Buyer's Guide to the Elektra Semiautomatica

[another_jim]

This thread started as a "initial impression", but based on reader interest and 1st-line agreement to sponsor a Buyer's Guide, Jim has agreed to add the necessary details for a stand-alone article. It will be published in mid-October.

This is Elektra's neat looking home HX machine, run basically out of the box with no changes to anything. And the test reveals an interesting factoid about espresso machine design:

-- The machine ran at around 93C for the 10 and 5 minute test
-- On the first 2 minute interval it rose to 100.
-- Thereafter, it ran at "champion stability," unfortunately at 101C.

By commercial standards this is a wimpy machine, 2 liter boiler filled with about 1.4 liters of water, a light group bolted to the boiler top, an L-shaped HX, and a measly 800 watt heater. What was happening is that the HX design is fine for heating the water on the fly, and once the group overheats, the faster one pulls shots, the hotter it gets.

When we did the home routine, pulling a few shots at a time, with long intervals between the the shot groups, and the usual cooling flush till the bubbling stops, the shot temperatures (tasted, not measured) were fine. But the machine is not set up for commercial use. Interestingly, it doesn't "run out of gas," instead it stabilizes at way too hot. Obviously, I could drop the pstat, let the HX scale bit, or do some HX tuning trick, and get it tuned for commercial use. But out of the box, it gets too hot, not too cool.

There are two morals:

1. the sequence of 10 second idle interval shots in the WBC test doesn't stress the thermal capacity of any commercial machine in the least.

2. Espresso machines are "set up" to transfer the more than available heat to the group and espresso water with different shot making rates in mind. The WBC test is going to reward the most flexible rate machines rather than the ones most stable within a limited range.

Final note: if you want to get real handy, real fast, on a pourover machine with a small drip tray; run the WBC test. You get to practice the required moves at slow speed first, then get dialed in as the pace quickens.

[another_jim]

On Tuesday, a very shiny new toy arrived at my door, the Elektra SXCD:



The back story is that I wanted to have a high tech machine, and after considering the Brewtus and GS3, I decided to convert my Tea into a lab machine. This will require a lot of hacking and downtime, so I wanted a new machine for everyday use. Now that I have 24 inch cabinets, there was one obvious choice, the machine I've always jonesed after, the Elektra Semiautomatica. A breakdown of the Tea's pump and Andy's arrival forced my hand, and I got it post haste.

Thanks to the good people at 1st-Line for making the buying and resolving a few teething problems a no hassle experience.

Mark Prince's review on this machine is a labor of love, a must read, and I won't have much to add in terms of performance. I will be giving my subjective impressions, some measuring data, and perhaps a few tips that Mark missed.

First off. This is not a commercial, semi-commercial, catering, or any other sort of business machine; it is a home machine, just one that happens to have a heat exchanger. Case in point: it weighs less than a Silvia, and takes up less counterspace (except upwards). Second case in point: the WBC protocol temperature test. The first two shots were fine; the next 12 were progressively more overheated -- this machine is set up to be good for an unflushed first shot, but once the group gets hot, the water out of the HX is too hot.

Does this mean the shots are bad? Absolutely not. The usual flushing "3 to 4 seconds past boil" works great, and I'll post more detailed "Home Use Protocol Temperature Test" data as soon as I get around to recording it. Just beware: this machine does not behave like an E61 box, and will probably need to be flushed at every shot when making a sequence of them.

I was concerned about another aspect of its being a home machine: the lack of an OPV. More on this below:

How good are the shots?
-- The doubles seem tastier than the Tea's, but more light bodied. No idea why, the water path is shorter, and the dwell time virtually non-existent (2 seconds). There is zero learning curve; I'd think it's a 10 on Dan's patented "morning after" rating. After more playing around, we found that this machine needs head space, and therefore a lower dose (the Elektra baskets look identical to the Faema style ones). Going ristretto improved matters, yielding soft sweet shots with good mouthfeel. Given the lack of an OPV, this has us scratching our heads. Something is preventing the pump from hitting the puck with 12 to 15 bar.
-- But the singles make this machine worth its price; they are transcendent, god-shot+, WBC 5 to 5.5 gems. I'm talking the Italian ristretto single, 10 to 12 mL (less than a 1/2 ounce) coating the bottom of the cup. In this case, the coating is 90% crema, has a pure butter mouthfeel, and smells and tastes like perfectly brewed COE coffees, mild and sweet. I discovered this today (sorry Andy!), and have had 6 shots so far. They've all been this way, so it's not just a one off fluke.

Folks, we have a mystery: a machine with a basket case HX, no pump regulation, pulling god-shot singles, which taste just as luscious as the machine looks. Stay tuned while I try to figure this out.

[HB]

Jim, you are full of surprises.

Given the lack of an OPV, this has us scratching our heads. Something is preventing the pump from hitting the puck with 12 to 15 bar.

What model pump and what's the flow rate? The performance charts on ULKA's website should cast some light on this. Or better yet, add a tee to your thermofilter and measure directly.

Folks, we have a mystery: a machine with a basket case HX, no pump regulation, pulling god-shot singles, which taste just as luscious as the machine looks. Stay tuned while I try to figure this out.

It will take me days to wrap my head around this unnatural event. I will follow your progress in unravelling this mystery very closely. And thanks for adding yet another machine to my espresso lust-list.

[another_jim]

What model pump and what's the flow rate? The performance charts on ULKA's website should cast some light on this. Or better yet, add a tee to your thermofilter and measure directly.

It's listed as the 52 watt Ulka, which has a somewhat more civilized curve than the 48 watt one. However, ristrettos singles on my old SL70 with the same pump without controls were undrinkable.

The group, basically the same as their pre-E61ish commercial group, doesn't seem to have anything to do with it either:

The group jet (part 19) is listed as 0.9 mm, and the water debit of 90mL is about right for that size. The group has almost no water storage. The little block (part 16) has tight passages from HX to Parker, and from there down to the bell; this is why there's no dwell time -- there's no place for the water to accumulate. I thought there might be some sort of OPV incorporated into the three way; but it doesn't put anything into the driptray when one backflushes.

The teething problem I mentioned was greatly exacerbated by my tweaking (just a little bit) a tiny screw. This gives me the feeling that the Elektra is emphatically not a hack and geek sort of machine. I will take measurments, but I may leave the dissassembly and gandering to someone with a lighter touch.

And thanks for adding yet another machine to my espresso lust-list.

Cheaper thrill than a GS3, A3, or Appia.

[Marshall]

-- But the singles make this machine worth its price; they are transcendent, god-shot+, WBC 5 to 5.5 gems. I'm talking the Italian ristretto single, 10 to 12 mL (less than a 1/2 ounce) coating the bottom of the cup. In this case, the coating is 90% crema, has a pure butter mouthfeel, and smells and tastes like perfectly brewed COE coffees, mild and sweet. I discovered this today (sorry Andy!), and have had 6 shots so far. They've all been this way, so it's not just a one off fluke.

Congratulations! It sounds like the design was optimized for the Italian home market: singles, singles, singles.

It's a beautiful machine, but something has to be done about that Tinkerbell finial. Is there an aftermarket for Mussolini-style eagles?

[miKe mcKoffee]

It's a beautiful machine, but something has to be done about that Tinkerbell finial.

I thought it was a little crystal ball for communing with the espresso spirit gods.

[cannonfodder]

I have said it many times, Elektra makes the most attractive machines on the market (Speedster vs Elektra is a tossup). Your performance observation is interesting, I look forward to your testing, and OOOOWWWW, pretty machine!

[another_jim]

It's a beautiful machine, but something has to be done about that Tinkerbell finial. Is there an aftermarket for Mussolini-style eagles?

They have two eagle styles: the taking flight one shown on the pics of other models, and a hunched shoulder fascist one shown on parts diagrams, presumably special order for those nostalgic for the good old days when the trains ran on time. Maybe I should have got it: Chicago's "Balbo Drive" may be the last street on earth named after one of the Mussolini crowd.

On the other hand, Elektra never quite managed the "sinister cool" fascist style:

[another_jim]

I thought long and hard about how to temperature test in a way that can reveal the everyday operating characteristics of a good home machine. After drawing a blank, I decided not to reinvent the wheel.
-- For instance, I objected to Greg and Barry's protocol because of the 10 second between shot finish. Well guess what, at parties, I've prepared 2 or 3 baskets ahead and gotten close to that pace, refilling spent baskets as the shots were going on. So to say a home machine will never be used at this pace is false. I decided to stick with the prescribed 14 shots at the accelerating pace.
-- However, while I appreciate not allowing baroque flushing regimes in a test; I think the blanket prescription of short 2 second flushes before and after is too curt. Rather than running out of steam, the Elektra overheated at the end of the test and ran at 101C; how would it do with longer flushes?

So I made just a two word modification to the WBC protocol -- the initial flush shall be two seconds past boil. This is equally straightforward, but allows the group to be cooled sensibly. If the water doesn't start at boil, the protocol reverts to the WBC's. Since Dan came up with the "x second past boil" flush, I propose calling this the "HB Test Variant"

Please note the following: I was not using a Scace Device. Instead I was using a single basket with all but one sieve hole glued shut and a k-type TC. The time to get 75mL varied from 23 to 33 seconds. This perhaps has the virtue of approximating some operator variation in dose and grind. Also although my logger records every second, it only updates 25 times per minute -- hence the rows of three dots on the graph.

Here's the graph:



As you see, when it comes to temperature, the Elektra is a humped back whale. There is no particular trend to the shots over time; and the major variation is the stretch or shrink of the same basic curve over the 75mL shot time. Once the curves are stretched to coincide in time, there's about a 1.5C to 2C inter-shot variation. Intra-shot variation is 5C.

What does this mean?

1. Given all the operating variations in flush and time, the curves match up pretty well. One can walk up to the machine, not know its state, do the "x seconds past boil" flush and get a predictable shot. However, it will never be very precise.

2. There is no point at all varying the temperature on this thing. The "hump" occupies the entire 6C range of acceptable espresso brewing.

Now, if this thing pulled awful shots, there'd be little point in considering this data. But the Elektra pulls terrific shots, especially the singles (although the profile may be different for these - can anyone suggest how I can block half a hole on my fake scace basket). This brings us to a paradox:

1. If a straight line is the best profile, adjustability is very precise, one can go anywhere from 90C to 95C.
2. If a hump that stays within the 90 to 95 range is best, then the bigger the hump, the less adjustment is possible. For instance a 3 C hump can only be shifted 3C up and down.
3. If Micheal Teahan is right, and the "best hump" is the 5C to 6C required to catch all the extraction possibilities, then one never adjusts the temperature at all. He would love this Elektra!

Curved shot profiles and doing a lot of temperature setting exclude each other. This is obvious in hindsight; but I have no idea how awareness of this fact will change our thinking about shot temperature control.

[HB]

Now, if this thing pulled awful shots, there'd be little point in considering this data. But the Elektra pulls terrific shots, especially the singles...

It's late, allow me to muse a bit... Wouldn't it be ironic if the Semiautomatica's mid-puck temperature matched the flat profile we've assumed was ideal and the GS3's mid-puck temperature looked indicative of an espresso bound to be sour?

Michael's made the point before that a high initial spike has the positive effect of rapidly raising the overall coffee cake temperature to the desired brew temperature. The precipitous decline you recorded may be necessary to keep the final mid-puck temperature from a major overshoot. A single coffee cake would presumably rise more quickly due to greater water headspace, potentially explaining the single versus double conundrum. That could be the fatigue talking, I'm not sure...

You haven't mentioned the coffees your using. Are there any particular attributes of the ones that perform well on your new toy?

[joatmon]

YOUR ATTENTION PLEASE

We have our first recorded case of "downgrade fever"!

Please protect yourselves accordingly.

[KarlSchneider]

Jim,

Great story. I can see some parallels to my experience with my Elektra -- especially in the single shot preference.

If you really want a eagle taking flight I can let you have mine. I much prefer the blue Murano glass ball on yours and on my Microcasa.

KS

[another_jim]

Jim,

Great story. I can see some parallels to my experience with my Elektra -- especially in the single shot preference.

If you really want a eagle taking flight I can let you have mine. I much prefer the blue Murano glass ball on yours and on my Microcasa.

KS

No thanks, the Murano "sort of" fits, the Eagle not at all. What it really needs is a miniature of a polished modern piece of sculpture, perhaps a mini-Millenium Park "Bean" from a souvenir store.

[another_jim]

We have our first recorded case of "downgrade fever"!

Technically, the Semi is a tad dearer than the Tea or other basic E61 boxes. Appearance-wise, it simply rules: A steel E61 box is pretty striking when it comes to "visual dominance." When I put in the M3 grinder, that was the first thing the eye went to; but the M3 looks plain-jane next to this espresso sculpture.

Also given that it heats up, rather than cools down, when making shots at a fast pace, it's "cups per hour" rating would be much higher than any semi-commercial machine. Remember, the shot temperature graphs I posted were done on the following wbc-test derived schedule (also I don't remember seeing any wbc test graphs, straight or modified, except Andy's Silvia, which I freely admit is more stable than this machine):

But when it comes to features like gauges, water taps, opvs, electronics, etc; it certainly is a downgrade.

Finally, my subjective opinion is that it improves significantly on the E61 box for shot taste. The "significant" part is quite unlikely to remain my opinion once the honeymoon is over. However, Mark spent a year reviewing this machine, and had lots of people over sampling it. They all thought it was a very strong performer compared to other semi-commercial machines. Once the Tea is repaired and instrumented, I'll set up a blind tasting shoot out with local espresso people, who can also bring by their contenders.

[Dogshot]

Once the curves are stretched to coincide in time, there's about a 1.5C to 2C inter-shot variation. Intra-shot variation is 5C.

It is a lovely machine; I'm with Karl - the Murano glass is the perfect topper.

Isn't this roughly the (temperature) performance expectation of a Gaggia?

A 1C temperature change on my BII is a clearly distinguishable difference - not just to me, but to everyone who has tried back-to-back shots from my machine. 5C variance within the shot pretty much says that temperature is irrelevant, as long as those 5C are within the coffee's comfort zone. And yet I can show how a shot can go from sweet to unpleasantly bitter within 3C on my machine with many coffees (as read from the BII's LED).

Or is Dan suggesting that what we are interpreting is a measurement phenomenon for Jim's PF setup?

I'm confused.

Mark

[another_jim]

It is a lovely machine; I'm with Karl - the Murano glass is the perfect topper.

Isn't this roughly the (temperature) performance expectation of a Gaggia?

Read the protocol, get a Scace device, or set up something similar for K type probes, run it with the 2 second or 2 second past boil flushes on your machine or a Gaggia, then talk. Machines have a way of looking rather bad on this test

A 1C temperature change on my BII is a clearly distinguishable difference - not just to me, but to everyone who has tried back-to-back shots from my machine. 5C variance within the shot pretty much says that temperature is irrelevant, as long as those 5C are within the coffee's comfort zone. And yet I can show how a shot can go from sweet to unpleasantly bitter within 3C on my machine with many coffees (as read from the BII's LED).

Or is Dan suggesting that what we are interpreting is a measurement phenomenon for Jim's PF setup?

This is where it gets interesting.

On machines with relatively straight line profiles, changes of 1C, or 1.5C for us dull tongued types, begin to systematically change the taste. As a profile gets more humped, the range of adjustabilty goes down, since pretty soon the hump or the low start and end points will be outside the brewing zone. On the Elektra, the hump is 5C, and any adjustment will take part of the curve out of the comfort zone. Hence it basically allows no temperature adjustment (compensating changes on boiler fill level and pstat setting may change the curve). The profile repeats fairly accurately shot to shot, so that multiple shots will taste roughly consistent; but if a coffee does poorly, one can only play with the shot volume and time.

Micheal Teahan prefers the hump, despite its adjustability problems for two reasons:

1. An initial spike may heat the puck more efficiently. This is what Dan was talking about. However, with respect, it's not what's happening here. The Elektra has a symmetric hump, rising slowly, and falling slowly. Its effect on the middle of the puck may be no different than a straight line profile through its average temperature point.

2. Brewing the puck through the entire temperature range may get more out of the coffee than any single temperature. This is what he was saying at his SCAA and Homecoming talks. In other words, you can adjust your BII or GS3 through the entire range, drink a dozen shots, and not get a single one as good as from some old line, "temperature all over the place" machine. In other words, Schomer's seminal article about "mediocrity by design," which started the trend towards tightly controlled brew temperatures, and all the subsequent work done here, may have been based on a false assumption, or at least one that needs to be heavily qualified.

One problem is when one asks Italian espresso people a question, they reply as if they are humoring little children; and this is especially true if they have no idea what the correct answer is. The Aurelia proves that the ones putting on this act know a lot about espresso machine design; but they consider this stuff as trade secrets, not as scientific or engineering knowledge that needs to be publicly reviewed.

My guess is that Schomer got the "you're a child" treatment too on his Italian Journey, and has been making it up ever since. The difference is that he's publishing, and we're testing and discussing. So eventually we'll figure it out.

[another_jim]

I tend to get analytical with numbers, so one last post on the Elektra temperature test, then I promise I'll stop.

I redid the data so the X-axis shows shot volume, rather than time. This allows the 22 and 32 second shots to be shown in the same scale. I also added an average shot line (thick black) so one can tell how variable the shots are:

As one can see, the peak temperature at roughly 25m/L has a respectable 1C dispersion. But the shots rise and fall at variable rates, so the dispersion increases to a rather shoddy 2C to 2.5C at the shots' beginnings and ends. This remains true even if one removes the three outlier shots (3, 7, & 9)

This machine will never be be particularly temperature stable, although I'll experiment with different boiler levels and pstat settings (there'll be no modding on this baby)

My guess is that one can read what will happen to ristrettos, singles and ristretto singles simply by cutting off the graph at a lower volume, and assuming the temperature be the remaining curve stretched over the time of the shot:
-- A ristretto will be 40-45 mL (15 in the puck, 25-30 in the glass) and show mostly rising temperatures with slight drop at the end.
-- A single at 32.5 mL (25 in the glass, 7.5 in the puck) will mostly climb and level off at the end of the shot.
-- A ristretto single at 20mL (12.5 in the glass, 7.5 in the puck) will have a purely rising temperature curve.

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.

[akin]

Hi Jim,

although the profile may be different for these - can anyone suggest how I can block half a hole on my fake scace basket

one or more single tiny copperwires in an electric cable do the trick for me. Just put it through the hole and bend it.
Hope it helps ...
Akin

[joatmon]

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".

Thanks for sharing your experience,

Jack

[another_jim]

Hi Jim,

although the profile may be different for these - can anyone suggest how I can block half a hole on my fake scace basket

one or more single tiny copperwires in an electric cable do the trick for me. Just put it through the hole and bend it.
Hope it helps ...
Akin

Dankeschoen. I'll try it and be the second person with a ristretto thermo-filter.