suggestions for high-resolution tiltmeter (inclinometer) sensor?

[rhb]

I'm getting ready to build some of these and have been studying the NASA Tech Brief schematic.  It *looked* obvious until I looked at it closely.

Did anyone work out a theory of operation?  I'm having trouble understanding the circuit concept.

[JBeale]

In general terms it is a differential capacitance meter. It does work, but my experience is that the ultimate sensitivity is significantly limited by mechanical (or possibly chemical) effects on the inside of the bubble capsule. I used four different capsules from 3 different sources and they all had "sticky" behavior which limited resolution to a much lower level than just the electronic noise would otherwise indicate.  They are still pretty sensitive though.

In case of interest- a few months ago I encountered a significantly better design, although the electronics are much more complicated. It has been developed by several people over a period of 10+ years and is called FMES (fluid mass electrolytic seismometer) and is intended as the name suggests, to be seismometer with a horizontal velocity output obtained by integrating and bandpass-filtering the tilt signal.  However, there is an internal node voltage that is proportional to DC tilt, so it can be used as a sensitive tiltmeter as well (although the signal does drift with temperature).  I have built three units now and I think they work pretty well. Note it takes a LONG time to fully settle down after initial startup, like three days or more. Maybe internal bubbles resolving and/or electrolyte concentration gradients diffusing.

Description of FMES: https://docs.google.com/viewer?a=v&pid=forums&srcid=MTEwNTY3NjcxMDQ3NjEzMjgwMDYBMTMxMjUwODI3MDY3MzY1NTcyMDEBUFdScUMtNnRFZ0FKATAuMgEBdjI&authuser=0

Discussion of FMES: https://groups.google.com/forum/#!topic/psnlist/ih4p__5cGW0

[rhb]

What did you use for the diode bridge in the NASA design?  Those appear to be used as matched 1 pF capacitors.  And unobtainable except as NOS parts.

I had originally thought it was a typical capacitance bridge, but on closer inspection, I'm quite baffled as to how it operates.

Thanks for the link to the other design.

BTW have you considered that the settling time might actually be very low frequency physical motion?  It might be interesting to examine data over several days after dropping a 50 lb bag of sand on the floor 10-20 ft away.  Big earthquakes make the earth ring for weeks.

Have Fun!
Reg

[RoGeorge]

Any diode will do it, as long as all 4 of them are as identical as possible.  Those are Ge diodes (and thus the +/-0.7V range == 2 x Vf for Ge) but Si should work just fine.  The circuit is supposed to be symmetrical in order for the DC output to be linear with the ratio between the two vial's capacitive arms.

Another thing, the device output voltage is linear only around zero, or as long as the output DC is very small in comparison with Vf.  For bigger capacitance differences between the vial arms, it won't be linear at all.

There will always be an AC component over the DC one, but the AC is supposed to be attenuated by the low pass RC filter on the right end of the schematic (0.1uF and 100K).

Another thing to consider is the impedance of a 0.01uF at 300KHz, which would be about 53 ohms.  It is 2 orders of magnitude lower than the 5K6 resistor series with the bridge.  If we set the generator for 300KHz and 5Vpp, this will be about 50mV of AC across the diode bridge.  Small enough to consider any diode as a linear device for a swing of only 50mV around whatever decent DC static point.  This small enough swing (and the symmetry of both the voltage waveform and the bridge arms) will make the whole thing a nice charge averaging circuit (one vial arm will have a bigger C because of the vial's tilt, thus its corresponding capacitor will accumulate a bigger charge then the other arm, because everything else is supposed to be symmetric).  The two charges Qleft arm and Q right arm of the vial will end up averaged by the 0.01uF put across the bridge.

The first 0.01uF, in series with the generator is there just to be sure no DC possible offset from the generator will go into the bridge, and vice-versa to be sure no DC generated by the bridge will go back into the generator (a DC voltage pushed back into the generator's output would be bad, because it might affect the waveform symmetry for example - depends of the generator, but most of all we don't want any DC offset from the generator to go to our Vout).  Symmetry in both the waveform and the bridge arms (identical transfer functions for all 4 diodes) are essential for this circuit.  The waveform itself is not important as long as it's symmetrical and has no DC.

The 0.01uF inside the bridge is to integrate the charge imbalance over many periods of the 600KHz bridge switching.  Ideally, the output voltage would be to be read across this diagonally 0.01uF inside the bridge, yet the author choose to read the DC voltage over the whole RC (5.6K and 0.01uF), thus adding the full 5Vpp swing from the generator over the DC we need.  This full AC will be diminished by the 100K and 0.1uF RC low pass filter at the output.

My guess is this AC pollution was done to keep a common ground between the AC generator, the ground electrode of the vial and the DC voltmeter (for better shielding and noise rejection of the whole thing).  If it were to have a floating DC voltmeter, will be much better to read the voltage across the bridge directly, without including the series 5k6 resistor.

Forgot to mention, the vial arms' capacitances are suppose to be very small in comparison with all other capacitors in the schematic, probably pF if not fractions of pF (no idea how small, didn't plugged any numbers).  The absolute value is not important, only the ratio C_left/C_right of the vial.


Later edit:
Trying to proof reading the text afterwards, and realized that, without proper schematic and numbering of the components, it is all just a big TL;DR.
 

https://www.docdroid.net/enKb2ed/nasaames-simpletiltmeter.pdf

[rhb]

Very helpful.  I bought 10 of the CA3039 diode arrays via eBay this morning.  I'm going to start work on a Wien bridge oscillator for the signal.  Probably the 2nd or 3rd iteration in Jim Williams' paper.

The thing that's messing with my mind is the topology of the diodes.  I went through over a dozen texts on electronics measurements today and could not find anything that remotely resembled the NASA circuit.

Everything had the detector across the arms of the bridge and no connection inline with the source and ground.

  I'd love to know more about the circuit's origins.  As I understand what you wrote, the diodes are being used as small matched capacitors and should never turn on.

[JBeale]

For what it's worth, this photo shows the parts that I used for the diodes:
https://www.eevblog.com/forum/projects/suggestions-for-high-resolution-tiltmeter-(inclinometer)-sensor/msg1583374/#msg1583374

[RoGeorge]

I didn't lookup the datasheet for CA3039 before, and wrongly assumed they are discrete Ge diodes, but now I see it's an array of Si diodes (Vf is 0.7V in the datasheet).  Now I don't understand why the expected DC range of Vout specified by the author is +/-0.7V and not +/-1.4V.  Will give it a second look tomorrow, do a SPICE simulation, and try to come up with a cleaner explanation.

Anyway, I considered those diodes rather as a bridge of (controlled) resistors than as a bridge of capacitors.

[rhb]

It occurs to me that there may be a NASA patent that explains it.

When I redraw the bridge part leaving out the input and output sections I'm dumbfounded.  The nodes that I'd expect to be connected to the detector are instead connected to ground via the sense capacitors.  It is unlike any bridge circuit in ~17 texts  which included the NBS electrical measurements volume from1963 and Terman & Pettis.  My library is rather disorganized, so I *might* have a book that describes the circuit, but at this point I think it's probably original.

I was impressed before.  Now I am simply in awe. This is not your Dad's capacitance bridge.  Whoever designed this was *really* good.

I'm left going, WTF?   We know from experiment that it works, but it looks like nothing I've ever seen.  At this point I have to rank this as being in Jim Williams or Bob Pease territory.  Or for the ignorant, an act of God.

Reg,
much befuddled.

BTW I found a half bridge circuit built around a Signetics NE5512 that looks as if it would work quite well in this application in "The Modern Measuring Circuit Encyclopedia" by Rudolf Graf.  I'm quite sure it came from a datasheet or app note.

[IDEngineer]

I was impressed before.  Now I am simply in awe. This is not your Dad's capacitance bridge.  Whoever designed this was *really* good.

This sort of offshoot results from investment in the space program. I used to receive NASA Tech Briefs, and every issue was packed with mind-expanding innovations. Many were patented, but patents expire (now) after 20 years, and then they're public domain. This sort of thing is exactly why I'd much prefer tax dollars being invested in the space program than in many of the other, ahem, things that occupy the federal budget. This kind of development benefits everyone on the planet.

[rhb]

I redrew the circuit as the 0.01 uF & 5.6K resistor forming a low pass filter with a diode bridge across it.  If we consider a 5 Vpp signal input at 300 KHz, there is 50 mV across the bridge which is not enough to turn on the diodes.  They still conduct enough though to charge the fF caps formed by the vial electrodes.

I'm not convinced that analysis is correct, but it's the best I've been able to do so far.  How that becomes a measurable DC voltage still eludes me.

I'm clearly going to be spending a lot of time studying the circuit and reviewing the behavior of the PN diode in minute detail especially at low forward potentials.

I'm also going to have to learn to use LTspice and similar.

[smithnerd]

I'm also going to have to learn to use LTspice and similar.

When this circuit was being discussed last year, I decided to model it in 'Qucs', in order to learn how to use that software:

https://en.wikipedia.org/wiki/Qucs

I can't say that I understand the circuit much better (I'm not an EE), but I did experience a few 'ah-ha' moments, while plugging different component values in.

[RoGeorge]

A simulation won't reveal how it works, but it is helpful here to quick check hypotheses:  assume it works in some way, then make some predictions according to the hypothesis, then check the predictions using the simulation in order to see if the prediction is fulfilled.

So far I get it how it works, can explain it in a short novel, but don't have yet that clean and intuitive explanation in just a few words.  Still looking for that, if I'll ever found one.

As a slightly offtopic thought, I found out why the reverse voltage of CA3039 diodes is so low (about 5V).  The diodes are most probably made from transistors with collector-base junction short circuited, like in this other array of diodes, CA3019 (see starting from page 94 of 379, ICAN-5299, Application of the RCA-3019 Diode Array http://the-eye.eu/public/Books/Electronic%20Archive/RcaLinearIntegratedCircuitsAndMosfetsApplications_text.pdf )

I feel that that diode bridge with capacitors on it is something that I should have known by now, and back in the days of analog era it was probably a basic topology for all kinds of stuff, from power rectifier to RF modulation, yet I'm bamboozled by this particular schematic.  It's easy if we think about each diode as a controlled switch series with an average resistor, but not clean enough for a nice intuitive explanation.

[rhb]

I put a 0.25" strip of copper foil and a 1.25" strip on the vial.  After cutting the 0.25" strip in the middle and doing a probe calibration  I get 3.9 pF when the bubble is at the unconnected end and 3.4 pF when it is at the connected pad end.  So the total capacitance change is around 500 fF.  I doubt those are accurate values, but they do give an order of magnitude estimate.  The values repeated when I rotated the sheet of cardboard that the vial and LCR45 were sitting on.

The attached photo shows the vial with the copper foil applied.  The foil proved very difficult to handle, but maybe with practice I'll get better.

There was never a patent issued to the authors, so no explanation from that direction. The CA3039 diodes have a reverse bias capacitance of 1 pF or less.  I've attached the datasheet for convenience.

I'll have 10 RCA CA3039s in a day or two and can complete construction of the basic circuit.  My initial focus is understanding the circuit operation, so I'll just do it dead bug style on a piece of single sided PCB with everything stuck down with hot melt glue.

Once I've convinced myself I understand how it works I'll tackle how to mount  the two vials I have and build a stable PCB with 2 bridges and an MCU to read the data and log it while I consider what to do next.

The aim of my project is to use two sets of orthogonal levels to measure the surface deviation of a surface plate.  One set of vials will be attached to the surface plate so that as the floor moves under the plate the MCU can calculate the difference between the planes determined by the reference vials and a movable tester with 3 point contact with the surface using sapphire swivel feet.  Initially I'll just display strike and dip of the plate surface along with relative elevations of the feet.  Later I'll figure out a way to get location so I can generate a contoured surface.

I'm using 2"/2 mm vials instead of the 5"/2 mm vials in the NASA Tech Brief.  Assuming 10 cm between supports on the test probe, I should get resolution on the order of 10 nm!   So I might have to buy some lower resolution vials.

I bought a 7" x 14" Chinese mini-lathe which I plan to rescrape to be accurate to 0.0001" and then install a lapped plain journal bearing to replace the OEM deep groove ball bearings.  That should provide me with a convenient and reasonably inexpensive high precision toolroom lathe.   

I bought an ER-32 collet chuck to replace the 3 jaw and a full set of collets.  I'm also going to make a drawbar MT 3 mount for fixture plates for holding larger work.  That setup should allow me to make good differential screws, unlike my drill press based attempt.

Reg

[mycroft]

I found this patent: Diode-quad bridge circuit means (US3883812A) https://patents.google.com/patent/US3883812.

[IDEngineer]

I get 3.9 pF when the bubble is at the unconnected end and 3.4 pF when it is at the connected pad end.  So the total capacitance change is around 500 fF.  I doubt those are accurate values, but they do give an order of magnitude estimate.

Varying the size of the bubble would likely change the dynamic range - not that doing so is really an option after-the-fact! 

Maybe dramatically shorten the leads, get the meter closer to the DUT? I'd expect 500 fF of change just by waving your hands near the test leads in the photo.

[rhb]

@mycroft found the rosetta stone.   I didn't craft my patent search very well.

The operation is gradually becoming more clear from just studying the circuit and considering an order of magnitude analysis.  The patent should make it all clear.

As I remarked, this is not your Dad's capacitance bridge.

This was designed to be very cheap. There are a lot of ways to measure small changes in capacitance.  It will be very interesting to compare different sensing circuits. For the purposes of measuring very small deviations from flatness of a surface plate, this circuit should work very well.

As the CA3039 is hard to obtain I plan to compare that to the implementation @JBeale linked using the 2 diode arrays.  It should work just as well with those.

I'll work out some vial attachment that makes it easy to test different bridge circuits.  After a year of buying test gear like a drunken sailor, it's time to build stuff.

[babysitter]

See
https://www.eevblog.com/forum/metrology/volt-nut-meeting-2019-in-stuttgartgermany/msg2253138/#msg2253138

EEVBLOG Forum User Branadic will talk about it.

[RoGeorge]

I found this patent: Diode-quad bridge circuit means (US3883812A) https://patents.google.com/patent/US3883812.

Thanks, starting from that one, there are a few more similar patents, some much older (1955):
https://patentimages.storage.googleapis.com/79/b2/c4/dfca62cf25a852/US3883812.pdf
http://www.freepatentsonline.com/3545275.pdf
https://patentimages.storage.googleapis.com/b2/5f/4f/7d758d81925b9e/US3271669.pdf
http://www.freepatentsonline.com/3318153.pdf
http://www.freepatentsonline.com/3260934.pdf
https://patentimages.storage.googleapis.com/44/e5/7b/5182b82aa9a452/US3012192.pdf
https://patentimages.storage.googleapis.com/c0/82/cd/7ece43283ed779/US2929020.pdf

Later edit
A new one, forgotten from the initial list:
https://patentimages.storage.googleapis.com/71/41/61/b940d5c5a526b3/US3869676.pdf

Some are almost the same, I don't know come they are considered distinct patents.

[mycroft]

This is a very interesting article by  the authors of patent US3883812 (Harrison and Dimeff): A Diode‐Quad Bridge Circuit for Use with Capacitance Transducers. You can find it at https://aip.scitation.org/doi/10.1063/1.1685975

[IDEngineer]

This is a very interesting article by  the authors of patent US3883812 (Harrison and Dimeff): A Diode‐Quad Bridge Circuit for Use with Capacitance Transducers. You can find it at https://aip.scitation.org/doi/10.1063/1.1685975

...and for $30 you can actually read it! 

Any chance you have the actual PDF file available outside of a paywall?

[RoGeorge]

This is a very interesting article by  the authors of patent US3883812 (Harrison and Dimeff): A Diode‐Quad Bridge Circuit for Use with Capacitance Transducers. You can find it at https://aip.scitation.org/doi/10.1063/1.1685975

Wow, thanks again!
Now, seriously, how do you find these?
What search terms/search engines do you use, please?

[RoGeorge]

...and for $30 you can actually read it! 

Any chance you have the actual PDF file available outside of a paywall?

"Diamonds Are A Girl's Best Friend", and Sci-hub is a researcher's best friend. 

[IDEngineer]

Now, seriously, how do you find these?

With respect to patents, once you have one patent, additional references are easy to find and you can quickly become buried in data. To just scratch the surface, read the patent itself - most contain a list of references to prior and/or related art. Chase those down and you can quickly develop a 3D network of dozens/hundreds of related references. Some will be patents (domestic and foreign), others will be standard publications of all kinds.

Then go down a 4th orthogonal axis into the same or related Patent Class(es), which is how the PTO categorizes applications. These will be related but perhaps not with 1:1 mapping to the original. And of course all of these searches can then stem from any of THOSE references too.

If you want to go further, hit the PTO's website and look at the file wrapper for any of the patents (original or referenced). That will reveal not just more patents, but even non-patent references that the Patent Examiner reviewed while prosecuting the patent application. Each of those, in turn, can reveal more references that THEY relied upon, ad infinitum.

You can go REALLY deep down some of these rabbit holes. I've lost entire afternoons and evenings chasing one thread or another like this. It's intellectually stimulating but the hours pass quickly....

EDIT: All of the above searches are totally free. The applicant or the PTO may have paid to find the references, but the results of their research is in the public domain.

[mycroft]

I searched by authors and diode bridge. There are some jewels in this article, for example:

Second, it makes little difference in the
performance of the circuit when used with capacitance
transducers if the excitation waveform is sinusoidal, sawtooth, triangular, square, rectangular, or whether it has a
symmetrical or an unsymmetrical duty cycle.

This is a very interesting article by  the authors of patent US3883812 (Harrison and Dimeff): A Diode‐Quad Bridge Circuit for Use with Capacitance Transducers. You can find it at https://aip.scitation.org/doi/10.1063/1.1685975

Wow, thanks again!
Now, seriously, how do you find these?
What search terms/search engines do you use, please?

[jmaja]

I make high resolution inclinometers mainly for marine use https://www.jmsensors.fi

Resolution is 0.001 degrees (RMS at 10 Hz, much lower after averaging). Accuracy 0.005 degrees + 0.5%. Long term stability (years) better than the accuracy. Measuring range +-17 degrees. The basic model is 1 k€ (USB connection, Windows software and IP68 aluminium enclosure). I don't think you can find cheaper ones with similar specs.

I also make an OEM model, which can be temperature calibrated for very good offset stability (typical +-0.01 and max +-0.02 degrees -20C - 80C).

I use a MEMS sensor and calibrate each sensors at 1 degree intervals.