suggestions for high-resolution tiltmeter (inclinometer) sensor?

[JBeale]

I would expect that a glass vial "spirit level" always uses alcohol. If it used plain water, it would likely break if it is ever stored below freezing. I guess you could test this by storing it in a freezer! I expect the "stickyness" of the bubble edges is related to the interior glass surface cleanliness. For example I'd expect the sticky behavior if there are random small patches on the glass that have different surface energies, due to different molecules locally exposed on the surface.  If the glass is perfectly smooth and also clean, the bubble should not have a preference for any position except balanced at the very top.

[branadic]

The content is either alcohol or hydrocarbon, such as Diethylether.
As expected the vials from UK arrived earlier. Yeah, time to start applying some copper electrodes and the PCap01 board. Results coming soon.

-branadic-

[rhb]

The inside of my 4"/2mm vials is frosted.  It is clearly visible with a 14x Hasting's triplet. So that is likely a source of sticking.  The need to fully polish the inside may be a significant cost factor in the manufacture of the commercial microradian sensors.

[branadic]

My setup works incredibly good and shows very high resolution. My setup currently lays on a table with some aluminium blocks on each side of the vial. What I can say for now is:

1. I need the V-block as fast as possible
2. I have to move the setup to our basement and to the grounding
3. The electrodes are slighly different, that can be seen by the response due to inclination of each single capacitor and the calculated difference, so glued electrodes are not the best choice for the resolution I get with the IC, but that is
4. I have to develope a temperature compensate for the setup

-branadic-

[JBeale]

The inside of my 4"/2mm vials is frosted.  It is clearly visible with a 14x Hasting's triplet. So that is likely a source of sticking.  The need to fully polish the inside may be a significant cost factor in the manufacture of the commercial microradian sensors.

Yes you are quite right. I see this also, now that I used a magnifier on it.  The inside surface of the vial has a bumpy or pebbly texture so it is much rougher than the outside surface. I think you can get away with that amount of roughness, if you only use the vial visually.

[metrologist]

I managed a 30x photo with the red line clearly hovering over and reflecting off the inside surface, that matches the photos presented above. The texture (highs/lows) seems orders of magnitude more granular than the measured response, 2/3 arcsec.

[rhb]

It will be interesting to hear what reports are on the UK made vial.  I suspect that a minor amount of post acquisition processing would clean up the steps in a satisfactory manner.  Especially so if they occur at the same spots.

0.0001"/10" is a very demanding spec for a hand level and you're pushing way beyond that.

[branadic]

The W. A. Moyer & Sons Vintage level vials  arrived today as well.
For comparison JBeale, what is you repitition rate? I currently configured PCap01 reading two capacitors plus compensation measurement. This results in a repitition rate of 6.1Hz including temperature measurement.

-branadic-

[JBeale]

I am only logging data once every 2.1 seconds. The Keithley 196 DMM itself converts and updates the local readout at maybe 5 Hz (?) but there is a digital low-pass filter option I have enabled, giving a final time constant of a few seconds.

[rhb]

It occurred to my while getting lots of face time with my pillow (norovirus :-( that a small solenoid that got pulsed just before a reading might solve the sticking problem.

Can we get closeups of different vials?

[JBeale]

Next week I should have some UK-made vials to compare.  The price difference is only about $20 per vial so for the small quantity I want, I can't justify a lot of extra design effort.  A vibration source should do something to reduce the sticking, but if you're trying for sub-micron positioning of the vial for long-term monitoring (my application) but then whacking it for every reading, it feels like an unavoidable tradeoff against how rigid and stable the mount can be made.  I should add the one Aliexpress-sourced, so-called 4"/2mm vial I measured was actually about 2x more sensitive than claimed. Maybe they ground the glass with coarse grit only and did it for 1/2 the proper time, but anyway just to underline I think that source's process and quality control is roughly zero.

EDIT: The photos on the link below are the type of vials I received.  The close-ups I posted above are just a higher-mag look at these same parts, showing the pebbly texture of the interior surface.
https://www.aliexpress.com/item/Size-14-96mm-Accuracy-4-2mm-Glass-High-Accuracy-inclinometer-Tubular-Bubble-Level-Cylindrical-spirit-level/32558917931.html

EDIT_2: FWIW I'm interested in measuring things generally, as in the old C.L.Stong AmSci columns that I enjoyed also, but in particular my motivation here is catching the earliest possible hint of a structure (my house, specifically) slip-sliding away downhill.  I know at high resolution there will be diurnal effects, temperature, humidity, irrigation, precipitation effects, seasonal effects etc. also, which is all OK.  The tiltmeters used on volcanoes may use other technology because they resolve better than 0.01 micro-radian, that is 0.002 arc-second. https://earthscience.stackexchange.com/questions/10220/how-are-data-from-tiltmeters-used-to-monitor-volcanic-activity

[rhb]

Did you ever post a photo of the Chinese vials as received?  I don't recall seeing one.

I certainly agree about the extra cost of the UK vials if they are made to a better standard. 

Do you live near a volcano or fault zone?  Those are the principle application for high precision tiltmeters aside from monitoring susbsidence from underground pumping of oil or water.

You've rekindled a childhood interest in whole earth geophysics generated by C.L. Stong's book of Amateur Scientist columns.  I am feeling the need to build a seismograph or two.

[branadic]

Measured a first characteristic curve of my UK vial  (3"/2mm) in the range of +/-0.06° and in steps of 0.001° today. Forgot to copy the values of the reference sensor to my USB stick for comparison, but will do that on monday. Measurement for temperature compensation algorithm and long term stability is actually running during weekend, so I will have this data on monday as well.

The x-axis equals a measurement time of 3:48h.

-branadic-

[metrologist]

That is amazingly linear. I'm more impressed with such a setup than the response of the vial. How is such accomplished?

±0.06° is ±216 arcseconds, or ±54 divisions on the vial, and in roughly 4 arcsecond steps. The bubble will be forced into the end long before the end of a cycle, so I am a bit surprised it's that linear. 

[branadic]

±0.06° is still within the grinded range of the 11mm x 96mm (100mm total length) vial, without the bubble moving to the ends of the vial. I measured against a Nivel230 on an inclination test bed we have for high resolution inclination sensors. To be honest, the vial wasn't adjusted, layed only on two grinded parallel spacers and showed an offset in zero position of the test bed.
The curve looks much noisier than it actually is, it's just a plot artefact. I have to verify the positions with the Nivel230 data to make a statement on accuracy, but it's quite promissing for such a cheap solution.

The test bench is a granite slab with three spheric mountings underneath, two of them are moved by pushers up and down and have glass scale sensors with 50nm resolution for control and the third mounting is fixed in height, thus the slab is inclined by the pushers around the third mounting.

-branadic-

[JBeale]

Measured a first characteristic curve of my UK vial  (3"/2mm) in the range of +/-0.06° and in steps of 0.001° today.

That plot looks quite good; it is a promising result. There may be some hints of nonlinearity over the displayed range but not easy to see on this plot. I'm curious what it looks like with even smaller steps.

[branadic]

Not to forget all the imperfections by the copper tape electrodes cutted by hand, but yes that looks really good. I prepared the second 3"/2mm vial with a little more attentation to detail while applying the copper tape electrodes. I also prepared one of the W. A. Moyer & Sons Vintage level vials and will measure them both soon.

-branadic-

[metrologist]

±0.06° is still within the grinded range of the 11mm x 96mm (100mm total length) vial, without the bubble moving to the ends of the vial.

Maybe you have a different vial. Mine has 9 lines each side and estimated 9 divisions between the graduations, a total of 27 divisions of 4 arcseconds. 108 total arcseconds the bubble moves between ends, or 0.03 degrees end-to-end, ±0.015.

At any rate, I hear my extra vial calling for such experimentation.

[branadic]

Yes my vial is different, I linked it in the post above.



The bubble can move more than just between the outer divisions without touching the ends of the vial.

-branadic-

[rhb]

That is a much finer grind on the inside surface than the Chinese vial I have.  It's not polished,  but your tests suggest it is good enough.

[branadic]

Exact, you can see the dim area of the vial.

-branadic-

[branadic]

Added plot of the reference sensor. I come to the conclusion, that 30s for measuring a single step is to short. The bubble or the test bench hasn't stabilized during that time.

-branadic-

[JBeale]

In case of interest, a visual reference for some of the vials recently discussed here.  All of them have some visible texture to the interior surface (eg. ground, but not really polished)

https://www.ebay.com/itm/W-A-Moyer-Sons-Vintage-Level-Vial-Type-L-1-A31314-Bubble-Level-NIB/192433183719

https://www.aliexpress.com/item/HACCURY-Size-14-100mm-Accuracy-4-2mm-Glass-High-Accuracy-inclinometer-Tubular-Bubble-Level-Cylindrical-spirit/32558917931.html

5726/101 – Ground vial, Ø15x95mm, sensitivity 0.02mm/m
5725/101 – Ground vial, Ø15x96mm, sensitivity 0.05mm/m
from https://www.leveldevelopments.com/products/vials/ground-vials/

EDIT: A preliminary overnight readout test of the 0.02mm/m vial from UK shows it also has "sticky" behavior. When the tilt is changing very slowly (eg 1 or 2 arcsec per hour) as in geophysical applications, the output displays sudden jumps on the order of 1 arc-second, similar to what I saw before with the 4"/2mm aliexpress vial.  I suspect you do not encounter this if you do a test with large enough or rapid enough steps that the bubble never quite stops moving. I am now thinking if I could make a mounting that has a continuing wobble of say 5 arcsec, to keep the bubble moving just slightly, and then average it out.  In the below graph, the tilt overnight through mid-morning is roughly tracking with ambient temperature, but I believe it to be a true tilt signal and not just a readout tempco, because it passes through the null point (output = 0.000 volts).

[branadic]

Got some feedback from W.A. Moyer Company today. They couldn't find anything, but the vial must have been a special Military Spec ordnance vial made for the U. S. army during WW2. The numbering is very close to the numbering system used by a company named Keuffel & Esser who made optical equipment.

Looks like I need some measurement to calculate sensitivity.

-branadic-

[rhb]

From the size I'd guess that the Moyer's vials were  made for laying in mortars in anticipation of the invasion of the Japanese mainland.  Experience in the island hopping campaign had made the military planners *very* pessimistic about casualties and material losses.  If we invaded, they did not want to run out of equipment.  So they ordered mind boggling quantities of stuff.  Thankfully it was never needed.