In a message dated 97-11-22 08:52:35 EST, vicwhit@octonline.com writes:
> This brings up an interesting topic. Can anyone explain the operation of
> the voltage stabilizer? In my mind, I always thought it operated similar
to
> a voltage regulator and could easily be replaced with an electronic
> version.
Vic:
I'll give it a try! Basically, the old style voltage regulator used in our
LBCs is nothing more than a switch. In operation, it is constantly switching
the voltage to the gauges on and off. If the battery voltage is 12 volts, and
the regulator switch is on 50 % of the time, and off 50 % of the time, the
average voltage from the regulator is 6 volts. If it is on 67 % and off 33 %,
the average voltage is 8 volts. Since the gauges are very slow to respond,
they don't "see' the on-off action, only the average.
One of the switch contacts is fastened to the end of a bi-metal strip. When
the strip is cold, the contacts are closed. When the strip is heated, it
bends, moving the contacts apart, and opening the switch. Around the bi-matal
strip is a coil of resistance, or "heater" wire. Current through the heater
wire causes it to warm slightly, heating the bi-metal strip, causing it to
bend and open the contacts. When this happens, current through the heater
wire is cut off, and the bi-metal strip cools, reclosing the contacts.
Current once again flows, and the process is repeated over and over again.
Assuming for the moment that every thing is in equilibrium, and the contacts
are opening at the rate of 67/33 %, producing an average voltage of 8 volts
with 12 volts from the battery. If you then turn on your lights, and the
battery voltage drops to, say 11 volts, less current will flow than before,
which will produce less heat at the bi-metal strip. The reduced heat will
cause the bi-metal strip to take longer to open, and the duty cycle will
change, in this example, to 73/27 percent. 73 % of the lower voltage should,
if everything is working properly, produce the same average 8 volts as
before.
The gauges work in a very similar manner, except the bi-metal strip is
mechanically linked to the needle such that the bending motion of the strip
is converted to rotory motion of the needle. The more current through the
heater wire, the warmer it gets, the more the bi-metal strip bends, and the
more the needle moves (that is why the sendors have a higher resistance when
empty or cold than they do when full or hot - just the opposite of what one
would think).
The amount of current through the meter is controlled by the resistance of
the sender unit. As you pointed out, the fuel senser (and the temperature
sensor as well) is nothing more than a variable resister. By its very nature,
though, any resistor is also a "heater." If you put enough current through it
by raising the voltage, and it doesn't burn out, it will get red hot, just
like your shop heater (The biggest difference between resistance wire in a
heater, and any other resistance wire, is the ability of the heater wire to
withstand the temperature required to glow without melting). In a typical
automotive application, the fuel sender dissipates around 1/4 - 1/2 watt.
> Same issue with the fuel sensor. The one that I know of is just a
> wire-wound variable resistor. The amount of current isn't large and I
doubt
> that it heats up and I don't think that you would want a heat source in
the
> fuel tank. The voltage source that drives the circuit (thru the resistor
> and gauge) is the from the stabilized voltage. I don't understand how one
> will offset the other.
The best answer I can give you for this is - they sorta do, and they sorta
don't! Let's take temperature compensation to start with. As with any
experiment, all variables except the one of concern must be kept constant. We
will assume the exact same voltage from the battery, and the exact same fuel
level. The only thing to change is ambient temperature. For the sake of
argument, assume a current flow through the circuit of 60 milli-amps, a
reasonable value for a half full tank, and a temperature of 75 degrees.
Further, assume the fuel gauge reads exactly 1/2 full. To provide this 1/2
tank reading, the 60 ma had to produce a certain amount of heat in each of
the three elements in the loop - stabilizer, gauge, and sender. This amount
of heat brought the bi-metal strips to a given temperature. If we now drop
the ambient temperatur to 20 degrees, for example, the heat produced by the
current will no longer maintain the temperature of the strips, and they will
cool off.
As the strips cool off, the gauge will tend to read lower than before, but
the cooling of the strip in the stabilizer will cause it to stay closed
longer. Since it now stays closed longer, the effective voltage at the gauge
will be higher, causing the current through the loop to be higher, causing
the bi-metal strip in the gauge to get hotter, and, if all goes well, the two
effects will cancel out, and the gauge will again read exactly 1/2 tank. The
amount of heat produced in the sendors is not a factor, as there are no
bi-metal strips associated with these.
Up to this point, Tom O'Malley and I are in perfect agreement. We differ a
little, though, on how well they compensate for temperature. His experience
has been that they compensate extremely well, and my experience has been that
they don't compensate so well. Why the difference, I can't explain, but my
experience with instrumentation has proven that there is no descrepancy here.
There are so many other factors at work, that his gauges could be right on,
and mine could be off.
What other factors? Let's say you go out to your car and start it up. Your
gas tank is almost empty, and the water in the engine is cold. Both the fuel
and the temperature gauges are reading very near the bottom of the scale.
Assume that your alternator is working just fine, and producing exactly 14.6
volts. Now you drive your car 15 miles to the nearest gas station and fill
up. The fuel gauge now reads full, and the temperature gauge is in the middle
of its range. With the alternator still producing 14.6 volts, the current
through the gauges has doubled or more, because of the change in engine
temperature and fuel level. Because of the additional current through the
stabilizer, the contacts will stay closed for a shorter period of time than
before, reducing the voltage to the gauges, causing them to read lower than
they should. The stabilizer has actually introduced an error! In this case,
will the effects of ambient temperature changes be compensated for? If so,
will they be compensated for when the gauges read low? What is the
temperature of the stabilizer strip compared to the strips in the gauges? Do
they both change the same amount with respect to changes in temperature? I
don't know!
Which brings us back to Tom's statement about solid state regulators. As Tom
said, there will be no temperature compensation with these, unless they were
deliberately designed for it. I don't know if any of the units commercially
available have this compensation or not. Does anyone on the list know?
Temperature compensation would certainly not be hard to do.
> Dan Masters, have you got some insight on this? Anyone else?
I have a fair amount of insight into this sort of thing. Back in the early
eighties, I was given the responsibilty for developing and implementing a
program at TVA for determining the accuracy of the instrumentation used for
safety functions at our Nuclear power plants. The simplist calculation for
instrumentation accuracy required over 40 manhours to perform, and the most
difficult took over 200 manhours! The calculations were usually fairly
simple, once we had the required data in hand, but knowing what data was
required, and where to get it, could be quite a daunting task. I would love
to do a calculation on the gauges in our cars, but there is no way I would
ever be able to get the data.
I have never looked at instruments the same since then. As an example,
consider driving accross country at 80 mph, and driving back at 20. You
odometer would show a longer distance coming back than going, because the
diameter of the tires is larger at higher speeds than at lower speeds. At 80,
you cover more distance with each rotation of the tires. Since the
speedometer only records rotation of the driveshaft, the additional distance
is not accounted for. Thoughts like that drive me crazy now!
I would be very surprised if the fuel and temperature gauges in our cars come
close to 10 % accuracy. With the exception of the speedometer, accuracy is
not as important as repeatability, anyway. If you have 2 gallons of gas left
when the gauge reads empty, that is no problem as long as you ALWAYS have two
gallons left. Who knows what the temperature of the water actually is - it's
only important to watch for a descrepancy from norm.
Hope this helps, and is not more than you asked for.
Dan Masters,
Alcoa, TN
'71 TR6---------3000mile/year driver, fully restored
'71 TR6---------undergoing full restoration and Ford 5.0 V8 insertion - see:
http://www.sky.net/~boballen/mg/Masters/
'74 MGBGT---3000mile/year driver, original condition
'68 MGBGT---organ donor for the '74
|