I realize we need a nice stable power supply which we have talked about & I will build.
Not 100% correct. You might be able to switch to a single supply IA:
https://www.analog.com/en/specialty...ation-amplifiers/ad8223/products/product.html and drop the negative supply.
Oopps, sorry.
Also reading the sensor ratiometricly using the Reference pin which I should be able to do.
Ratiometric was a big misunderstanding on your part.
Reference voltages (set points) but as you say these can be the cause of errors & not really required.
So instead of BCD Switches or Rotary encoders etc etc. Can't I just use a 12 digit numerical Keypad?
Great, more stuff on the table. The way to think. Yea!
Your bringing back college projects of the 80's. Scan a keyboard and display 4 digits all in Motorola 6800 assembly language.
I was trying to "stay away" from a full blown HMI (Human Machine Interface),
Your original specs were to easily select like 5 setpoints. I took your potentiometer thing, ground it up, spit it out and said - Hey, use a BCD switch (requires 4-bits)
This pretty much satisfies what you need, but "it would be nice" to display the "measured value".
How?
1) An analog meter with an external reference (The meter usually requires another isolated power supply)
2) have the uC do it.
3) Display could be switched to measured value, SP and optionally deviation (somehow)
So, I took it two steps further:
Reserve one position for - manual
But, I came up with two versions of manual and not mutually exclusive either.
1) Take a 3 digit or so BCD switch and use it directly to set pressure in Engineering units.
You had one already, so reading another 3 is virtually cake.
So, for instance position 0, means use the other 3 digit BCD wheel.
The use of an extra switch here or there is part of the HMI. SO if you wanted a toggle switch set horizontally yo point at the left (setpoint) or right (arrray of setpoints)m that's an HMI decision and probably a good one at the expense of a bit.
Then got "giggles", it would be neat to "tweek". So, the rotary encoder is not essential. It's VALUE ADDED. It takes whatever setpoint you have set and each click increments or decrements it a bit. And again for another value added function and probably not necessary, is to make it a 0 to 100% continuously variable.
Can't get tables to work, Anyway layout something like this:
a:xx.x BCD .....,,,,,,,,,,,,,,,...<A switch B>..............b:x BCD...........................DISPLAY
.........................................................................................................................MV
...................................Tweek encoder........................................................DEV (switch)...........On/Off E-stop
........................................................................................................ .................SP
The left switch selects what BCD thingy to use.
The display can be set to measured, Setpoint and deviation is optional.
The "tweek" encoder just modifies any setpoint.
Even without the tweek encoder, I think it satisfies everything.
BCD switches (in this case 4) are easy to read.
A switch is easy.
A 3 position switch, like DPDT (center off) a bit harder.
Flip the design a bit maybe for ergonomics. It's probably better for a left-handed person. It came out in order of importance.
Deviation might be a little tough to implement. It's hard to implement with a separate independent meter. meaning the use of another IA to do the subtraction and an analog multiplexer would be required.
So, the uC spitting out the values on a serial or BCD display would be the easiest.
The encoder would be a pain. The idea would be to pretend the encoder exists from the design point of view. Make room for it on the panel even, but don't implement it unless you think it's necessary.
You still need buttons, like clear, enter, delete, go, stop
Other options: Increment/Decrement (array) with two displays measured and setpoint or a single serial display.
Increment/decrement digit as pushbuttons using a single display.
What bothers me, in general, is that serial MIGHT interfere with PID/interrupts/Phase angle firing and so could an encoder for that matter.
So, other options are possible, Explain how your keyboard might work?
When you design stuff, there are a lot of things you can mimimize and they are not mutually exclusive and not the same for everyone. Cost to produce, time to market, cost to develop, easy to assemble. Space applications of solar cells are more concerned about weight and efficiency, For terrestrial cells, you need to get them up a ladder and cost is more of an issue than efficiency or weight.