malachi wrote:Just finished first testing session. Assume the following is VERY preliminary.
My tentative initial numbers are looking like no-flush ±0.3F.
Note that the PID has not been tuned.
I really need at least two more days of testing and some tuning before I have conclusive numbers, but the goal I would love to see is ±0.15F.

Do you mean ±0.3F stability INTERshot (from one shot to the next) or INTRAshot (within each shot)? We normally expect PID tuning to improve intershot stability, but the physical design of the machine is what really controls intrashot stability. With its relatively large boiler and massive build, it's likely that the GS3 acts "normally."

I assumed the plastic cap was for insulation. You may want to check your measurements with and without the cap.

If you're targeting to measure 0.15F, is the resolution of your meter in the tens place or the hundredths place? When the resolution gets this low, thermocouples can be too noisy. There are semiconductor devices that have higher accuracy and stability. I have one epoxied to my boiler and I can resolve a stable ten-thousandths place with a good multimeter. (For my purposes this is just for fun, not for cup quality per se.)

Keith

Abe Carmeli wrote:
Thanks Chris. A second by second graph would be more telling than a final number. It would be nice if you could post a few shots once you've concluded the PID tuning. And by the way, does the PID have an auto tune, and does it use Fuzzy Logic ?

With this machine I'm having to do a TON of measurements.
It's going to take some time. I'm hoping to have some results by the end of the week.
Results from a few shots are going to be rather meaningless. The margin of error in measurement of the thermocouple and the Scace and the Fluke (combined) are going to present an issue with this machine I think.

I'm not doing the PID tuning - I'm getting numbers from Marzocco.
I'm afraid I can't share a ton of details about the PID right now.

AndyS wrote:Do you mean ±0.3F stability INTERshot (from one shot to the next) or INTRAshot (within each shot)? We normally expect PID tuning to improve intershot stability, but the physical design of the machine is what really controls intrashot stability. With its relatively large boiler and massive build, it's likely that the GS3 acts "normally."

Sorry.

I should have given more detail.
I was speaking Inter-shot.

Here is the deal... this machine seems to be stable to a degree that I'm not used to. And it seems like the margin of error of the measurement rig is contributing significantly to the variance.
In addition, with this machine I found that, after an initial post-idle flush, I was not seeing the usual "bad first shot" temp results.

As a result, my theory (and please, correct me if I'm wrong on any of my assumptions, methodologies or math as I'm no engineer) is that I'm going to need to get a very, very large number of results in order to minimize the influence of the measurement rig.

Thus, the ±0.15F number noted above was the result of a set of 140 results.

The plan is to have more than 500 results to work from.

Does this make sense?
Am I going about this all wrong?
Should I just say "It's stable to 0.5F or less" and leave it at that ('cause I can, in fact, conclude that right now)?

I'd actually love to just leave it at that because I have to tell you all this measurement crap is really boring.

Day 4:

Having burned through the last of the Hines espresso in a marathon session yesterday evening with rockstar Kyle Larson, it was time to roll in the Hairbender. I'm excited to be testing the machine with an espresso I have such familiarity with. This should provide me with a better frame of reference.

First note... if you set the machine to turn on and off automatically you want to make sure you set it to the right times. It was a bit frustrating to wake up this morning and find the machine off. Pure operator error.

Anyway, I dialed the temp to a touch over 198F and pulled the usual garbage shot.
A little tweaking of the grind and...

A very, very familiar shot.
Gorgeous looking with tight and incredibly persistent crema.
Wonderful aromatics (cocoa, vanilla, dried fruit, tropical flowers and some sort of spice I can't get a handle on).
Rich and sweet, with a ton of dark belgian chocolate, some dutch processed cocoa and even a little milk chocolate.
Dried stone-fruit, a little hint of melon, some raw honeycomb... and some tropical spices (cardamom?) and a touch of sweet citrus zest to top it off.
Very complex - and yet very integrated and rounded.
Perfect reproduction in the cup - great clarity.
This is a very dense and heavy espresso for the Hairbender. It's something I'm noting is consistent across all espressos so far with this machine. There is a concentration of flavour in the mouthfeel that is almost like French Press - a sort of feel on the palate that is almost like there are more solids in the espresso than normal. It's almost like the espresso is "gritty" feeling. I know that might sound like a bad thing - but it's not. It makes the espresso taste heavier and more syrupy. Coupled with the incredible clarity in the cup - this allows for an "amplification" of flavours that is incredible.


Anyway - a quick cappuccino... And finally I'm really getting the hang of the steaming set-up. The "lever" control didn't take long, but the limited articulation of the wand has forced me to change how I hold the pitcher. While I complained about this the other day - it actually hasn't taken that long to address. I just have to pay more attention and not go by memory. Really nice microfoam - nearly instantly. The stretching phase is over almost before it starts. If you have to adjust wand angle or depth or pitcher angle you're probably going to over-stretch the milk. That's how fast this beast is.

In milk the Hairbender softens up, the chocolates become more cocoa and milk chocolate dominated and the raw honey becomes a little more polished. The citrus and spice cut through the fats nicely... a lovely drink.


The espresso - better than all but a handful I've pulled on commercial non temp stabilized dual boiler commercial machines. And as good as all but a couple I've pulled on temp stabilized commercial machines. To my palate, it was simply noticeably better than any I've had, pulled or been served on any HX machine (home or commercial).

The cappuccino - not quite as noticeable a difference as compared to the commercial machines. Definitely the sort of thing I'd have been proud of making, serving or being served in any Stumptown location. But I've had capps from Kyle et al off the temp-stabilized Mistral that were better. I think I still need to dial in the milk to get the sweetness there. To my palate, however, it was also better than anything from any HX machine.


I think that espressos like the Hairbender are going to show off the strengths of this machine more than a lot of coffees. Coffees that are hard to work with - that are unforgiving or are finicky or particular are, I think, going to be where we see this sort of really clear difference. The more forgiving coffees are probably going to be less exciting to work with.

I'm very interested in checking out some of the other finicky coffees like the Vivace espressos.
I'm also kind of interested in checking out some "flawed" espressos to see if the clarity issue can, in fact, turn out to be a negative.


I never thought I'd say this - but I don't care that this thing is automatic. I want it.
Damn.

malachi wrote:I never thought I'd say this - but I don't care that this thing is automatic. I want it.
Damn.

Oh but so soon? This must be the shortest foreplay ever. I thought at least you're going to tease us for another week or two before taking us to the last page. And what happened to the "love you till the day I die" pledge you gave the Speedster just a week ago?

I don't have the Speedster in my kitchen.

sigh.

Day 4 (continued):

And a long, long day winds down...

I burned through a whole lot of coffee. Four other folks were by pulling shots, tasting shots and (in general) flipping out over the machine. These were all coffee professionals and baristas. These are jaded, cynical people who drink a lot of coffee and have pulled a lot of shots. And they all started talking about how they wanted one of these at home.



One thing... there has been a lot of chatter about this machine being "ugly" on various sites.
I don't get it. Perhaps it doesn't photograph well. This thing, to me, is an attractive piece of equipment.



Anyway...

I continue to be amazed. I pulled a lot of shots, tasted a lot of coffees. I continue to get incredible clarity. I continue to get amazing mouthfeel. Flavour reproduction is incredible. With every coffee I'm tasting things I'd not tasted before. With every coffee I'm getting shots that are as good as I've ever had.



I also continue to run through millions of boring measurements.
Honestly - I'm really tempted to say "screw it." This machine is at least as temperature stable as any of us want, desire and/or need. But I'm soldiering on. Another couple days and I'll have data. Of course, the data will all be nearly meaningless beyond being able to say, "yes - this machine provides stability of equal to or greater than 0.5F." And I am confident that I can say that right now.



To keep myself sane, I'm breaking up the "boy scientist" silliness with milk practice. And I'm glad to say that I'm continuing to refine my technique with this machine (as you can see from the following 5.5oz cappuccino photo). At this point I'm confident in my belief that milk with this machine is on par with espresso with the machine.



In any event, I'm loving this machine. In fact, I'm head over heals in love with it. The idea that someone is going to, soon, take it from me makes me really, really sad.


If anyone has questions or would like me to take any specific photos for them - or describe anything for them - just ask.

Oh... and if you're in Portland OR this Friday morning and want to come by for coffee... let me know.

malachi wrote:... this machine seems to be stable to a degree that I'm not used to. And it seems like the margin of error of the measurement rig is contributing significantly to the variance.
In addition, with this machine I found that, after an initial post-idle flush, I was not seeing the usual "bad first shot" temp results.

As a result, my theory (and please, correct me if I'm wrong on any of my assumptions, methodologies or math as I'm no engineer) is that I'm going to need to get a very, very large number of results in order to minimize the influence of the measurement rig.

Thus, the ±0.3F number noted above was the result of a set of 140 results.

The plan is to have more than 500 results to work from.

Personally, I think you should stop measuring temperatures IMMEDIATELY. As we say in the hedonistic world of espresso, F#ck the Fluke!

± 0.3F from idling is spectacular. It's a dream machine!

Sit back and enjoy the coffee....

malachi wrote: Rich and sweet, with a ton of dark belgian chocolate, some dutch processed cocoa and even a little milk chocolate.

What, no Chilean cocoa butter? Seriously, though, love your descriptors. Few people can detect flavor characteristics that well, and I've worked with many judges (albeit judging beer).