DIY from spare/repair parts: make a recommendation

Need help with equipment usage or want to share your latest discovery?
User avatar
Luftmensch

#1: Post by Luftmensch »

Hi there Home-Barista,

tl;dr

I am considering making an espresso machine using spare/repair parts. It has full license to be a franken-machine: I am happy to pick parts from any manufacturer. What parts would you recommend? Bonus points if they are readily available (particularly in Australia).



Longer version

My 'post-it note specs' are a dual boiler PID system. I would like to have the capacity to mindlessly bang out coffee's for guests during brunch or dinner, but daily use will see two to three coffees per day for two people. I love milky coffees so a large capacity steam boiler is a must.

The current leading contenders are a build based around Profitec Pro 700 boilers or a La Marzocco group. One fun idea might be to use the Profitec Pro 700 boilers and the Profitec Pro 300 group (PP300). The PP300 group could be used solely as an active group - a small mass of water to maintain a consistent temperature at the portafilter. The brew water could be drawn from the PP700 brew boiler.

At minimum I would like to include a rotary pump (quieter)... if the pathway to controlling a gear pump looks achievable, it might be fun to go down that route.

Background

I know basic digital electronics (might struggle with the AC). I know basic control theory. I can competently program (embedded and x86). I am also reasonably handy. I have no doubt I could build a basic machine like an HX from spare parts.

Caveats

I know this could be a money sink... I know it could be worse value than simply buying a machine.... I know I risk getting bored halfway through :wink: ... and most significantly I know there may be no in-the-cup gains (or worse) from DIY. For now this is all besides the point - the primary thing that interests me is the fun of the project. An equal goal is to make more consistent coffee than my ageing ECM Giotto (not too difficult I am sure).



Cheers HB, looking forward to learning from the community!

Jeff
Team HB

#2: Post by Jeff »

Have you ever rebuilt or restored a machine?

If not, I'd start there.

That can be very rewarding, provide great insight into how these machines work, and generally give you a machine that performs comparable to how it did new.

I'd probably stop there as well.

These machines are thermodynamic masterpieces or nightmares, depending on your perspective. That they are filled with superheated steam, boiling water, run at 150 psi, and are plugged into a 240 V, high-current outlet makes a bit more than a hodgepodge of parts a necessity.

User avatar
Luftmensch

#3: Post by Luftmensch »

Thanks Jeff! I appreciate the reply.
Have you ever rebuilt or restored a machine?
I service my own Giotto. :D

The typical E61 stuff... descaling, replacement of gaskets and valve seals etc... Its a simple machine. Close to as simple as they come. I have done a couple of strip-down, rebuilds to deep clean it. So no problem in that department. The hydraulic and electrical circuits are very simple.

I hear you on both the thermodynamic design and AC. The nice thing about using a stock boiler is that you know they are rated to a specified pressure and can be paired with their original over-pressure safety valves and thermal safety switches (Profitec has two!). I am hoping i can mitigate poor thermodynamic design with closed-loop control (PID). I am also hoping to stay on the simple side of AC design - many of the components run at mains voltage. I could keep them on a simple/dumb circuit and switch them using an Arduino/Raspberry Pi through a solid-state-relay.

But you are right.... safety should be taken seriously...

Jeff
Team HB

#4: Post by Jeff »

There are some challenges with just PID control of the boilers. First, at least in my opinion, a pressurestat is about as good as you need for a conventional steam boiler. A possible enhancement might be detecting the steam valve open as predictive of needing heat.

There's a lot of lag in the components of the brew boiler that make control challenging, You've got the mass of the heating element, convection currents in the boiler, the mass of the boiler itself, then, especially with an E61 group (rather than saturated), poor conductivity of heat, mass of the group head, and loss of heat to the environment. Add to that no "negative" control (you can only add heat with the heater) and you've got quite a difficult system to manage on the order of tens of seconds that has time constants on the order of tens of minutes.

I don't think that there is anything magical about PID control (being used literally, not as "some kind of electronic control"), other than that espresso machine manufacturers, who weren't masters of active control at the time, could buy them off the shelf.

User avatar
Luftmensch

#5: Post by Luftmensch »

Jeff wrote:There are some challenges with just PID control of the boilers. First, at least in my opinion, a pressurestat is about as good as you need for a conventional steam boiler. A possible enhancement might be detecting the steam valve open as predictive of needing heat.
I agree... pressure control is not particularly sensitive. But I suppose if you are developing fine grained control for the brew boiler... you may as well use it on the steam boiler! Pressurestats will do the job just fine, but they arent as reliable as a temperature probe and a heating element pulsed through a solid-state relay.

Jeff wrote:There's a lot of lag in the components of the brew boiler that make control challenging, You've got the mass of the heating element, convection currents in the boiler, the mass of the boiler itself, then, especially with an E61 group (rather than saturated), poor conductivity of heat, mass of the group head, and loss of heat to the environment. Add to that no "negative" control (you can only add heat with the heater) and you've got quite a difficult system to manage on the order of tens of seconds that has time constants on the order of tens of minutes.
Excellently put! A real challenge for sure. It got me thinking about the relative merits between a large saturated group (more stable, lots of inertia) versus a small 'active group' like a ring group (less stable, more responsive). If constant temperature is 'the be all and end all'... then lots of thermal mass is the way to go. On the other-hand...if a temperature swing of a couple of degrees or less is seen is acceptable, then there are some nice and compact designs out there.

Jeff wrote:I don't think that there is anything magical about PID control (being used literally, not as "some kind of electronic control"), other than that espresso machine manufacturers, who weren't masters of active control at the time, could buy them off the shelf.
Magical? No! Very 'dumb'... pretty much 'Control 101'. I think it was Home-Barista? Someone implemented a kalman filter to help regulate their brew-boiler. THAT impressed me 8)

If you were so inclined, you might be able to build a model to learn the thermal response of the machine to inputs given ambient conditions and conditions of the boiler.



Right now I am reading posts about flow rate valves and seeing if there are some commercial options that look suitable and easy to automate...

Jeff
Team HB

#6: Post by Jeff »

If constant temperature is 'the be all and end all'...
Even two years ago, it was. Now people are talking about exploring the merits of temperature profiling during a shot. One of the thoughts about "lever flavor" is some of it may be due to the declining temperature profile. As there isn't yet a widely available test bed, there's a lot of conjecture as to how temperature profiling will impact flavor.

The next iteration of the DE1 group head (Ultem replacing brass for certain components) will probably lead to some very interesting explorations and debates. Apparently it will be available for retrofit into existing machines.

User avatar
Luftmensch

#7: Post by Luftmensch » replying to Jeff »

Hehe... Right?? I have read about that... Funny thing... There is this strange romance about replicating characteristics about lever machines. As if old must be better? Yet many people are happy to recognise that the coffees those machines were running on might be considered strong and bitter in a contemporary sense. Seems like a selective view of history!

I am all for playing with variables and exploring a hobby. Whatever floats your boat! For what it is worth... I like pre-infusion and low flow rates at the beginning. But I don't think any of that needs to be pinned to a pseudo mythology.

User avatar
Luftmensch

#8: Post by Luftmensch »

Jeff wrote:The next iteration of the DE1 group head (Ultem replacing brass for certain components) will probably lead to some very interesting explorations and debates. Apparently it will be available for retrofit into existing machines.
The shower screen:

Marcelnl
Supporter ♡

#9: Post by Marcelnl »

I beg to differ in the notion that levers are believed to be good simply because of nostalgic sentiment, you do have a point in saying that modern roasts may not be best suited for a lever.

I'd say it all starts with instrospection; what coffees do you like best, what roast level etc. If you like nordic roasts you probably want a different grinder and you'll want a different machine with controls you otherwise do not likely need.
Reading you like milky drinks I somehow think that 'nordic' is not your flavor profile. I always thought my taste was leaning towards more old fashioned cofees, until I recently roasted a batch a bit darker than normal and I combined that with my rejection of rescue can's of Illy red label (if demand overwhelms my roasting schedule I used to buy some escape beans rather than grinding fresh roasted beans) as the 'roastiness' appals me in a day or two. So I think I'm in the medium/light camp of roasting after all, the lever handles it perfectly with pre infusion and a declining pressure slope.

I guess the debate is similar to the good old transistor versus tubes discussion in audio, transistors were not invented because they are soudning better they are lighter, consume less power and last longer. Extracting espresso works under pressure, how the pressure is created does not matter but the method comes with advantages and disadvantages and a pressure profile, and you have to pick your poison. I'd recommend you put together a list of things you want the machine to be able to do starting with what coffees you like, more tick boxes increase complexity/price.

BTW; your grinder may have all to do with varying results in the cup and does more in the cup than your espresso machine, what are you using, is it aligned, what burrs, how old?
LMWDP #483

rktcyntst

#10: Post by rktcyntst »

Jeff wrote:There are some challenges with just PID control of the boilers. First, at least in my opinion, a pressurestat is about as good as you need for a conventional steam boiler. A possible enhancement might be detecting the steam valve open as predictive of needing heat.

There's a lot of lag in the components of the brew boiler that make control challenging, You've got the mass of the heating element, convection currents in the boiler, the mass of the boiler itself, then, especially with an E61 group (rather than saturated), poor conductivity of heat, mass of the group head, and loss of heat to the environment. Add to that no "negative" control (you can only add heat with the heater) and you've got quite a difficult system to manage on the order of tens of seconds that has time constants on the order of tens of minutes.

I don't think that there is anything magical about PID control (being used literally, not as "some kind of electronic control"), other than that espresso machine manufacturers, who weren't masters of active control at the time, could buy them off the shelf.
Or this project could be completely overcomplicated by addinng an active cooling loop including cold water source and radiator in order to bring down temperatures on command through a separately controlled valve! Is it unnecessary, absolutely! Would it make this Frankenstein machine more interesting, absolutely! Want to make it more interesting? Add a water chiller to your cooling loop.

Plus I think it would be the first instance that I know of where both heating and cooling aspects could be incorporated. Just more stuff to control, and more stuff to break!

And who knows, this might give birth to the first temperature profiling machine during a shot pull! We've already got pressure profiling.... :) (I should go patent this)