Sunday, July 7, 2019

My Projects 2018




During the past year or so I have been building some electronic projects at home during my spare time.

These are small hand-held test instruments. Top right is an event counter module.

The event counter module will count the number of events (off to on transition) during a set length of time.


Bottom left is a voltage source module.

In a vehicle, voltages can represent fuel level, air pressure, ride height, etcetera.

For my test bench, I needed a module that could simulate these voltage inputs to a control system.

The voltage source module has four resistors, four voltage displays, and four circuit protection items, i.e. 250 milli-ampere fuses.

Grab one of these ten-turn knobs and select any voltage from zero through five volts.

Read the four voltages from the endplate window.

Kind of like an early tricorder.

From Wagon Train. Before there was a Star Trek.

The next project started out simple enough.

Generic 4 - 20 milliamp current loop input is a very low voltage which goes through a variable resistor to ground.

Connected to "Instrument Ground" in the vehicle.  Not "Chassis Ground".

The box seemed kind of empty. Needed some more things to fill it up.
It has three resistors and three switches, for use when things go right.
It has three of the 200 mA fuses for when things go wrong.

Why not add some milliameters?

Why not add some nifty 0 - 100 % nameplates?

Each resistor has a switch that may connect one end, or the other end, of the resistor to "Instrument Ground."

Some fuel level sensors have five ohms when empty, 90 ohms when full.

Some fuel level sensors have 90 ohms when empty, five ohms when full.

In any given prototype vehicle, we don't know which type of fuel level sensors we are going to get.  Using this test box, we can just flip a switch to select which end "0%" or "100%" represents "Fuel Tank Full."

RESOLUTION

Off-the shelf behavior of the milliamp meter would indicate zero for any amount of current less than ten milliamperes.

We need to measure four milliamperes and up. Therefore we choose to change the resolution of the milliamp meter.

We will change the resolution ten-to-one, so that  the ammeter will read all the way down to one milliampere.

The original shunt wire probably had a resistance of 0.005 ohms or so.

Replacing that with a 0.05 ohm resistor resulted in some readings that were a little too high. A little bit more that 10 to 1. In other words the readings were inaccurate.

Experimenting with some small value resistors, a parallel combination of 0.49 ohms and 0.05 ohms brings about 0.049 ohms.

The results are pretty good.  We add 0.39 ohm (black orange white gold) and 0.01 ohm (black brown black gold).

I did not hack in far enough to move the decimal point on the display.

0.39 ohms and 0.1 ohms makes 0.49 ohms approximately. 
0.49 ohms in parallel with 0.05 ohms results in 0.49 ohms approximately.



Reading is 0.5 milliamps low. Good enough for bench work.


[Projects_3F.JPG]

[Projects_3H.JPG]

[Projects_3I.JPG]

fin