Glenn T wrote:Obviously, they had to dumb things down for me. I was told that current should never flow through the ground circuit. The Germans learned to make allowances for Americans, especially ME's!

The only thing more dangerous than a mechanical engineer with a schematic is an electrical engineer with a caliper...

(Extrapolated from my usual "the only thing more dangerous than a software engineer with a soldering iron is a hardware engineer with an algorithm". If you've got a better suggestion for what to fill in, please share)

It is very common to use the chassis of a vehicle as the common ground; a taillight has one wire running to it; the return is the chassis.

Some capacitive sensing circuits transfer no net charge: they use AC.

Just my $0.019999993929993 (still using that pentium 2)

Joe

Postby pocojoe on Apr 17, 2012, 11:14 am
It is very common to use the chassis of a vehicle as the common ground; a taillight has one wire running to it; the return is the chassis.

Some capacitive sensing circuits transfer no net charge: they use AC.

You know, I love how much I'm learning by doing this project! This little min-course on electrical and sensor engineering is bit off topic, but I hope it's instructional to other readers too.

You're certainly right about car wiring! Cars use the chassis to as the common or return and have no separate ground circuit. I suspect the electrical codes and subsequent wiring practices are quite different with higher voltages, though.

I had no idea capacitive sensors are used this way. That's kind of embarrassing when you consider I worked for several years in industrial automation. Then again, the applications engineers relied on me to understand what the machine DOES and to define the customer needs while I relied upon them to understand the wiring diagrams and how signals flowed to accomplish our objectives. I understand how a capacitive sensor would have "no net charge" on an AC system because the sine wave continually discharges the cap.

My eyes glaze over when I try to understand how a SENSOR input, which normally looks for continuity/non-continuity, works when the cap is constantly charging/discharging through the chassis? Wiring-wise, I expect the single wire from the sensor goes to one lead while the other lead attaches to ground. How do multiple cap sensors not interfere with each other when they're all discharging through the same ground?

============
Edit:

Praise the Lord and pass the ammunition! I found me an 'Merican parts supplier!
http://www.cafeparts.com/listCategories ... oryID=1549

Rather than manually crossing all the parts, I can do things the much easier way.

============
Edit:

Remember the braided flex line that originally broke? Well, I found what looks like a replacement source. See: http://www.espressoparts.com/variouseur ... nlesshoses
1. How does one tell if the fittings are BSP, metric or SAE (fractional inch) without buying a bunch of thread gauges?
2. What is the correlation between wrench flat size and fitting size? Generally, SAE fittings are sized according to the tube diameter.
3. Many stainless flex lines have an inner lining. Will a high concentration citric acid bath damage anything on the good flex line?

============
Edit:

by allon on Apr 16, 2012, 8:34 pm
The only thing more dangerous than a mechanical engineer with a schematic is an electrical engineer with a caliper...

I like it!

Over the course of my career, I more often heard: "How do you tell if a salesman (or marketing guy or manager) is lying?...

"His lips are moving!"

Hi Glenn,

I am learning as I go, reading here and there - but in general, it seems that there are two types of circuits that are employed (or at least mentioned) when it comes to water level sensing.

The first type senses conductance by ionic transport; these circuits measure current flow through the electrolytes in solution within the boiler. The probe that sticks into the water has a voltage applied to it, referenced to the chassis. The system must be grounded for electrical safety, especially given that a common failure mode is that a heating wire can "burn out" and touch the shell that surrounds it. A very small current flows (microamps, I would think) because of the very high resistance of the water to current flow (megohms, I would think). This type of circuit is fooled when distilled water is placed in the boiler- no dissolved mineral content equates to no electrolytes. It is also fooled when there are bubbles or froth- they can find their way up to the tip of the probe and depending upon the sensitivity that is present, can trigger a false fill cycle.

The second type is based upon capacitive sensing. I am building one of these up for my La San Marco. It has a sight glass that is about 2 cm in diameter and about 10 cm long. There is a rod that runs down the middle of the tube to hold the system together. This rod is at ground potential.



If a small piece of copper foil is placed on the outside of the sight glass, along the length of the tube, and is not grounded, then a capacitor to ground is formed. As the water level changes, the dielectric between the plates changes.

In reading about how to measure the capacitance, it seems that there has been a big shift in how this variable capacitance is measured over the past decade or two (since the advent of PIC microcontrollers).

PRE INTEL 4004:

If a single threshold value is desired (like, turn on the water when the boiler says I'm empty) an astable multivibrator (think 555 timer) would have a voltage step that would drive an appropriate resistor that would allow the capacitor to charge. If a voltage threshold were reached before the astable reset, then a signal would indicate "fill me". (Less water, less capacitance, less time to charge the capacitor; more water, greater capacitance, longer time for voltage to climb to max).

For instances when a level (continuous read of level vs threshold detection of a preset level) was needed, then something different was done. The variable capacitor (that is what the varying fluid level presents) would be in the tank circuit of an LC oscillator. A variable frequency oscillator would result; as the water level went up the frequency went up. This AC would then be frequency filtered, rectified, and the average voltage would drive a meter. Maybe you have heard of the Theremin http://en.wikipedia.org/wiki/Theremin - a very science-fictiony-musical instrument that made "The Day the Earth Stood Still" sound so creepy. It uses a proximity sensor base upon the variable LC oscillator.

POST INTEL 4004:

Since microcontrollers have become so cheap and ubuiquitos, a different method is now used - it seems like all the time. The controller uses an R-C circuit that exponentially charges voltage over time. However, instead of just sensing a threshold, it reads the voltage just before reset and looks up the fluid level associated with it. It does it over and over in real time. With calibration it sounds like it works pretty well.

The fun thing about this is that for my La San Marco, I should be able to figure out in advance - at least roughly - what the capacitance is with no water, and when full of water, because the geometry is so well defined. Air has a dielectric constant of 1 (almost the same as vacuum); water about 1.75. If I continued the copper foil all the way around, I would have a coaxial cable. If I just put a piece of wire up the back, I have a pair of parallel lines. Both have formulas published for the distance of separation and the dielectric. It is at least a starting point.

Like you, I am finding this project fascinating. Good luck with getting your machine running. Fortunately for me, I have coffee to drink while this beast gets constructed, so I don't have to be in a terrible rush to get it fixed. I'm having fun and learning a lot.

Regards,

Joe