summarize:

(1) First you would have to decide what magnitude of temperature variation it is that you want to test. At the current state of equipment, I think it would be overly optimistic to try to get the shot temperatures being tested closer than 1 degree F. apart. Even that would be a challenge to pull off, and you would want some real-time validation that you were actually accomplishing this on a shot by shot basis. The result would be to add one more variable that could cause you to pitch the two shots produced, in addition to the likely occurrence of shot pairs with one substandard shot in them, requiring the tossing of both shots and repeating.

(2) There are very many different sorts of shot profiles produced by different types of equipment, everything from a typical humped heat exchanger profile varying several degrees F over the time duration of the shot, vs. something ultra flat, and everything in between. Although some in the marquee cafe crowd might like to assume that drinkable espresso only comes out of an LM or Synesso, others (including myself) would beg to differ. Can you extrapolate findings from a flat profiled Synesso to a Cimbali or E61 HX machine? I don't think so.

(3) What kind of coffee are you going to use and at what shot temperatures? If you picked something regarded as finicky, like one of Schomer's MM containing blends, would the findings with this coffee at 203F (or thereabouts) mean anything whatsoever to the sorts of coffees most people seem to prefer? My reading of various posts on this website and others leads me to believe that most people seem to prefer most coffees between maybe 198F and 200F. I think this would obligate you to test at least 2 or 3 different coffees or blends, and with more than one set of temperature differences for each coffee. I'm not as concerned as Dave with the issue of the coffee aging and having multiple tasters; I think you could balance for that in the study design, but you would want to conduct the tastings over a period of 2-4 days, maximum (for a given coffee), incorporating the time when the coffees could be anticipated to be at peak, after roasting.

(4) The mechanics of doing such a study would require two identical espresso machines and grinders, with experienced operator(s) and hopefully, tasters. Having done this sort of thing a number of times now, I can tell you it would be overly optimistic to expect to get more than 8 paired shot comparisons per taster per day. You would need to get at least 5 or 6 comparisons of each condition (coffee and temperature) to have much chance of statistically meaningful results. You'd have to plan it all out, but I'd guess you'd need a minimum of 6 tasters for 3-4 days. This is assuming that you would be testing each of 3 coffees at each of 2 or 3 different temperature variations. Each day's tasting session would last 2 or more likely 3 hours, after you had gotten the machines and grinders all set up and ready to go.

(5) what I am describing would really just be a comparison of two identical machines, and given the differences in shot temperature profiles, the conclusions you could reach would be very narrow and hard to extrapolate to other machines with much different shot temperature curve shapes.

This sort of study is a huge PITA to actually execute. To say it is not fun is to underestimate the aggravation involved for both those who pull the shots and those who taste them. I would not volunteer to be in a study like this, that much is for sure.

ken

[DavidMLewis]

I thought there were a number of commercial machines that allow the groups to be set to different temperatures. The NS Aurelia is one, but I thought I'd seen that on others. That would simplify matters, if someone with such a machine in their shop could be persuaded to participate. Then, you could do triangle tasting, since you're just trying to determine whether someone can reliably taste the difference. I'd pick a single-origin rather than a blend, as something more likely to show a difference (without getting into it with my friend Ken about MM-containing blends, which I happen to like and he clearly despises).

On my Techno, which claims the ability to adjust brew boiler temperature in 1 degree C increments, I can easily taste a one-setting difference on most single-origins. Whether the temperature is actually changing in greater steps, or whether I have a palate that's more sensitive, less sensitive, or of average sensitivity, I have no idea. How that translates to finer increments I also have no idea.

Best,
David

[JimWright]

Don't the Synesso machines allow this?

[iZappa]

I thought there were a number of commercial machines that allow the groups to be set to different temperatures. The NS Aurelia is one,

The NS Aurelia does not have that ability. You can insert three gicleurs in it to influence the temperature (two in the group and one on the flowmeter exit). But since it is a huge HX you will not be able to get the same temperature shot after shot. It is very stable though through a shot. Replication of the exact temperature is hard on the Aurelia.

[pdx]

Don't the Synesso machines allow this?

Yes, each group on a Synesso has its own boiler & pid setting. So a 3-group could give 3 simultaneous temperatures.

[bernie]

FWIW, I have an idle LM 4group that is scheduled to have dual PIDs installed this summer. Not sure when I will get to that as we are putting the finishing touches on a new bakery and bagel kitchen. It will be kicking my butt for a while I'm certain. But, when the machine is done I will offer it up for such a test. It is one big muthah. I'd need someone to arrange moving it to wherever the testing takes place. Maybe a road trip from someone.
Bernie

[John P]

Short answer.

Yes.

Having used a LM Linea for 15 months and then using a Synesso for 2+ years now, I can say not only does the stability matter, but the control over it matters as well. That's not to say that a specific declining temp profile may not do wonders as well, but for targeting select sweet spots in various espresso it makes a remarkable difference.

This statement by Malachi during his GS3 testing days has always stuck out to me.

You're right about the massive difference a single degree temp change brings. I'm actually noticing huge differences at even 0.3F changes. And that is an eye-opening statement.

http://www.home-barista.com/forum...tive-t759-120.html

[HB]

This statement by Malachi during his GS3 testing days has always stuck out to me...

A couple points make me really wonder about the statement that anyone can "notice huge differences at even 0.3F changes."

First, that's smaller than the error limits of most lab quality temperature measurement equipment (e.g., my type T thermocouple, Fluke 54-II), let alone the onboard sensors on the GS3. Just because it says 200.3 on an LCD display doesn't mean the actual boiler temperature is precisely the same. Secondly, the temperature spread between the top of the puck and the bottom of the puck is several degrees. The two temperatures don't meet until the final seconds of the extraction. So given the extraction temperatures (plural) progress over a range and that range doesn't narrow until the last third of the extraction, is it logically plausible that a fraction of a degree over this multi-degree temperature spectrum is critical? Not everything can be reduced to numbers, but intuitively I would say no.

Furthermore, are there many individuals who (a) have sufficiently developed barista technique such that minute differences in brew temperatures are the leading factor behind flavor profile shifts, and (b) are capable of tasting the subtle change in taste caused by a fraction of a degree change in temperature? Again, intuitively I would say no.

[AndyS]

This statement by Malachi during his GS3 testing days has always stuck out to me.

Of course, that statement has always stuck out to many of us. Particularly since it is doubtful that the GS3 prototypes in circulation at the time were delivering temperatures that were repeatable to 0.3F

And John Bicht made the statement that a 0.05 bar change in brew pressure made a significant difference in taste.

But IIRC, neither John's nor Chris's observation was the result of blind tasting.

[HB]

And John Bicht made the statement that a 0.05 change in brew pressure made a significant difference in taste.

I apply the same layman's analysis to John's statement. The brew pressure at the top of the puck is 9 bar... and a lot less than 9 bar at the bottom of the puck. So a brew pressure spectrum varying from 9 bar to the exit brew pressure at the bottom of the puck is impacted by a 0.05 change in pressure? Yet again, intuitively I would say no.

[AndyS]

Yet again, intuitively I would say no.

<sigh> Another "thought experiment."


(Your "intuition" is not the result of blind testing).

[HB]

(Your "intuition" is not the result of blind testing).

That is my exactly my point: Unless it can be demonstrated in a blind taste test, such statements should be treated with skepticism on the basis of a "thought experiment" alone. In contrast, if someone had claimed "I can taste a brew temperature difference of 3F", I would accept it without demonstration.