OK. not a super automatic.
Does this happen with both group heads or only one?

Volume in the cup is measured by a "flowmeter" also known as a "volumetric measuring device". Water shoots in through a small hole and spins a little wheel that has two magnets in the top. A sensor in the cap with three wires going to it, counts how many pulses flow has generated and the computer shut off the flow after the determined number has been reached.

The two flowmeters can be identified on your schematic as two devices off by themselves, one for each group, each with three wires going between them and the brain.

On the machine itself the flowmeters can be identified, again one for each group, three wires or sheilding or conduit attached to the top of each one, and a water inlet and outlet connection for each. Outlet going to the boiler/heat exchanger. There should be directional flow arrows on each flowmeter.

I assume in this case the flow meter is ok because it accepts programming.
But what seems to be happening is the machine is "forgetting" not sensing what it's been told to do.

Could be an intermittant ground or signal connection or a defect in the computers memory. Maybe just a bad flowmeter. sometimes the little magnets go bad? Need to know if the problem is one or both groups.

"Looks perfect" doesn't cut it. Physically check every wire, carefully tug on them to make sure they're not loose, tighten every screw holding a wire, check for continuity, everything. Do any wires look burned? Are you sure all the ground wires are there? Burned spots on the board?

Easy if not cheap the just throw another board in and pray, but if it's not the board you have to do all the work anyway. Any other users of this make and model experiance this problem?
Or as you say it's a small problem, is it something you can live with, using a group stop switch?

i agree with normriff.. like he said the flowmeter only sends pule's to the control card.. it would be nice to know if the control card in your model is built with microprocessor or uC in it.. i know some of the older stuff is build using 4000 series logic gates and potentiometers to adjust the number of counts the control card has to wait for until it cuts power to the group solenoid, so th control card is not measuring time, its measuring volume in the form of pulse's from the flowmeter.. 1 count will equal a certain amount of volume that has passed through the flowmeter.. the schematics you have would be good? and or a Hi-res pic of both sides of the control card.

The machine is from 2007 and the first years it's been standing in a "HotDog-stand" (more or less).
Haven't been used for too many shots.
One of the owners had a interrest in coffee, but when he left, the machine just stood there.
Eventually the owner sold it to me.
The machine is like new inside.
It's a microprocessor card. Perfect optical stat. Clear coated solderings with no signs of heat, moisture och defective component.

Haven't been using it's right group more than once or twice and haven't programmed it yet.
I usually use the left one as a "main", and the right when here's a lot of people and the demand is higher.

The previous owner ONLY used the right one and therefor I thought I try the left one out if there was a reason for the selection, or just as me, old habit.

Gonna program the right one today and see if it behave's the same way when left for a couple of hours.

So far, the right group has been working perfectly all day.

Could it be a bad wire or membrane keypad causing an incorrect signal to be sent to the CPU; so it might be the double button instead of the single button.

Maybe try programming all the buttons on the left side?

Good idea.
I will try it!

Often, one of the lights or LEDs on the button pad will begin to blink when there is a fault with flow or sensing. is this happening?

If it was the membrane pad then it would dose the correct amount, intermittantly. Not forget how much it was supposed to dose.

I'm inclinded to think it's the control board at this point. You really have to check every thing else before replacing it however. A good close look at both sides of the control board might reveal a defect.

There is no LEDs on the button pad.

Don't forget that it works for a while if I program it.
If I am to make a string of shots, I can program it on the first and all the following is gonna be the same.
Then it chooses to forget.
The circuit board is perfect even under a magnifier.

normriff wrote: If it was the membrane pad then it would dose the correct amount, intermittantly. Not forget how much it was supposed to dose.

My thinking in suggesting the pad as a possible cause is that the "forgetting" might simply be a misinterpretation --

If you push button A but the CPU senses button B was pressed and dispenses the dose programmed for B, which is not the same as the dose you programmed for A, you might think that the CPU forgot the programming for button A.

If the connection is flakey, it might not be consistent.

I hear ya.

But let's say that I on button "one" sets 2cl.
Button "two" 4cl, button "three" 6cl and button "four" 8cl.
I always use button "four". But all of a sudden it gives me like 12cl. None of the above.

I have programmed some of the other buttons now (into shot glases) and it delivers exactly the programmed amount each time.

I will give it a couple of hours now and see if it forgets.