Buyer's Guide to the Elektra Semiautomatica - Page 2

[another_jim]

Jim,

I agree that it certainly looks like a great machine. Frankly, this is one I had not considered. I guess my attempt at humor failed. But, here I am, lusting after $2.5 - 3k machines and something a tad less expensive appears as a giant dot in the center of my radar screen. If I had just picked up a new Appia or Junior, I'd be having "buyers blues".

It was funny, but I'm in one of those "new owner, don't diss" moods, where all humor escapes me.

The back story is that I'm probably going to spend a lot of hours and around $1K turning the Tea into a temperature and pressure profile control machine. It won't be all that accurate, but it should be capable of limited profiles accurate to ± 0.5 bar and 0.5C. This will allow me to do espresso testing to my heart's content. When the super-techs come out with better profiling machines, they can then properly test any speculations I come up with.

In any case, why not have something completely different for an everyday machine? And it doesn't get any more different than the Elektra for a fairly convenient, good tasting, 24/7 brew and steam machine.

Like Andy, Greg, Ken, and some others, I've done so much inconclusive taste testing on machine adjustments that I've become a complete agnostic about pursuing godshots by tightening and improving their easily measurable performance. Keeping the temperature and pressure within consistent and reasonable bounds shot to shot has worked well; but going beyond that has not payed off so far. I suspect we'll need to use numerical models of the entire water path from mains to cup, along with some taste chemistry models, before the ultra-precise approach starts striking gold. If this turns out to be the case, very few of us will be able to take part in the research (on the taste chemistry part, it looks like we'll all be dead of old age first!).

And except for research, I have little interest in high tech machines.

[ilVecchio]

Somehow the weather here in Chicago contributes to any sense of humor at all. We sit inside with our two collies in a heat related miasma. Only the occasional espresso helps the mood.

[joatmon]

I wish I could share some of these experiences. I'm on an island here, no friends that share this passion. What I think is a good shot may be a sink shot to many here. Heck, to my friends, the "edge" is putting their styro cup under the brew basket during the brewing process as the Bunn takes 170 degree water and spits it all over 2 year old Folgers.

I do enjoy these threads and live out some of my "coffee dreams" vicariously through your research/gatherings.

Jack

[another_jim]

I wish I could share some of these experiences. I'm on an island here, no friends that share this passion. What I think is a good shot may be a sink shot to many here. Heck, to my friends, the "edge" is putting their styro cup under the brew basket during the brewing process as the Bunn takes 170 degree water and spits it all over 2 year old Folgers.

I do enjoy these threads and live out some of my "coffee dreams" vicariously through your research/gatherings.

Jack

Yeah, this is the insanely frustrating aspect of being in an on-line community that's centered on the taste of something. I can talk until I'm blue in the face about how the espresso tastes; but it's all, quite literally, nonsense -- something nobody else can sense, check out, or share.

The biggest thing on my coffee calender is actually meeting the people. The annual SCAA convention (May 2007, Long Beach CA is the next one - make it if you can) is the biggest one, and it attracts about 20 or so on-line people (we need about 10 times that) as well as thousands of coffee pros. This is always a blast. However, I like the meetups in people's homes the best, where one can get a sense of how they do espresso or coffee, and learn stuff that's much more applicable to my home-barista-ing.

In any case, dear readers, if life ever takes you to Chicago, I'd love to invite you over.

[espressoperson]

The ristretto singles from these machines are perhaps the best shots, certainly the softest and butteriest, I've ever had. Yet I've previously gone on record saying that a rising profile cannot be good, since the low temperature will accentuate the sours at the start, the high temperatures the bitters at the end.

This machine is very confusing.

Some non-analytical types might just kick back, enjoy the espresso, and damn the theories!

[another_jim]

Some non-analytical types might just kick back, enjoy the espresso, and damn the theories!

That was exactly my intention. But it's a new toy, and I want to play a bit first.

[joatmon]

Yeah, this is the insanely frustrating aspect of being in an on-line community that's centered on the taste of something. I can talk until I'm blue in the face about how the espresso tastes; but it's all, quite literally, nonsense -- something nobody else can sense, check out, or share.

The biggest thing on my coffee calender is actually meeting the people. The annual SCAA convention (May 2007, Long Beach CA is the next one - make it if you can) is the biggest one, and it attracts about 20 or so on-line people (we need about 10 times that) as well as thousands of coffee pros. This is always a blast. However, I like the meetups in people's homes the best, where one can get a sense of how they do espresso or coffee, and learn stuff that's much more applicable to my home-barista-ing.

In any case, dear readers, if life ever takes you to Chicago, I'd love to invite you over.

Well, I used to get to Chicago a bit, but not lately.

I've been more concerned about temp control after having 8 different SOs in 2 months from Intelligentsia. I can usually dial-in Black Cat and Code Brown and get decent shots. But I know the SOs could really respond to slight changes in temp. I've ordered Eric's thermocouple adaptor and am awaiting all the components so I can start to document, discover and effectively reproduce shots that please me.

I'm close to taking this hobby over the line to passion. Next stop. <backspace><backspace><backspace><backspace><backspace> (No need to offend).

Jack

[joatmon]

Some non-analytical types might just kick back, enjoy the espresso, and damn the theories!

Yep. There's a word for that. newbie.

[another_jim]

Day 2 of trying to master the temperature control on this beauty. It gets geeky from here on in, so if this stuff bores you to tears, skip it. If you're considering buying one of these, this post has some worthwhile reading:

RTFM. The manual says fill the boiler 3/4 of the way up the sightglass; I've had it at half way up since I got it. Today I moved it up.

(Side note on water: No need for insanely long FAQs on this baby. The boiler fill is manual, and there's no water tap. Fill it every few days with distilled water so the boiler never scales, use regular water for brewing, and descale the HX like any single boiler home machine)

I repeated the modified WBC test (flush 2 seconds past boiling before the start) and got this:

Looks like the same mess as yesterday. But wait! If one looks at the last six shots, spaced 10 seconds apart (in blue), they show world class temperature stability. Can this be exploited for regular home use? I tried a double flush. First a longer flush to simulate the previous shot, then 10 seconds later, the 2 seconds past boil flush. This can be done conveniently while grinding and dosing. I did six "walk-up" shots, a la Ken Fox, using this technique, and here's what I got:

Not as good as the final WBC set, but except for the outlier, not bad at all. We'll see how it develops as I get more practiced.

Did at make the least bit difference to the taste of the espresso? Not so I could notice. The hump is still 3C wide, just like yesterday (it looked bigger yesterday because of the wide shot to shot variation, but the average curve hasn't changed much); it could be that this makes reproducing the exact timing of each temperature point from shot to shot somewhat less critical.

[another_jim]

Let's take a look under the hood:

Here's the parts diagram:

The components are mounted on the underside of the chrome plated part of the base, and the plastic under-panel removes with a single screw (Millenium owners may get envious at this point):



Once the machine has cooled off, one can pick it up, turn it on its side, drain the tank, open the steam valve, and drain the boiler. This works since the vacuum breaker is open, allowing in air. It just takes a bit of time. I only drained out enough to set it on its side without getting leaks from the breaker. The base panel has the boring detailed name plate:



Pop it off, and you get the innards:

There's been some changes since the diagram. The hosing is flexible braid, The safety stat has been replaced by a capillary that cuts off a switch in series with the pstat (it's the box next to the terminal strip for the incoming power wire). The pstat is still a mater, but a heftier looking model than the standard one shown.

No OPV, but plenty of room for one if installed closed circuit so the overpressure discharge goes directly to the suction side of the pump. Was there a hint of overextraction in that last ristretto? Hmmm.

This is as simple as an HX heat exchanger machine can get.

A few notes. The pstat is on the water side of the boiler. The boiler fill is the second pipe underneath, while the HX entrance is an external pipe into the side of the boiler. The heater terminals are concealed by the two pipes.

[cannonfodder]

Yeah, this is the insanely frustrating aspect of being in an on-line community that's centered on the taste of something. I can talk until I'm blue in the face about how the espresso tastes; but it's all, quite literally, nonsense -- something nobody else can sense, check out, or share.

Call your ISP, tell them you want the 'T' channel on your data line turned on. Get that special sensory monitor and lick the screen. MMMM, taste-a-vision, the next virtual breakthrough.

[cannonfodder]

I am missing something in the schematic. I don't see the HX line. So it is externally fed through the base, up one of the tubes along side the boiler, then passes through the boiler and into the top of the group via the parker solenoid?

How large is the heat exchanger?

[another_jim]

I am missing something in the schematic. I don't see the HX line. So it is externally fed through the base, up one of the tubes along side the boiler, then passes through the boiler and into the top of the group via the parker solenoid?

It's partially obscured on the pic I posted; it's the external tube on the right of the boiler feeding in near the bottom. You can get a clearer view on the official pics the vendor sites have.

How large is the heat exchanger?

Good question! I haven't disassembled the boiler (it's a lot more than one screw), and there are no drawings to be found. I surmise it's tiny and oval cross-sectioned like the Tea's or other E61 boxes with the horizontal 1.2 liter boiler, and works like an instant hot, rather than the storage heater like HXs found in the Cimbali or Aurelia. My reason for saying this is that the pstat is really closely coupled to the brew cycle -- the pressure drops rapidly when one brews, and heat turns on a few seconds after brew switch. This means there's not much buffering between the cold water coming into the HX and the boiler.

I have a pf pressure gauge coming. If the reading's as bad as I expect, I'll test, and hopefully install, a closed circuit opv. There's plenty of room to do it.

The temperature work is coming along, and I'm getting better at it. With the higher boiler fill and better flushing, the non-outlier shots peak at 95 early, drop to around 92.5 at the end, with all shots clustering in a 1C band This is roughly what I got on my Tea, so I probably won't fool with the pressurestat at all. I'm still trying to get handle on why about 1/4 of the shots are outliers on the temperature profile, and won't post anything more detailed until I have the problem solved.

Once I'm though, I'll write a longer piece, the espresso-nut's version of the user manual, so to speak.

[another_jim]

I've mastered controlling temperature on the Semi; and it bears out what Ken has been doing. Here's the graph and explanation:

The blue cluster of lines in the middle are the last six (10 second idle time) shots from the WBC test I did a few days ago. They represent my target, the tight clustering of inter-shot temperatures I wanted when making my everyday shots.

The red cluster on top (I added 2C to each line to space them above the WBC set) represent the best I've gotten using a double flushing method to simulate the effect of the immediately prior shot in the WBC set. As you can see, this didn't work. A close examination of the lines shows that five of the eight cluster well, and three are outliers. I spent three days trying to make this double flush technique work, but the results were always the same -- mostly good shots, with about 1/3 being outliers.

The black cluster on the bottom (2C subtracted from each line to space them below the WBC set) shows the walkup shots today. They are about a hair away from being as good as the WBC set. Notice there are no outlier shots. For this set I abandoned double flushes and went to a single flush, 3 seconds past boil, immediately prior to making the shot. Then I turned on the steam valve for a second or two and hit the brew switch.

What's this; a new wrinkle in HX flushes? Yes. I noticed in the Elektra, the boiler cycle responds very quickly to the shot being pulled, but not quite instantly. At the end of the WBC test, the boiler is running continuously, so there is no boiler cycle and nothing to create outliers. So the new technique is to flush the HX to put it and the group in a predetermined state, then turn on the steam valve to put the boiler in a predetermined state. Same state for the entire thermal path should equate to the same temperature profile for each shot. It more or less does.

Nods to Ken Fox. In his "walkup" thermometry, he noticed that the boiler cycles on the vibe were messing with consistency, and he PIDed it, he also noticed the autofill on the rotary was messing with it, so he's added an override. Would a steam flush on the Junior have done just as well as a PID. I doubt it; the HX is huge, and the boiler state for the past hour is probably key, rather than just the one during the shot. However, a steam flush may add consistency to the small HX/small boiler E61 boxes.

[another_jim]

On the Elektra's pressure; the portafilter gauge reads 14 to 16 bar at no flow. However, the ristretto singles continue to be soft and sweet, which according to my previous posts is impossible.

No problem! The great French biologist, Buffon, who defended the fixity of species in the 1830s, was puttering in his garden when he noticed a strange bug, which to his expert eye, instantly proved that species do evolve. So he stepped on it. I'll be installing a closed circuit OPV and stepping on the inexplicably tasty high pressure singles from the Elektra.

[miKe mcKoffee]

What's this; a new wrinkle in HX flushes? Yes. I noticed in the Elektra, the boiler cycle responds very quickly to the shot being pulled, but not quite instantly. At the end of the WBC test, the boiler is running continuously, so there is no boiler cycle and nothing to create outliers. So the new technique is to flush the HX to put it and the group in a predetermined state, then turn on the steam valve to put the boiler in a predetermined state. Same state for the entire thermal path should equate to the same temperature profile for each shot. It more or less does.

Interestingly that's similar to what I do with the Bricoletta for flush-n-go. I briefly open the steam wand to kick on the boiler heater and then pull to flash-count to temp-n-go. Just reversed order of what you're doing but result would be similar I'd think. My testing found it more accurate intershot and more stable intrashot than random boiler heater state to begin the sequence.

[another_jim]

Interestingly that's similar to what I do with the Bricoletta.

Sorry Mike, I must have missed that post; the credit is yours.

[miKe mcKoffee]

Sorry Mike, I must have missed that post; the credit is yours.

No apologies necessary and who cares about credit! More or less just stumbled on the technique. It just seemed the right thing to do early on with the Bric' using the Thermofilter to dial in shot temp surf. I never posted specifically about that revised HX surf method just mentioned it a few times when elaborating my flush-n-go shot temp sequence. Since no graphs were included maybe nobody picked up on it.

But I do give you credit for making descaling the Bric' a breeze. Fascinating how the Bric's lever action is smoother and totally quiet after descaling the HX system including descale backflushing. Smoother and quieter even than when new. Be interesting to see if it stiffens back up and gets some squeak back after my next detergent backflush in a day or so.

[another_jim]

Be interesting to see if it stiffens back up and gets some squeak back after my next detergent backflush in a day or so.

The chemical backflush removes the lubricating coffee oils, and you always get the squeak for a few shots. The citric will remove any calcium grit, and may make the action smoother, although I've never noticed this.

[RapidCoffee]

Interestingly that's similar to what I do with the Bricoletta for flush-n-go. I briefly open the steam wand to kick on the boiler heater and then pull to flash-count to temp-n-go. Just reversed order of what you're doing but result would be similar I'd think. My testing found it more accurate intershot and more stable intrashot than random boiler heater state to begin the sequence.

Jim, as I understand it, your heating element is on when you pull the shot (please correct me if I got this wrong). Mike, is that also the case with your Bric? Or does your heating element cycle off before you pull the shot? Cycle time must be pretty short on the Bric...
________
John