Making a thermo-filter

I’ve been making a thermo-filter, like a SCACE device, to make sure that my machine works as well as I think it does. Up till now, I’ve just been pushing a thermocouple up into the group-head outlet pipe, but this does nothing to control the flow rate or mimic the conditions inside the filter basket.


I’ve made the plastic puc from acetal, turned down on the lathe, then drilled and tapped a hole in the centre. The acetal turns really nicely, and the thread feels very strong.



I made the bottomless porta-filter earlier on my milling machine. The valve is a needle valve from eBay, which will enable me to control the flow rate and mimic different extraction times.

I have a bare-wire thermocouple to take the temperature measurement. They respond much faster than enclosed-type thermocouples, as the heat does not need to be conducted through the probe shell. Considering the extraction is over in about 25 seconds, if it takes the thermocouple 5 seconds to respond it won’t be a particularly revealing test.

I’ll fit the probe later. I’m away at work now, so it will be a couple of weeks before I can make sure it fits my machine and glue the probe in position.

How my machine works – an overview

My new machine harnesses the properties of steam to provide very accurate and quick heating of the brew-water and group-head.

To produce decent coffee, we need to push water through the beans at a steady temperature. There is some disagreement as to which temperature is best, but lets for the sake of argument settle on 94C. 100C is too hot, and will make a terrible espresso.

Now, every body knows that water boiles at 100C at sealevel, and that a kettle boiled on top of Everest doesn’t boil at 100C but will in fact boil at around 71C. As it happens, if you boil a kettle at around 6,000 feet altitude, it will boil at precisely 94C: perfect for espresso! The summit of Cerro La Campana in Chile would be ideal.

It takes an enormous amount of energy to turn water into steam. Immagine boiling a kettle on top of Cerro La Campana: it will quickly reach 94C and start boiling vigorously. If you keep the kettle on the stove, it will stay at exactly 94C until all of the water has evaporated, which will take ages and ages and a lot of gas. Now, the steam coming out of the spout will be at exactly 94C, but will be carrying away all of the heat the stove’s flame is putting into the water!

The same thing happens in reverse. When that steam hits a cold surface, it will condense, and turn back into water. When it condenses, it releases all of that stored up heat, yet doesn’t change temperature whilst doing so. Afterwards, when it is water again, it might cool down more, but at the instant of turning back into water, it is at a constant 94C.

My new machine works by creating inside itself the same conditions found on top of that mountain: reduced pressure. The machine takes steam from the high pressure, milk-frothing boiler and reduces its pressure and temperature. That low pressure steam then heats up the brew water and warms up the group-head.

Because the steam has loads of energy, yet releases it all at the same temperature, the brew water or group-head never overheat, and the small volumes of water needed to make espresso can be heated up “on the fly”. There is no need for an actively heated group-head, or a electrically heated large seperate brew-boiler. The machine warms up and stabilises really quickly – my working model is ready to go for steam and espresso  in just under fifteen minutes. There are a number of changes to make, and I invisage it going from cold to ready in under ten minutes.

When compared to a traditional heat-exchange unit, as the brew-water is heated by the steam at 94C (or similar), the brew-water cannot overheat so no “cooling flush” is ever required.

Another advantage is that the steam can be transported quite long distances through relatively small hoses, so the group-head can be far away from the boiler and pump yet still heat up effectively. This allows the machine to be built as a number of seperate, over the counter dispensing units connected an out of the way boiler and pump unit.

New Espresso machine – first working model

Well, I’ve actually built an espresso machine. I built it to try out an idea I had, so it really doesn’t look pretty, but it does work rather well, and makes decent coffee to boot!

It is a working model for a possible new product. The working model is to prove that an idea works, unlike a prototype which shows how a product will look and feel.

It’s an under-the-counter model, similar to a Modbar or Mavam.


This is bit that goes above the counter. Its really boxy and massive, with all sorts of bolts and gaskets sticking out. The next version will be much smaller and curvy. For scale, the shiney round thing in front of the wall socket is the group-head.


This is the bit that would be placed under the counter. Its got a load of control valves, a small boiler, a pump, a filter and a control board. It is connected via hoses to the counter-top unit.

The laptop in the background is for programming and monitoring the control-board. I’ll post some more about the controller and software later.

The under-the-counter part is currently just built on a sheet of plywood. I am in the proccess of designing the next version, which will incorporate all of the lessons I learned building and testing this machine and also look like it might belong in a kitchen rather than a scrapyard.

sg101222 sg101220

Close-ups of the boiler, and the pump and control valves. The boiler is salvaged from one of those steam-generator iron base units. Its very compact and good for about five bar, but I’m only running it at one bar. I’ve had to add some ports for level switches and pressure gauges etc.

The gauges are for me for testing. All of the paramaters are under software control, but physical gauges are usefull for debugging and calibrating.

sg101226  sg101227

These are the internals of the over-the-counter unit. I built it with loads of room to spare for modifications etc, so it is much bigger than it needs to be.

The group-head was salvaged from an old Brasilia Portofino, then machined down and fins soldered on. The plate heat exchanger visible in the left hand photo is used to heat up the brewing water before it passes to the group-head. The heat-exchanger will be under the counter in the next version.

The counter-top unit is made of tig-welded laser-cut stainless steel. Laser-cutting is an incredible process – I designed the unit in Fusion 360, emailed off some files to the laser-cutters, and picked up all of the components next day. They needed only a very tiny amount of cleaning up of the edges before welding.

I’ll be put some pages up shortly to explain how it all works – its all rather different to anything I’ve seen before.