Thanks Gary for the great intro and review!

Since I have more experience with the Faema Baby, which is almost a dead ringer for the Cafelat Robot, a lot of my expectation/comparison will be made based upon that. We will also briefly go through some of the differences and improvements between the two.


Twins from different eras!

First impressions

When I received the parcel, my immediate reaction was "wow, that's heavier than I expected!" Upon unpacking, the Robot sits reasonably compact on the counter, similar to the Baby. The Robot weighs about 2.4 kg / 5.3 lb (without tamper). For this reason, the Faema Baby almost feels like a toy - the Baby is mostly made of alloy and is much lighter @ 1.3 kg /2.9 lb.


'I am taller' 'Noooo, I am taller!'

All of the Robot's parts are extremely well-made. If you are someone who appreciates fine details, you will come to be impressed by the impeccable finish on all the stainless parts - including the lever arms, which have some tricky curves for polishing, as well as the portafilter. The workmanship that Paul Pratt is famous for in restoring vintage machines is certainly carried over to these Robots. That alone makes it well worth the price!

Except for the cast aluminium body, almost every part of the Robot is made of stainless steel. This of course includes the accessories such as baskets, portafilter, shower screen, dosing spoon (!), etc. Really impressive.


Cafelat Robot's accessories on the left (100% stainless steel), Faema Baby's on the right (mostly alloy).

The most critical part - the basket

As mentioned by Gary, the unique characteristic of the Robot is that the basket and the piston's cylinder are a one-piece design (see picture/diagram below). In a classic lever machine, the basket and cylinder are two separate components where the cylinder is the hefty lever group head made of brass and is designed to cool down superheated water from the boiler.

Having this one-piece design provides several advantages for the Faema Baby. For one, there is very little mess when grinding directly into the basket (think of it as a free built-in grind dosing funnel!). Secondly, there is no need to preheat a separate cylinder.

In this one-piece design, getting the basket thickness right is critical as too thick of a basket will result in significant heat loss - causing a low brew temperature or necessitating an extensive preheat routine. Alternatively, too thin of a basket will allow deformation under pressure. The Robot's basket weighs a measly 83 g / 2.9 oz, which is a reasonable feat when you consider that a typical basket weighs around 25-35 g / 0.9-1.2 oz and a typical lever group head (i.e., cylinder) around 3-5 kg / 6-11 lb!


Left : Comparison of the Robot's two-in-one basket and a conventional basket design (this picture is from Cafelat on the Robot's Kickstarter web page). Right: The working principle of the Cafelat Robot's special basket.

In case you are thinking of getting an aftermarket tamper, please note that the Robot will require a special tamper size. Despite being a '58mm' basket, my regular 58mm tamper (i.e. not the oversized 58.x mm tamper for VST, etc.) did not fit. Measured with a caliper, the tamper that came with the Robot measures about 57.30 mm and the Robot basket diameter around 57.85 mm (which tapers down to about 57.5 mm). However, there is no actual need to obtain an aftermarket tamper - the tamper supplied with the Robot works reasonably well and you get the cool factor where the tamper can be hung onto the Robot when not being used.

Other miscellaneous notes

You may notice a slight metal rubbing sound when moving the lever arms up and down. This is perfectly normal as it is inherent as part of the design - there are some moving parts in the arm mechanism that will slide over each other to allow movement. As the arms are made from stainless steel, wearing out is very unlikely.

A PTFE shim is sandwiched between the flat moving part between the two lever arms to provide smooth lever movement as well as to minimise lever wobble. For some of you observant ones, you may notice there is minor movement in the lever arms - again, it is part of the design requirement to allow for alignment of the piston/portafilter when in use and for ease of assembly. How did I find out? Well, I 'cleverly' tried to shim the arms to fill up the small gap (0.1mm or so), but it was almost impossible to reinstall! Note to self: don't try to reinvent the wheel.


Note the white PTFE shim piece that was sandwiched between the two arms.

Parts compatibility with vintage Faema Babies?

One of my questions coming into the review was whether we are finally getting a new spare part source for the vintage Faema Babies. As some of you know, sometimes these old Faema Babies can be found on the secondhand market but often with missing parts such the basket, portafilter, seals, etc. Unfortunately, I reluctantly report that most of the parts on the Robot are not cross-compatible with the Baby. This is most likely due to the design improvement implemented on the Robot (different portafilter bayonet system) and also possibly because the Robot was designed around the silicone seal (i.e. , incompatibility was not by intention). The shower screen is the only part that is interchangeable between the Robot and the Baby.

How easy it is to use?

Compared to many other portable devices, the Robot is very straightforward to use. Back-to-back shots can be managed with a reasonably quick turnaround time due to a relatively minimal routine. In my experience, the most 'challenging' part in getting used to on the Robot is when locking in the portafilter. This is where you are required to lift the lever arms with one hand while holding the portafilter (filled with hot water) in another. However, the learning curve was short and I was very comfortable from shot 3 onward.

Brew pressure and pressure gauge - usage impressions

Since I've got the 'Barista' version, it has the pressure gauge fitted to show real-time brew pressure. The pressure gauge is small (and cute), and it is surprisingly very readable. Although the pressure gauge itself is workable, I feel that its placement or the direction of the gauge could be improved. I often lean my body over the Robot and use some of my body weight to help pulling the shot - meaning I will be looking over the Robot from top-down. In this position, it is difficult to look at the gauge during the shot extraction when it matters. Obviously, I could also pull a shot while staying in front of the Robot and facing the gauge. However, in that position, I could only comfortably exert around 5-6 bar as the stainless steel arms can start to painfully pressing into your hands when applying extreme force. Fortunately, I found that even a 5-6 bar shot can be very delicious. In my opinion, adding silicone sleeves around the arms could make them more comfortable to hold. In an email correspondence with Paul, he is currently looking into a new gauge/bracket design that can be swiveled to any position.

With the minor caveat above, the main advantage of the Robot over the Faema Baby is its versatility, with a full range of pressure available (up to 9-11 bar *) and a real-time brew pressure readout. The characteristic of the Robot's shots can be changed by manipulating the brew pressure. With the Faema Baby, I am often limited to 5-7 bar (assuming the mechanical advantage of the levers on both the Robot and the Baby is the same) as I was wary about breaking it (especially the thin basket), resulting shots that tends to taste softer and have lesser crema in comparison to the shots from the Robot.

*Edit 20 Oct 2018 : Cafelat does not recommend going much beyond 8 bar, although you have the ability to do so. Adhering to the guide should ensure you get many years of faithful service out of the Robot.

Preheating - temperature performance

For a pourover/unpowered espresso device, the main challenge is often the heat loss or the ability to reach a reasonable brew temperature. Without any preheating procedure, the brew temperature of the Robot reads about 90 °C / 194 °F by immersing a thermocouple into the hot water above the tamped coffee. This brew temperature can be perfectly acceptable for darker-roasted coffees to minimize potential bitterness. For lighter roasts, you can increase this brew temperature by preheating the portafilter/basket assembly by briefly soaking them in hot water (such as in a latte cup).


Preheating the portafilter assembly in a latte cup.

Another preheat technique I've learned from using the Faema Baby is the so-called double-pour technique. This is done by filling the portafilter (tamped coffee bed and shower screen in place) with off-boil water and allowing it to sit 5-10 sec. Instead of brewing directly, discard the water and refill the basket with fresh off-boil water. By doing that, a brew temperature of 92-94 °C / 198-201 °F can be achieved easily without any laborious preheat routine.

I was using a medium-light roasted Colombian coffee for some of the test brews - it can taste harsh/acidic when not pulled properly. Using the double-pour technique, the taste seemed about right to my palate. I have also purchased some ultra-light filter roasts for testing (the Kenyan was retailing for AUD$88/kg!) and I was able to get great results with both the light roasts (albeit with minimal crema, not dissimilar to what I get using the commercial lever and Linea Mini).


Ultra-light filter roasts used for testing the Robot.

This double-pour technique may seem unconventional to some, as it involves discarding some 'coffee water' that is light-brown in color. However, this discarded preheating water proved to contain very little coffee material and therefore the loss was negligible - I couldn't get an accurate reading on my Atago refractometer (<0.05% TDS), representing less than 0.1% loss of the total extracted/soluble coffee solids.


The discarded preheating water - not enough to matter...note that the refractometer isn't really up to task for accurately measuring such low TDS %.

Initial impression of shot quality

Obviously, it is not possible to judge a machine (especially a manual lever with pressure profiling capacity) only after a short period. However, the shots have generally been very good, and it is very easy to get great shots once the grind is dialed in. The body of the shots generally seems a hair thinner compared to my commercial lever (Brugnetti Aurora), but the flavour layering was really good (comparable or even better). I can taste the sweetness and the nuances of the medium or light roasts coming through easily. Although the recommended dose out-of-the-box was 10-16g, I found that 18g worked best when I was test driving the Robot. To pull a ristretto (for darker roasts for example), I find that it's better to updose to 20-21g instead of grinding finer. However, this may depend on your pull preference and the type of the coffee/roast.


One of my first few shots - very easy to obtain good shots once the grind is dialed in.


Another shot from this morning

Thoughts

Throughout the testing period, I have been really impressed with the ease of use of the Cafelat Robot (for an unpowered device) and its build quality. It was built on the Faema Baby design (an excellent design as it is), and Cafelat has incorporated many modern features such as stainless steel construction and a brew pressure gauge to make the Robot practical for daily use. The Robot seems to work best if I just pull the shots by feeling (rather than being extremely meticulous by going with a timer and executing a pressure profile to the nth degree). For this reason, I can see those who enjoy the hands-on nature of levers and typically have a more casual approach towards coffee (go-with-the-flow type and as long as the coffee is delicious) will love it. Overall, I find that the Robot is a well-executed modern iteration of the classic lever experience - the relatively forgiving nature of levers and the charm of engaging with a lever are two aspects of the Robot that I really enjoyed.

p/s: For the next installment of the review, I am hoping to summarize some key points/guides that was helpful to me in using the Robot.

Since some of the readers of this review will have already purchased a Robot, this part of the review is intended to serve as documentation to provide some dos & don'ts that may help you in getting the most out of the Robot. If you have not already, going over the manual written by Paul first is highly recommended.

A word about brew temperature and preheating

When brewing with the Robot (or any other unpowered espresso device), it is OK to use boiling water - don't worry, you won't burn the coffee! By the time the water has come in contact with the coffee, some of the heat will have already been absorbed by the basket/shower screen.

With no pre-heating, the Robot achieved about 90 °C/194°F in my experience (depending on your ambient temperature and the measurement method). For some darker roasts, this is a perfectly acceptable temperature range. In the past, I had found that using a lower brew temperature can sometimes be surprisingly good, bringing out a different characteristic of the coffee. So keep an open mind, try it without preheating, and see if you like the result.


This was how I estimated the brew temperature - it is very sensitive and the temperature can jump around a lot!

Attaching the spout

All Robots will come with a detachable spout that you may choose to use or not. Insertion of the spout attachment may be tricky without silicone grease. One of the tips I discovered (which I later found was also mentioned in the manual) is to wet both the spout attachment and the portafilter with some water prior to inserting the spout. Using a back and forth twisting motion while gently applying some force will also help. This way you don't have to apply the silicone grease if you choose not to.

However, definitely do not try to fit them dry (and without silicone grease) as it could potentially damage the o-ring.

Coffee dosage

On the box, the recommended dose is 10-16g. However, unlike a conventional machine, there are really no hard-and-fast rules for dosing the Robot. I personally prefer results using 16-18 g. Below that, a fine grind is required and a good grinder is highly recommended. You may also updose to 22 g, or even more. Just be careful when the piston is on the downward travel, as there is a chance the piston may press on the shower screen and potentially damage the screen (particularly if the screen is shifted and not located in the center).

For fine-tuning the shots, Jim (another_jim) has a guide to adjust dose and grind by taste that is almost like it was written for the Robot, since the Robot can accommodate such a large range of dosage. This is a highly recommended reading/refresher and a perfect exercise for the Robot Espresso 101: How to Adjust Dose and Grind Setting by Taste

Tamping

With the deep basket, it may require a different technique than with a regular basket/tamper. With the Robot, I personally like to twist the tamper back and forth (to feel whether it's level) while applying some downward force until the coffee compresses. Don't try to tamp too hard as the tamper might be tilted in the process.


Lock-in

Fun fact : Did you know that the portafilter can be locked in both directions, either towards the left or the right? I didn't!

However, it is still best to lock in towards the right as the ears won't be fully engaged in the left position.


How much brew pressure?

As Paul has mentioned in more than a few occasions, the Robot doesn't need to operate at 9 bar to get great results. I have been pulling shots between 5-8 bars and finding the same.

If you have a pressure gauge (Barista Robot), you may be tempted to go to 9 bar (or even higher) as I did initially. After some experimentation, I find that getting an accurate 9 bar pressure is much less critical than applying a smooth force. There should be a reasonable (but not extreme) resistance from the coffee bed itself. If not, adjust the grind -> too easy to push? Grind a tad finer. Too hard to push? Grind coarser.

I typically like to lean over the Robot to apply some downward force comfortably. If you feel strained at all using this technique, that's likely too much pressure!

While the Robot is rather forgiving, I find that I like to keep the brew pressure above 5 bar, otherwise the result is not as good as it can be.


'Proper' way to apply pressure

One of my early mistakes was holding the claw when applying pressure. My thought was by holding the lever arm near its end, I'd have a longer leverage. However, as I found using the Robot that way can cause some discomfort to your palm. The key here is to apply force on the flat bar (just before the 'claw'). If you still find that it is painful, most likely you are applying too much force. Alternatively, you may also add some sleeves to the flat bar for a more comfortable experience - though it likely won't be needed with the right technique.


This is NOT the way to hold the lever arm


I have some clear vinyl tubes (that were cut along the length) that I used as a sleeve for the Robot - they blended in rather well there. You probably don't need them if you are applying pressure using the right technique mentioned above.

Draining the excess brew water

I find it extremely helpful to have a dedicated 'drip tray' to catch the excess water. Another tip you may use is to fill the basket with slightly less water so you don't generate that much excess brew water in the first place.

If you are not making a second shot right after, you can leave the drip tray under the portafilter to catch the drip. The gravity will be pulling down on the Robot's arm/piston to get rid of the excess brew water. So you can just leave it and go enjoy the hard-earned shot you have just made. Come back 1-2 minutes later to a dry puck. So one less step without needing you to do anything!

Note: if you are using fine grind (for ristretto for example), you will find that it takes longer to purge the excess water. Using a coarser grind will be faster.

Technique for removing the basket post-brewing

After pulling a shot, you might find that the basket can be stuck onto the piston seal. To avoid this, lift the lever arms, tilt the portafilter forward (or backward) and the basket will come off easily.


Lift the arm and tilt the portafilter - easy!

Cleaning the basket post-brewing

The coffee puck does knock out in one piece. However, most of the time there will be some coffee retained in the basket. That is normal and is almost inevitable. You can either wash it in the sink, or use a paper towel to clean it if you are making back-to-back shots. For this reason, I find it easiest to use the Robot near a sink and faucet.


This is typically how the basket looks like after knocking out the puck.

Pressurized Portafilter, yay or nay?

Initially, I did not expect to like the pressurized basket much (it is usually known as a way to make 'espresso' using less than optimal coffee). However, after some tips from Dr Gary, this basket has performed better than what I had expected, especially for some of the lighter roasted coffees that don't work well for espresso. The mouthfeel of the foam obtained using the pressurized portafilter is indeed different from the non-pressurized basket, but it doesn't have to be the same!

Here's a recipe that I use. This is by no means a fully optimized recipe, but so far it has worked wonderfully! The result tasted like a dessert coffee drink with beer foam mouthfeel (with the fruity light roast I have).

(i) Grind slightly coarser than espresso grind. I used a 13g dose.
(ii) Preheat the basket and portafilter once, and fill with boiling water.
(iii) Lock in the portafilter, and place a cup underneath.
(iv) Some liquid may start dripping into the cup. If not, give it a gentle push to kickstart the process.
(v) Wait for 1-1:30 min, it's ok if the coffee keeps dripping,
(vi) Finish with 'espresso' extraction (10-15 sec) by pressing on the lever arms - generating 7-9bar.

What you get is something between filter and espresso. It is more flavorful and concentrated than a filter drink, with the beer-like cream that provides additional texture/mouthfeel. The lingering aftertaste does resemble espresso but the taste balance is closer to filter. I am normally not a filter coffee drinker, but this is REALLY good.


Lower the lever arms and allow their weight to pull down on the piston. Alternatively, you can also slot in the tamper for additional weight.

The was the result. Perhaps it was not as photogenic as an espresso, but tasted pretty good!

One note : I always pull a dummy shot using hot water after a shot to keep the pressurized basket clean. This is because there is a chamber within the basket that can't be accessed easily. Not sure if this helps, but there's no harm in keeping coffee parts clean!

Disassembly, should you?

While you may be tempted to disassemble the Robot seeing it is only fastened by some allen keys, I don't recommend doing that unless there is a compelling reason to do so (i.e., diagnosis of a leak or blockage, etc). This is because reassembly may be frustrating and time-consuming for first timers (specifically aligning the lever arms and piston). Disconnecting the quick-connect tubing might make it leak-prone down the road (the connectors can bite into the tubing, leaving marks) and you may be required to replace the tube. If you are really interested in learning how the Robot is put together, Paul has great videos on the Robot's Kickstarter webpage (in the update section #4 for disassembly and #14 for assembly).


My favourite approach to use the Robot?
Obviously there are many ways to use the Robot and there is no one right way. For me, I found the most enjoyment in the Robot by going with the feel. To do that, I fix a dose (18g), get a shot glass (2oz), pull by volume, adjust the grind by the resistance and taste. No timer, no scale. I don't really aim for a particular shot time or shot beverage, although ironically it does help to do that a few times to calibrate yourself.

As the Robot can easily pull over 40-60 g beverage mass, I find that I tend to overdo the volume. While it is not the end of the world (you might even like the result!), I find that I can prevent that by using an appropriate shot glass to prevent myself from overpulling.

When I first heard about the Robot I thought 'interesting gimmicky toy for under extracted cold shots'. A few years later and after a couple weeks of hands-on experience with the machine, my initial opinion has changed.

As with anything Paul produces, the build is top notch. Solid construction, no cheap plastic parts and a solid fit, no wiggly bendy bits. Out of the box the build quality is evident. When I picked up the box, I was honestly surprised at how heavy it was. Thick gauge metal casting, Teflon gaskets and thrust bearings, a non-slip sticky base (very important for a lever espresso machine) and, *gasp* a tamper that actually works included with the machine.

The portafilter has a removable bottom similar to Cafelat All-In-One portafilter design, although with a manual espresso machine that requires two hands and some effort to operate, you will not be sticking your head under the machine to watch the extraction. You have a normal filter basket, an optional pressurized basket and some spare gaskets. The included tamper is nice and frankly all you need for the machine. The little details like the rounded edges on the top are nice. Again, attention to detail, squared edges would be uncomfortable when tamping. The deep portafilter basket requires using your fingertips on the top of the tamper. On my sample machine, the tamper only fits in the left arm, the opening in the right is smaller and the tamper will not fit.

A note on the pressurized basket. If you are using it, chances are you should not be using this machine. If you are not willing to get a good grinder and use fresh beans, you probably do not have the patience and dedication needed to use the Robot.

I was surprised by the size of the Robot. While the footprint is relatively small, it requires a large working area. Traditional lever machines are narrow but deep with the lever protruding out the front. The Robot is shallow but requires a lot of lateral clearance. Point of reference, the arms at 90 degrees and fully extended requires more counter space than my big Elektra A3. While storing requires little room, using requires ample space and overhead clearance.


Pulling shots

I would pre-heat the basket assembly while preparing. Usually 2 fills directly from the boiler of my Elektra would get the basket toasty to the touch. My experience with non-powered espresso machines has always been heat management, or rather the lack thereof. They do require some fiddling with, it is not a grind, pull and go machine.


Due to the spouts and tall basket, a work surface that will hold the portafilter is a good idea. My old Cafelat tamping station held the assembly while I prepared the shot and tamped. You are not going to grind directly into the filter basket due to its height. Also dosing by volume is a problem again due to the size of the basket. My quick fix was simply grinding into a normal filter basket and using a funnel, dump the grounds into the Robot. Tamp, insert the water dispersion disk in the basket, then fill the portafilter with water directly from my boiler, lock in and pull.

Shots were quite good. I was using a Brazil peaberry and Ugandan blend I had roasted a few days ago. Using water from a tea kettle produced a sour cold shot even with a pre-heat. The Robot will probably work best with blends that require a lower brew temperature. The water cools quickly once poured into the basket. To get temperatures high enough for the coffee I was using, I was dumping directly from the boiler of my espresso machine. That water is much hotter than you will get from a typical electric kettle.








The shots were lighter in body, which is normal from a lever machine, but quite good. It does make real espresso. It takes some effort to extract. You will want the machine on a solid surface and slightly lower than shoulder height so you can use your body weight to assist with the extraction. I was using the same grind and dose that I used in my Elektra as a reference sample. The shots were brighter in acidity with more fruity notes in the cup which is typical of most lever-based espresso machines.


After the shot you will need a catch cup to put under the portafilter for any remaining water in the piston and the post extraction drips. The little silicone cup holder sits solidly on the base of the machine and wiggling was not an issue due to the nice non-slip base on the Robot. The filter basket may stick to the piston after you remove the portafilter. Not an issue, you just pull it off separately, remove the water dispersion disk from the basket and knock out the puck. Then off to the sink with all the parts to wash.




So, who would be interested in a Robot espresso machine? Someone with time and patience that is willing to put more work into a shot. While requiring little room to store, you need some counter space to use the machine. A solid work surface and a nearby sink or basin to put parts into while using is a big plus. You will obviously need an external heat source for your water and due to the rapid cooling of the water a coffee that works best in the lower 195-200 temperature range.

It is a novel little machine and will produce a real shot of espresso, if the barista is willing to put a little time into using it. It is not recommended for someone who wants a pull and go espresso machine or wants the occasional milk-based drink. I will also point out that while I would not consider myself old, I am not 40. Time has started to take a toll on my hands with some arthritis starting to rear its ugly head. At times I had difficulty grasping the bits and pieces, so if you have arthritic hands or reduced dexterity, you may want to pass.