Christian 'Chris' Schwabe, the creator of KaffeePod, "Entertainment for coffee nerds", has made a nice podcast about the leva! firmware and ito 2. His podcasts are recorded in German language:

Superhirn für Espressomachinen (May 21)
ITOv2 (Apr 23)

So Dietmar wouldn't be the only one posting lately I'll add my thoughts. I purchased the entire kit without giving any second thought what is in it and what is not. It turned out to be the best DIY kit I have ever purchased. Even though I had to buy the pressure sensor from sager electronics, everything went smoothly.

I did not quite like the lack of a brew switch. First I went with the standard method of taking the L-sense from the brew switch and leaving it on all the time. My OCD kicked in this weekend and I wanted to do something about this. There were two contact cables dangling around from the pressure sensor connector. I thought why not use that for the brew switch. I disconnected all the high voltage cables from there and connected the L-sense directly to L. I blanked off the remaining cable. Now the issue was the lack of a light in the switch. However, the brew switch can be disassembled with a flat head screwdriver inserted into the side of it. There is a liggle glim lamp inside which is the same size as the spare LED which came with the kit. After chipping off a piece of plastic from the backside of the switch and routing the LED cables through the switch's empty holes I got myself an LED lit 5V switch. I connected the red wire to the 3rd pin of the LED cable and the black wire to the black faston terminal at the PID LED and voilla. My Silvia looks original and for now I don't even need the display or the rotary encoder to have my coffee.

All I need now is the custom top and the water level sensor and it's finished. Oh yeah, and I need a source of good beans. Any ideas in NJ?

Dietmar's kit is awesome. I purchased one years ago and upgraded my Spaziale Mini Vivaldi. Granted, I only cared about the pressure profiling and the upgrade delivered. It's been years and I'm very happy with it. I posted on this blog a few years my upgrade journey. It's a great little upgrade for vibratory pumps.

It would be nice if flow control was available somehow. I know it's not straightforward to do it. The boiler needs to be at 0bar when starting the sequence and the flowmeter should be after the OPV, but still. It would be nice to have this as a feature. I know I would certainly install that pressure resistant flow meter which is in the leva! manual.

Also, when setting up the LEDs, I tried to directly utilize all pins of the 4pin plug but then one LED is going to be innverted. It would be nice to have that as a feature in the leva! to invert the behavior of LEDs.

> flow meter

Flow meters are problematic. They output only 1 or 2 ticks per second at espresso flow rates. They don't output the flow rate directly. The time between ticks must be measured to calculate the flow rate. Such a low update rate is not enough for a pump control loop that is expected to adjust pump power fifty times per second (60 in the USA). The flow rate graph in the Status Monitor may not look so bad but that is created by interpolating between those sparse measurements. That means it's partly fiction and, more importantly, it can only be calculated after the fact. The interpolated flow rate is not available in real time, for use by a control loop. In addition, flow meters are inaccurate, especially at low flow rates. With pulsed flow, the error can easily exceed 15%. Last but not least, they measure an internal flow of water, not the flow through the coffee bed.

A Bluetooth scale cannot solve the problem of having no good flow sensor. It will measure the correct flow (espresso, not pumped water) and it will be faster (usually 10 measurements per second) but this is canceled out by low resolution, usually 0.1g. Which means that weight measurements have to be spaced at least 1s apart to measure flow rates in the 1g/s to 2g/s range with (barely) sufficient resolution. There is also the problem of dripping, i.e. there is no steady flow in some phases.

Pressure sensors, on the other hand, measure quickly (hundreds of measurements per second), thus allowing the firmware to react to changes in a fraction of a second. They are accurate (1% FSS error with Honeywell MIP series) and they measure the right thing (pressure affecting coffee). I think users are better served with pressure profiling.

What could be done with a flow meter/scale is a cascaded control loop: Let the flow control loop adjust the setpoint of the pressure control loop. However, this would probably only allow poor flow control (slow gradual changes; large error).

One way around some of these problems is to estimate the flow rather than measure it. This is possible if the pump is always running at full power. In this case we have a known approximate flow=f(pressure) curve for the output side of the pump in the pump's data sheet. The pressure, which is easy to measure, can then be converted to flow using a formula. Profiling machines with a hammering sound - which regulate pressure by variable pauses between pump strokes, not by changing the power of the pump - can do this mathematical calculation of flow. This is not very accurate, it requires regular calibration (as the pump ages its characteristics change) and it only gives the estimated flow through the pump, not the flow through the coffee bed. The leva! firmware does not always run the pump at full power. It can adjust the pump's power in fine steps. This allows more precise pressure profiling with less oscillations. The trade-off is that the flow through the pump cannot be calculated from the pressure.

Thank you, Dietmar, for the very detailed explanation! I was just thinking whether there was a way to save the shot if it starts flowing like crazy. So far the setup has been treating me very well. I am about to install the display today but even without it, it is wonderful. I cannot wait for the bluetooth scale and the water level sensor.

> I was just thinking whether there was a way to save the shot if it starts flowing like crazy

Turning the encoder counterclockwise during a shot will decrease the pressure/flow. The current pressure profile will continue but a certain amount, set by the encoder in 0.1bar increments, will be subtracted. For orientation, the display shows flow Q in [g/sec] if a wireless scale is attached.

Sounds good, I'll try it out next time. I've just installed the display today so now I have the chance to play with it a little more. I don't have the scale yet. It's impossible to get the tiny Eureka one in the US.

@Sandc from the Netherlands there is no availability of the Honeywell MIPAN2XX250PSAAX until august. I have one ordered since march...

The MIPAN2XX500PSAAX is available and if I remember correct it the 500psi is the absolute max Leva! can handle. What are the main disadvantages when going for this sensor range and is it even recommended in the first place? Thanks in advance!

I would expect a lower dynamic range since the pressure range is 0-500psi but the digitizing of the ratiometric output has the same bit depth. This means that the resolution of the PID control would be lower which would result in less smooth profiles. I don't know the resolution. I would probably wait with the right part, but @Sandc could chime in here, as well.