Hi all,

I'd like to share with you an experiment we've done. Using an Arduino Uno and a simple scale, to measure instantenous flow while making espresso.
The figure show 35 grams in 35 seconds, but looking inside shows start at a stable 0.8 gr/sec increasing to 1 gr/sec at about 17 second.
Machine: Bezzera Unica

Next step is understading how this level of control affect the taste in the cup!

Probably too much for every coffee - but I was always curious

Have fun,

Ehud and Tamir

Very cool!
Just wondering what is the response time of your balance? I've found that the inexpensive balances I've used need a settling time before getting an accurate readout.

I guess the next step is to w\ire your system in to the espresso machine for automated gravimetric vs. traditional volumetric extractions? Be fun to also throttle the pump as a function of measured extraction weigh with time. I'll bet someone has done this....

The g/s looks like unstable cos it too oscillatory. U using a PID controller ? U trying to control flow right? Try using a PI controller. PID controller is very sensitive to noise level, especially comes to flow control. Also the system responsed very slowly to the change . Most likely caused by noise level if u using PID controller.

So far I've not controlled the flow, just measured it. Contolling will be a later phase II. It looks like the scale responses pretty quickly, but I have not compared it to a standard. I'm planning to use it to see if it can aid in assessing channeling by a sudden increase in flow - I have not seen this kind of data anywhere, so trying first to gather data to see if it can lead to something that will shed some light into my routine.

ok nice . but anyway if u need help in PID controller feel free to contact me . i have studied about process control and instrumentation which PID controller is included in my study which it will be fun if i can apply it here. Btw are you able to get the PID algorithm?

Can you share details on the model of scale used and where you are hooking in to read values from the load cell? This is an interesting project, you've essentially created a home-brew acaia scale here.

The scale is an un-identified chinese scale bought for about $12. While looking for solutions I've found http://www.avdweb.nl/arduino/hardware-i ... rface.html. However, in the scale I had all analog circuit is encapsulated, so I had to connect to the LCD connectors, and decipher the multiplexer - so practically reading the display.
I have a total of 8 connections to the scale - 6 signals (2 for each digit), 1 synch (the first of 4 commons), and one ground.
Now I'm just trying to figure out the right enclosure to put it in, so the E61 drainage at the end of the shot does not soak it all...

ehudco wrote:...I had to connect to the LCD connectors, and decipher the multiplexer - so practically reading the display.

That's actually a pretty interesting approach if it could be somehow generalized. You could use the same thing to grab group temperature from erics thermometers or maybe a tds meter in the reservoir.

I'm very interested to here what you guys come up with, this is interesting work that I think could be applied to a number of useful metrics.