DIY questions: PID pressure and how much water in portafilter, etc?

#1: Post by **diy_hacker** » November 2nd, 2016, 11:53 am

Hi,

I've been designing my own espresso machine from parts of old ones, I now have my brew boiler working, and I'm tuning.

-My boiler is the spring kind, will be full at 4 bar, anything after that, water is coming out
-Inside of my brew boiler, I have a thermocouple touching the water, so I get accurate measurements
-I have a pressure sensor
-I have a flow meter
-I have the ability to PID both temperature and pressure, which I am doing
-I have an OPV as well, I thought of adding feedback control too, but senseless, I just turn it off for dry puck or if pressure is far too high > 12 bar and abort shot - grind looser!

1. First I pressurize my boiler to approx ~4 bar
2. Wait for temperature to stabilize to set point
3. Start preinfusion
4. Wait for preinfusion timeout
5. Start pour (PID of temp is still on) and targeting pressure of 10 bar via PID control on pump, it's pretty darn close, maintains temp throughout the entire shot and pressure as well.

I am getting an amazing shot compared to what I got from my $100 machine but want to fine tune.

Questions and feedback:

1. How much is ideal millilitres to pre-infuse?
2. How much water is usually taken by portafilter in coffee expansion before the coffee comes out? Should this be the entire preinfusion amount?
3. What is the best timeout for preinfusion
4. What is the best target for pressure?
5. Since I don't have much if any temperature swing during pour, what is the ideal temperature as my set point?

Any other feedback would be appreciated.

#2: Post by **ira** » November 2nd, 2016, 1:03 pm

There are a number of threads about the Decent espresso machine. He's been describing the development of that machine for what seems like a year or more and many of the questions you're asking have been discussed and answered there. That machine is closest to what your trying to do and at least what the users sees and why has been discussed at great length. Given that it's a commercial machine intended to be sold for a profit the implementation details are occasionally left to your imagination, but it doesn't sound like you'll find that to be a problem.

Conceptually pre-infusion should end around when drops of espresso appear on the basket. You might consider trying much lower pressures for both pre-infusion and brew, or at least make them easily adjustable so you can experiment. Learn about how a Slayer works for some insight into the potential benefits of very long pre-infusion and total shot times.

And congrats, it's always nice to see technology pushing the bar, hopefully higher.

Ira

November 2nd, 2016, 2:21 pm

Thanks for the reply. I looked at decent espresso, I'm surprised, this guy is doing exactly what I am, except I'm doing this personally and just for fun.

All the parts cost me a grand total of less than $200, and I'm achieving the same result, except I have a dual boiler setup, so steam and coffee at the same time, and doesn't trip my 20A breaker.

Yeah, pre-infusion to get your basket further pressurized when flow falls off a cliff is the right detection method with limits on pressure and volume, that'll be easy enough. I'm thinking of pulling out my calculus textbooks from university and multi-variable calculus to improve upon my multi PID. I'm inclined to add another thermocouple post boiler now.

An SSR is $5, a thermocouple is $3, a flow meter is $3, an analogue pressure meter is like $5, and this is at retail, a microcontroller to actually control this in real time is $2. Imagine the wholesale volume prices. If someone was truly entrepreneurial, you could produce a machine with all said features for $200-300 USD albeit with higher volumes. This isn't really rocket science nor are all the technology components prohibitive. Infact all the sensors are much less cost than any of the hard metal components.

#4: Post by **ira** » November 2nd, 2016, 4:24 pm

diy_hacker wrote: If someone was truly entrepreneurial, you could produce a machine with all said features for $200-300 USD albeit with higher volumes. This isn't really rocket science nor are all the technology components prohibitive. In fact, all the sensors are much less cost than any of the hard metal components.

And if you produced it for $300, you'd probably have to sell if for $999 to cover service, insurance, future development, salaries, rent, marketing, UL and ETL approval and all that other stuff that comes with having a company.

But yes, the technology to totally automate an espresso is easily available and relatively simple to implement today.

I would question if PID is the correct answer to boiler temperature. You know so much about your system, rate of rise of boiler, temperature of incoming water, flow of incoming water that you should be able to do a lot better than a PID if you incorporate all of the stuff you seem to know.

There's a lot of things about espresso machines that are historic and Decent seems to be the first person to be asking, what is possible if we start from scratch.

Ira