weird voltage reference design idea

[robert.rozee]

hi,
    i happened upon the following (none too impressive) project:
http://essentialscrap.com/vref/
which mentioned creating a current standard using a Kibble Balance:
https://en.wikipedia.org/wiki/Kibble_balance
part of a Kibble Balance being a Gravimeter:
https://en.wikipedia.org/wiki/Gravimetry#Gravimeter

(btw, the mention of a Kibble Balance reminded me of a documentary i watched recently about the redefining of the kilogram)

i wonder if a simplified version of this sort of setup could be used as an alternative way of creating a precision voltage reference? NOT an accurate reference, but just a precise one. ie, like the LM399.

imagine a spherical neodymium magnet suspended above an energized coil at a precise height. how well defined would the required current through the coil be? how easily could we determine the exact resistance of that coil? and how precisely could a modern MEMS Accelerometer determine local gravity (or weight of the magnet)? if all three could be measured with the required level of precision...


these are just some half-baked ideas at 4am in the morning. but i'd be interested to know if the idea is interesting!


cheers,
rob   :-)

[TimFox]

The strength of the permanent magnet depends on temperature.
Some information:  https://www.kjmagnetics.com/blog.asp?p=temperature-and-neodymium-magnets

[Kleinstein]

Magnets also get weaker over time, as they age. Modern magnets can be relatively stable, but are still not perfect.

Another problem is that the gravity is not constant, but there are natural variations, e.g. from the moon causing tides. Chances are than on a short time scale the system would also detect seismic motion.  The gravity sensors may be of short time, but can also drift. So a compensation of changes there is not very feasable.

So I don't think this would be a practical solution.

[Conrad Hoffman]

I can think of various ways to do this that don't use permanent magnets, but they all seem to be ways of producing a temperature sensitive and uncertain reference, at great complexity and high cost!

[robert.rozee]

I can think of various ways to do this that don't use permanent magnets, but they all seem to be ways of producing a temperature sensitive and uncertain reference, at great complexity and high cost!

i must confess, i was inspired by the idea (in the first link) of a Kibble Balance being used, along with a standard weight, to obtain a precise reference voltage. the 'real deal' is a piece of incredibly complicated mechanical equipment whose cost (and ongoing care) would far exceed the utility of a voltage reference used my the average folks on this forum.

however, i wondered if by severely compromising on accuracy, while maintaining precision, one could arrive at some comparatively simple mechanical construction to achieve a precise (yet uncalibrated) voltage; for example, a small chunk of steel might have a mass that varies over time by fractions of a PPM, the distance between two metal plates separated by glass standoffs may be similarly precise. and, of course, thanks to GPS we can obtain the most precise timing intervals.

could these easily obtainable things be combined to obtain a precise/invariant voltage?


cheers,
rob   :-)

[Conrad Hoffman]

I have to point out the obvious- there's a completely mechanical solution that's quite accurate and easily available, or at least it used to be. The D'Arsonval meter movement!

[TimFox]

A "current balance", without ferromagnetic materials or permanent magnets, comprises two coils in series.  The current through the two coils produces a force between the two coils.  One is fixed and the other is on one beam of a gravitational balance (like a laboratory analytic balance).  This allows calibration of a current (not voltage) by calculating the geometry of the coils and gravitational balancing against standard weights.  (Of course, the actual weight of the standards depends on the local gravitational acceleration.  That can be measured with a Kater's pendulum, requiring a time measurement and measuring the distance between the two knife-edge suspension points.) 
The classic d'Arsonval ammeter usually uses a permanent magnet and a watch spring, with their temperature and aging problems.
Refs:  https://en.wikipedia.org/wiki/Kater%27s_pendulum  and   https://nistdigitalarchives.contentdm.oclc.org/digital/collection/p15421coll3/id/231  https://en.wikipedia.org/wiki/Ampere_balance

[mendip_discovery]

Have you got any lego lying about?

https://www.nist.gov/si-redefinition/kilogram/nist-do-it-yourself-kibble-balance

[Conrad Hoffman]

For our purposes, what I'd worry about with the coils is a change in dimensions with temperature. I have to assume that would change the force and ultimately, the voltage.

[TimFox]

When NBS used current balances, they knew how to minimize and correct for thermal expansion, and how to maintain the temperature in the calibration laboratory.
The tricky bit is to measure the physical dimensions accurately.
The physics definition of the Ampere comes from the force per length L between two wires of length L with uniform spacing d carrying the same current I:
F= µ₀ I² L / (2pi d) =(2 x 10^-7) x I² L/d
Therefore, if in the thermal expansion, all dimensions of the assembly increase by the same factor, the overall change in the total force F cancels out, since the L/d ratio is constant for a uniform dilation.

Again, current balances are only used to measure current.  See pp 854-855 of the 2nd edition of Physics, parts I and II, D Halliday and R Resnick, J Wiley and Sons, 1966.

[robert.rozee]

i've read that gravity on earth's surface varies by approximately 0.5%. the earth is a 'squashed sphere' shape due to rotation, and no doubt the local shape is also affected by the current location of the moon.

if we obtain our location and altitude via GPS (accurate horizontally to within a few meters, vertically about 10m), does this provide enough information to correct for gravitational variations?

we can also measure temperature quite accurately, if our balance is made out of suitable materials that we are assured return to the original dimensions for any given temperature, can we calibrate out any dimensional changes due to temperature?

as a mechanical reference point: an automatic surveying level (with a suspended prism within) is accurate to within about +/- 1.5mm over 30 metres (+/- 1/16" over 100'). this translates to roughly +/- 50PPM without any temperature corrections and designed to operate in outdoors conditions.

addendum: we can just use a precision resistor to convert the coil current into our reference voltage. again, we can calculate out temperature errors. i envision a small computer being attached to carry out all the corrections!


cheers,
rob   :-)

[TimFox]

Local gravity also varies with subsurface material, such as iron ore deposits.
Kater's pendulum is an elegant way to determine g locally.
Also note that if the coils are designed to expand and contract isotropically, the dimensional variation cancels out, making it a good current reference.
To stabilize the current at a value corresponding to the weights on the other side of the balance, one can mount a mirror on the moving bar and make an optical lever to a pair of photosensors, using the difference as an error signal to the servo:  no digital computations needed.  This can be made very sensitive by using a long "lever arm" (distance from mirror to sensors).

[robert.rozee]

from here:
https://earthobservatory.nasa.gov/images/3666/earths-gravity-field

yes, it appears that the earth's gravitational field varies by up to +/- 60mGal (from the scale on the graph at the above link), and that "one mGal (milligal) is equal to an acceleration of one thousandth of a centimeter per second per second, which is approximately one millionth of the average strength of Earth’s gravity field (980 centimeters per second per second)."

so equivalent to around +/- 60PPM, sufficient to be significant. i guess a Kater's Pendulum can not be readily avoided, at least for calibration purposes whenever our voltage reference is moved to a different location... unless one can obtain an extremely accurate gravity almanac!

how close is the initial idea getting to being unfeasible?   


cheers,
rob   :-)

[mendip_discovery]

Not s problem, local gravity can be accounted for.

https://www.isobudgets.com/how-to-calculate-local-gravity/

[Conrad Hoffman]

Near the bottom of this page is a link to data and gravity maps- https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/grl.50838