Date: Wed, 1 Jan 1997 08:54:19 -0800
From: Mike West <mwest@cdsnet.net>
Subject: Gauges, Accuracy

The accuracy of gauges.

 One of the brothers recently pointed out some fairly cheap
ways to get a gauge in the system.

Kudos to Mike Jones for bringing it up.

I want to cover another aspect of gauges tho, that has nothing
to do with their price etc.

In the world at large, it is considered foolish to put a remote
gauge in a system without a direct reading gauge right at the
unit being measured to check it against.

This is where your life or the life of some expensive machinery
is at risk.

So right along side that remote sending unit will be a big direct
reading mercury thermometer. Or a Bourdon tube press. gauge.

VW engines don't really fall in that category unless you're
poor like me.  :-)

The gauges we use, ie. elect. sender to elect. gauge are remote
gauges. It isn't too practical to put in a direct gauge as vibration
will eat the thing.

The temp sender is alright as far as construction is concerned
but the pressure sensor is a type called "Aneroid", a diaphragm
chingadero. I was warned about those in grade school.

Ok, we use them and rely on them, I suppose.

Let's see what we can do here.

One more item of interest: all gauges read the most accurate at
the 60% range. Which means if you want to read a 60 psi circuit
you put in a 100 psi gauge and sender.

Now we look at calibration.

Temperature gauges are easy, we have two standards in our lives
called "frozen" and "boiling".

Stick the sensor in ice water, (with cubes) to check the acurracy
of that end and in boiling(just rolling) water to check the other.

We're not so lucky when it comes to standards to check a pressure
sensor and gauge.

In the industry, there are 2 ways. One is against a master gauge.
One that has been calibrated and is certified.

The second method is on a "dead-weight" tester, which is how
the master gauge got calibrated.

A dead-weight tester is a one inch piston with fluid under it,
and weights are stacked on it pushing the fluid into the gauge
you're testing.

I wish I could pull a rabbit out of the hat and tell you how
to whip one up out of cardboard but there is no chance.

Your best shot here would be to buy a very good bourdon tube gauge
for a "master" and take it to a lab and have it calibrated.

Then you could test your sensors and other gauges against it.

I suppose I should mention that your gauges are reliant on a
voltage source. Some more than others.

You have low voltage, I would expect your gauge to read low.

Since they work in the millivolt range I doubt that's a big issue.

Location of senders has already been covered by various members,
no point in belaboring that again.

Recomendations? Don't count on your gauges too much.

Periodically check them against a standard.

Put in a fitting where you can take a direct reading to check
them one against the other.

that's all, west  :-)