EEVblog

Did you know you can use your frequency counter to detect gravity? You've likely done it before and you didn't even know it!
Dave demonstrates the phenomenon of 2g-tipover on quartz crystal oscillators in an Agilent 53131A frequency counter.
Related videos:
How a rubidium frequency standard works: • EEVblog #235 - Rubidiu...
FE-5680A Rubidium Standard Teardown: • EEVblog #236 - FE-5680...
Forum: www.eevblog.com/forum/blog/eev...
EEVblog Main Web Site:
www.eevblog.com
EEVblog Amazon Store:
astore.amazon.com/eevblogstore-20
Donations:
www.eevblog.com/donations/
Projects:
www.eevblog.com/projects/
Electronics Info Wiki:
www.eevblog.com/wiki/

[@sdgelectronics]

I can detect gravity with my frequency counter by carefully holding the frequency counter then releasing it at a distance of 1m vertically from a concrete surface. If it smashes and stops working then gravity was correctly detected. You can also use signal generators, TVs and hi-fi equipment too.

[@PhillipRhodes]

I prefer to use a rack mount UPS and a human foot. Drop the UPS directly over the foot, and if you hear a loud scream and see blood, then gravity was successfully detected.

[@MrCarlsonsLab]

Great explanation Dave!

[@yoonki]

Fun fact: The same effect is seen on watch movements. This lead to the advent of tourbillon movements, which continuously changes the orientation of the balance wheel to net out the effect of gravity.

[@EEVblog]

I had completely forgotten about those. Wonderful stuff for 1795!

[@foxtrottNB]

So you can overclock your computer by flipping it up side down? :D

[@EEVblog]

OMG, don't tell the gamer kiddies!

[@teardowndan5364]

LN2 extreme overclocking competition attempts: flip your LN2-cooled board upside-down for that +0.000000001% lead!

[@pocoapoco2]

Of course, the next logical step is to swing the frequency counter around by the cord.

[@queazocotal]

Another smaller effect - gravitational time dilation. Raise it 1m, and it's 10^-16 lower. To get to 10^-12 change you'd need 10km altitude change though.

[@VoidHalo]

Yeah but that would have to be relative to you, meaning the cables would also have to be 10km long. So that would defeat the point of it. Rule of thumb I've heard is about 3-5ns per foot. So for the sake of convenience, let's call it 10ns per meter. Which for a 10km, or 10,000 meter cable works out to an extra 100us delay. Which is pretty huge when you consider even a 1ghz clock will oscillate every 1 nanosecond. So in the time it takes for the signal to travel down that line, your computer will have had 100,000 cycles.

[@w2aew]

fascinating - I had no idea!

[@jongmassey]

Are you sure it's not just because the electrons fall out of the crystal? ;)

[@jasong1984]

Very interesting phenomenon. Great explanation with the DaveCAD too :)

[@sykskysyk]

I *really* enjoyed this video Dave, BIG thumbs up!

[@iangenelly]

Fantastic video!! Thanks so much! As many will say: I use DaveCAD on a daily basis!

[@Lskaggs63]

Very interesting, I knew about temp effects on quartz crystals but I never considered gravity effects, great video!

[@fr0nage]

Wow, super informative video. Thanks!

[@capsbr2100]

That was very interesting for me! Many thx for the video! :)

[@krnlg]

Wow, very interesting - cheers Dave!

[@parnordlund2931]

I love you dave for being so bloody nerdy, top caster at scienceheap for sure :)

[@robertcalkjr.8325]

Very interesting. Thanks Dave.

[@Pwaak]

Very Interesting! Thanks Dave!

[@gregoryhall9276]

even in the same orientation, the frequency will change with altitude. Gravitational time dilation.

[@metaforest]

Thank you for that.... fantastic info for "young players."

[@thingyee1118]

Awesome video Dave. Made for some great night time viewing. Lol. Hope the manufacturer includes a warning or a side note in the instruction manual. Bet they don't often. Very interesting how a design feature of a frequency counter effects its accuracy. I think I have heard about this before. Cheers from England. Tally ho.

[@Minifig666]

I just had to check my calendar to make sure it wasn't April... Very interesting Dave!

[@heinzk023]

I always wonder: How can these counters have a resolution of a millihertz without a gate time of 1000 seconds? Can they measure fractions of a cycle? Does that only work with certain waveforms, e.g. with sine and triangle and not with square? Will Fundamentals Friday ever reveal that mystery :-) ?

[@twomorestars]

Nice, learned something new here. I like it.

[@lb5sh]

"Excuse me lads, I need to tilt my PC... just need to speed up some calculations a bit."

[@Magicride]

I think the real explanation for why this happens Dave, is because when you're turning the frequency counter upside down, the electrons are starting to fall out!

[@hikaru-live]

This is how MEMS accelerometers and MEMS gyroscopes work. Those stuff just measure more exact frequency of the vibration of pure silicon.

[@lawrencemiller3829]

Thanks Dave, I had to check the date to make sure it was not April.

[@rentAscout]

Definitely fascinating. I worked in calibration and never thought about gravity in the uncertainty principle. Wonder if the phenomena is used elsewhere in research...

[@teapotwar]

I love using DaveCAD!

[@redtails]

well if the deviation of the crystal by doing the tilting bale up is known, and it really mattered, it would have been really simply to put a little microswitch in there to add a little correction in the value before it was displayed on the screen

[@raymundhofmann7661]

What about a gravity compensated ovenized crystal oscillator? Will multiple crystals in different orientations connected together do it?

[@sorin.n]

Raymund Hofmann i was wondering myself that too... is it possible?

[@friedmule5403]

That reminds me of the good old CRT TV's sometimes you had to adjust the cannon to compensate for change in the magnetic field, from i.e. China to Denmark.

[@krnlprime]

Просто невероятно. Сасибо за Ваши ролики. Это значит что калибруемый инструмент должен применяться там- где был откалиброван.

[@Cybeonix]

Very cool :)

[@SuperFinGuy]

Cool, it can be used to measure tilt(direction) and acceleration then.

[@TheEPROM9]

It would be rather interesting to do this test on a rolercoster.

[@Pooua]

When I saw the link to this video on EFY Group's Electronicforu, I was expecting something more sophisticated than just the fact that changing the crystal's orientation would change its reading. I was hoping that someone had devised an actual project, making good use of this principle. For example, maybe measuring the force of gravity.

[@arends1874]

Luckily it still works in Australia despite the fact it's upside down! (sorry for the bad joke, I couldn't resist)

[@kennytheamazing]

Could you get a reading of the actual value of G with that kind of lab equipment?

[@slackerbelmont]

Cool Video, I noticed a small "chip" in the crystal near one of the pins when you showed it out of the case. Is that from opening it or was that a part of factory calibration?

[@EEVblog]

Not sure, came out like that so can only assume it was factory.

[@DaedalusYoung]

This is heavy, Doc!

[@USWaterRockets]

Interesting. Could this phenomenon be significant enough to lock up or glitch a digital video camera on a high pressure water rocket that subjects it to 250g-300g for a couple of milliseconds at launch? We have lad some videos cut out at the moment of launch in the past.

[@dinkc64]

Excellent video, but I want to know what happens if you lift the entire unit up instead of just tilting it, what happens then?

[@itrstt66]

do the nowadays funtcion generators have integrated frequency counters?

[@pcfreak1992]

What does that GPS receiver do in that frequency generator?

[@chuxxsss]

Look like you have the Garden Island 10 Mhz reference Oscillator, and I think it was on the Cal truck in 1999. Don't forget the GPS lock on the rubidium oscillators which governs it all. But you know that anyway.

[@EEVblog]

I used to work at Garden Island!

[@jackisthebombdrummer]

Dave can you please do a teardown of the Oricom UHF058 cb radio? They are very cheap on ebay and I would like to know what you think of it.

[@mukundalini]

at 2-m height of difference laser interferometers an interference would appear in 24 hours of hatchet. I think the shown effect deals with magnetic field

[@Vidicon31]

so it works a bit like a efficient mems accelerometer ?

[@nraynaud]

We're talking about a very small drift here, at such high resolution, isn't the atomic clock becoming a viable option? Otherwise, as discussed in other comments, gluing an MEMS accelerometer on the crystal box and mapping the drift seems like a simple solution (as long as we don't go into vibrations territory), what would be the cost impact? doubling the price?

[@ChumpusRex]

What sorcery is that frequency counter doing to measure with 1mHz resolution with a 1second integration time?

[@jordanrodrigues8265]

Phase detection of a nice clean square wave.

[@GregBurrowa]

Gate time and resolution: I would like to have someone help me understand how counters arrive at their resolution. When displaying the difference between two clocks the the counter updates all ten decimal digits each second, shouldn't this measurement require one thousand seconds to perform?

[@SimonJ57]

It's the electopns flowing fast when you tip it back.

[@TheGrunt76]

It is quite intriguing that gravity can affect such a small and light object that drastically (drastically here is very relative, of course). I really didn't know this before, but I'm not electronics expert either. I wonder how big part gravity attributes to regular quartz crystal wrist watch time error in normal use, which is of course quite small, but still? Those crystals must also be calibrated in certain position, am I correct?

[@yaghiyahbrenner8902]

Dave Cad now supports gravity simulations.

[@DelTapparo]

Very interesting. But what was with the blinking red LED in the box and the white paper? flapping around?

[@EEVblog]

That's the LED on the rubidium, you can see it through the fan!

[@paulrpg]

Does this mean having really stable quartz crystals in space would be quite challenging? How do you get precision reference crystals in an environment where the orientation can vary?

[@MrGrimv1G]

You've moved it closer to your lights... There is a gradient to account for.

[@victornpb]

I'm surprise you didn't mentioned that the earth gravitational field vary depending on were you are, like he says "it can be a big deal".

[@mooseteets]

it nearly zero'd out when it was sideways, is there no way to turn the internal oven sideways and get it to zero perfectly ?

[@RiverWyvrn]

Scary. they should mount bubble levels to the cases of these things, or digitally detect if they're being tilted.

[@MsHUGSaLOT]

How do crystals like this measure it's vibration with just 2 leads? Does it work like switch going on and off at a predictable frequency?

[@KeenanTims]

Since the physical construction of the crystal is (largely) symmetric, and we're measuring what I'd assume to be a "bulk" property, I'd expect the 180' flip to produce the same frequency as the deformation etc. would be the same magnitude. Is this due to inevitable asymmetry in the construction? The way the crystal is driven electrically? Curious effect. Thanks :)

[@theonlyari]

So, can the 53131A see the difference in gravity if you say put a very large mass of say, iron, next to the instrument? I suppose that could give you an interesting way to measure the supposed "gravity waves"-- assuming you put your measuring devices very, very far apart.

[@user-pm9xo7eo9q]

So you need to use a bubble level to do correct measurement?

[@yazicib1]

What kind of application (other than calibration) may require milliHz accuracy? Could you give a couple of examples?

[@jordanrodrigues8265]

Satellite navigation and surveying is the big one.

[@pvc988]

I wonder how much this affects quartz hand watches. These things are exposed to some significant accellerations while being worn.

[@TheCheesyProductions]

In what applications would one require such high accuracy for frequency? Millihertz seems ridiculously precise.

[@TheShadow1975100]

Gravity......Queen of physical forces of the Universe !

[@RealationGames]

That's another thing to account for in a space shuttle. I guess it's the 7236412445th thing in the list IIRC.

[@uriituw]

I know Agilent are a spin-off of HP, but I noticed the HP logo on the handle.

[@cliffjones7868]

My question is, by not having the top and bottom covers in place, would the effect the wires and capacitance of the circuitry influence the test demonstration? Would there be an added influence of the earths magnetic field. Or even the gravitational field of the sun, as we orbit in an ellipsoid path, or the Moon in the same respect? Just curious. ;)

[@Cnctrldotcom]

I'd be interested to now if it's possible to measure the strength of the local gravity accurately using this method. I'm guessing not as it looks like it's not far outside the noise of the oscillator. Presumably you can get higher frequency and more accurate oscillators though?

[@redtails]

higher frequency won't give you more digits on your counter. We're already talking parts per billion with measuring 2G difference. Measuring for instance 0.001G would therefore require parts per trillion accuracy :D. Probably possible... though not in this setup

[@Cnctrldotcom]

***** Thanks, and I feel a bit stupid for suggesting using a higher frequency, ppb is of course independent of frequency as you say. Good video though as it prompted me to read a bit on gravimetry. It looks like 1g is give or take 1000gal and the earths gravitational field varies by about +/-50mgal. A ppt sensitive frequency counter would need to be calibrated for it's location on earth and it's height! Totally impractical I'm sure but interesting none the less.

[@X-OR_]

Tell us more about the Frequency Standard ( Rubidium ? )

[@EEVblog]

I've linked in a video of that.

[@X-OR_]

EEVblog Thanks Dave

[@martinda7446]

I would have guessed after scratching head, that it was due to magnetic field...Shows what I know. Both being weak forces tho, wonder if magnetic field has measurable effect?.

[@karolismilieska3863]

That means if you want to measure gravity with frequency counter, you better not use SMD crystal oscillator, because it's mass is greater and has stronger effect with gravity. :)

[@Mpthedawn]

I am coming from the Amp hour podcast

[@gaborendredi8161]

I am mesmerized by the fact that a frequency counter can resolve a 10 MHz signal up to milihertzs in one or two seconds. Can it happen that something else caused the change on the display?

[@jfdjsksgeusj]

I dunno, unless you take it into a place of nil gravity, how do you prove that it's gravity creating the effect EEVblog

[@malgailany]

The frequency counter is Agilent and the handle is HP!

[@teardowndan5364]

HP spun off their instrumentation business under the Agilent brand about 10 years ago. They got rid of their electronic components division around that time too.

[@she0051]

Hi Dave thanks for another informative video :)..... you mention that the serial it's for a printer, would I be correct in guessing it just outputs the display value every few seconds down the standard rs232 8,n,1,9600 bauld.... if so would it not be trivial to write a program that receives the data and appends the time and date to the value and saves it as a csv file for later analysis or graph plotting..... And thanks again spent way too much time watching your videos :) I would be happy to write a program for you if it helps in anyway :) Regards Ian

[@EEVblog]

Ian Sheppard It might be usable in that way.

[@UberAlphaSirus]

Well it does matter to me that the rubidium standard, and such small numbers might have an error with relativity. SO what was done to correct and correlate that data to prove it.

[@PhattyMo]

Does the magnetic field of the earth have any effect? Maybe that's the key to the change,and not gravity? I dunno,just a thought.

[@octavmandru9219]

So the Aussie calibrated frequency counters are basically calibrated upside down :D

[@lucasmun1066]

Just wondering, is there anyway of compensating that problem? I know we can't physically change what "g" is but can't oscilloscopes detect the difference between control value and increase or decrease depending on the displacement of the control value? It just seems very odd to me that engineers can forgot to implement something to compensate this fault. So techincally, a 10khz reading in below the sea level will be different to a 10khz reading 10khz in top of the mt.Everest?

[@SuperFinGuy]

Try moving it around and see how it responds as an accelerometer.

[@cornboy3]

So does this mean if I shake my laptop I actually change my processor clock frequency and speed it up or slow it down a small amount? Very interesting.

[@sydnius]

I’ll have to disagree that there is any effect from gravity on the rubidium hyperfine spectrum. It’s not an oscillation in the same sense as a quartz crystal, hence its general utility for acting as a standard.

[@ctrlaltdel02]

So, if I get calibrated device in, say USA, and move it to Europe, it means it is off its calibration? If I want spot on calibration, it needs to be done where the dewice will be working =)

[@Mythricia1988]

Gravity will be the same regardless...

[@puddingpimp]

Gravity is not the same regardless, but this isn't a demonstration of Gravitational time dilation, the frequency shift is caused by the change in strain on the quartz crystal from changing it's orientation and can be observed transiently during any acceleration. GTD also causes a change in relative frequency but the effect is somewhat smaller (~1ppt for terrestrial variations) than this demonstrated strain induced frequency shift (~300ppt), but is observed by atomic clocks sitting on the top and bottom of a rack (with enough digits of counter). Unless you have a Caesium fountain clock or something more exotic, GTD is below the uncertainty of your calibration. It's also worth noting that the SI definition of a second does not make reference to gravity. Your clock will still count seconds in 100G as it does in 1G, though comparison of the two clocks frequency will show a divergence, leading to the definition and distinction of coordinate time (time elapsed in a coordinated reference frame) and proper time (time elapsed in the clock's reference frame). In the UTC/TAI system (the two systems are syntonic though with a leap second adjustment in broken down time), individual atomic time standards count SI seconds, and their outputs are adjusted with respect to calibrations of their local frame (slowed by 1ppt on average, because most atomic clocks are not on the geoid) to be syntonic, and the output is combined by a weighted average (relative to their uncertainty) to form an ensemble which defines the TAI second which is a coordinate time delineated in SI seconds on the earth's geoid, however defined by scheduled publications of BIPM.

[@Mythricia1988]

puddingpimp No disrespect for your knowledge or understanding about this, but that's a huge amount of content that has no relevance with the original comment at all. Gravity, is, the same either way. The original comment said nothing about elevation difference. He just talked about moving from the item from the US to Europe, which is perfectly possible without changing elevation between those two endpoints. And at the same elevation, the gravity will be the same, and even if there is an elevation change, he's talking about a practical situation, not a theoretical one, so even if there was a change, it wouldn't matter, the change is incredibly small, and massively irrelevant for the vast majority of users of this kind of equipment. Some people just can't answer a practical question with a practical answer.

[@EEVblog]

Not unless you change altitude(small effect), it should remain the same, as the original calibration vector direction remains the same orientation (toward centre of earth) regardless of where you are on the planet.

[@puddingpimp]

My answer was at the start. The change caused by differing GTD is so slight that it won't affect the calibration of normal lab instruments. Neither would the strain shift due to small gravitational changes in altitude. These shifts in gravitational potential are going to be on the order 1/1000, which is far less than the extreme shift in force angle demonstrated in the video. Even this extreme shift is only 300ppt, which is smaller than the error of almost all OCXOs which have a typical manufactured frequency error of 50ppb, though 1ppb parts are available. Anyone who cares about a 300ppt frequency error is going to be using an external reference clock. I would be skeptical if I saw a +/- 300ppt calibration on a frequency counter with an OCXO.since the top-end Agilent frequency counters give a 1 day aging of 300ppt and 30 day of 15ppb (Option 010 on a 53200A Factory cal is 50ppb). With the standard TCXO the factory cal is 0.5ppm with 1ppm/year aging, so the lowest 3 digits are garbage anyway without an external ref clock.

[@xfcisco]

intresting....

[@TonySmith-en4br]

Hey Dave, why not just mount the crystal on its SIDE, then the side weight loading would always be equal if it was tilted. Yes still not perfectly level but I'll bet a huge improvement. How about testing that ? Still interesting. Why didn't you try it on your frequency standard just to see how far it would drift and for how long ? Maybe they should make special padded shock mounts for crystals. Better yet, rotate them ourselves and average the results with a slow servo circuit like the frequency standard does. The oven temp my also change slightly with orientation because of different heat sink characteristics. Maybe they should make the package as a sphere for equal air movement from all directions from gravity. I see a new oven design in the future. A small sphere package packed inside a large soft foam insulation and shock protection.

[@VoidHalo]

For some reason this reminds me of the way CRT displays would go all wonky if you tilted them, or especially if you turned them upside down. Makes me wonder what caused that to happen. I wouldn't imagine the coils would move much, and the electron beams wouldn't be affected significantly by gravity.

[@sinuspl]

It's the earths magnetic field

[@VoidHalo]

@@sinuspl No it's not. The Earth's magnetic field is incredibly weak at the surface and spread over the entirety of the magnetosphere. When was the last time you saw a paperclip stick to the earth? At best you can get it to rotate a needle with very little mass under very little friction, ie a compass. But that's easily negated by pretty much anything significantly magnetic. Even a paperclip that's been magnetized will spin a compass away from the Earth's poles. The Earth's magnetic field cannot and will not EVER have ANY effect on a cathode ray. There has never been any experiment in which a cathode ray has been observed to be deflected by the Earth's magnetic field. I'm intimately familiar with the operation of a cathode ray tubes and electron beams (cathode rays). I will say that yes, they will be deflected a VERY VERY tiny amount. But it would be near impossible to measure without very precise equipment, and the slightest magnetic or electrostatic noise, which is all around us as it is, will disturb the results. The noise would drown out any actual measurement of deflection caused by the field. But I'm sure there have been some experiments to measure the effect of the Earth's magnetic field on electron beams. In fact, I'm 100% positive it's been studied. Particle accellerators, like the Large Hadron Collider, normally use beams of charged particles, in the case of the LHC it's either protons or lead nucleii. Which would react similarly to magnetic and electrostatic fields. And with the size of the LHC, and how accurate the experiments need to be, I'm sure they looked into how much the beam would be deflected by the earth's magnetic field. Like I said, it would be insifignificant in most cases, but with high precision experiments even a perturbation of a fraction of a percent would be devastating to the results. Really makes you appreciate the lengths they must have gone to to make not just that, but all huge particle accelerators work properly.

[@bangonkali]

I wonder how much change occurs when Fighter Jets make high G turns.

[@WAVETUBE84]

So if you had multiple oscillators mounted at various angles, you could use them to detect directional change relative to gravity?

[@EEVblog]

Yep. Wouldn't be terrific but would certainly work.

[@WAVETUBE84]

There's probably, something more accurate?

[@puddingpimp]

MEMS accelerometers and gyrometers work precisely this way. They are electro-mechanical oscillators designed specifically to have large coefficients for frequency shift under acceleration. They also measure the asymmetry of the positive going and negative going wave-halfs to figure out the sign of the acceleration. Turns out counting frequency has less noise than trying to measure the magnitude of deflection by voltage. Quartz is chosen for frequency reference oscillators because it has low coefficients of change for temperature, pressure, acceleration, humidity etc. If you're building an oscillator to measure some other quantity, then you want it to have a high coefficient for whatever change you want to measure. I think typically the resonators in MEMS devices are built as silicon fingers etched into silicon wafers because it's cheap to manufacture, you can build CMOS logic on the same substrate and silicon is virtually immune to damage from repetitive strain. There's always something more accurate (even if no one's built it yet), but it's hard to beat the price of a MEMS accelerometer: ~$3 each 1qty.

[@puddingpimp]

puddingpimp So I got curious and did some research: Not all MEMS sensors work this way, and it seems from a bit of research that gyros typically work this way, while accelerometers tend to work by suspending a conductive mass on a beam between two capacitor plates, and integrating the differential capacitance. As the mass shifts due to acceleration, the capacitance of one side of the balance will increase while the other decreases.

[@christianullrich2923]

So you have to be careful when calibrating frequency references during an earthquake. Good to know.

[@gamingSlasher]

I dont understand why you get the 2g effect when you turn it 180 degrees. Would have made sense if it changed each 90 degree of turn because the physical layout of the oscillator seems to be the same if you turn it up side down. The exception would be if crystals have a one way direction sensitivity to gravity which would be really interesting indeed. That would be something for Ben Krasnow to look into :)

[@EEVblog]

It's about vectors and some complex math and gravitational theory if you really want to get into it. Smart people have already done that and you can look up the info if you are keen.

[@MeakerSE]

You'll start to see atomic clocks in jets now you can get them surface mountable.

[@Razor2048]

Is it possible for a device to be calibrated to deal with multiple angles. e.g, a sensor to detect the orientation and then automatically compensate for it?

[@EEVblog]

Interesting thought, and yes certainly possible, as it is totally charactertisable reproducible, just like TXCO map a cal map that adjusts for temperature, you could also have a GXTO I guess.

[@EEVblog]

***** You'd still need to map for the in-between angles. Much simpler to just characterise it and then map it I suspect.

[@EEVblog]

EEVblog But simpler still to just whack whack a "This Way Up" stick on the product :->

[@teardowndan5364]

There is an easy way to "compensate" for angles if you absolutely need that much precision: use an external reference that is not (as) susceptible to mounting orientation which you toss in your in one of your equipment racks and hardly ever need to touch again... like a rubidium standard.

[@PeregrineBF]

Use an accelerometer to sense orientation, then create the compensation data.

[@RTL8187]

Cool stuff! Is this effect also experienced when crystals are subjected to acceleration? Can you throw your equipment from the top of your building and see what happened? :-D When a fundamental Friday on crystal oscillators?

[@EEVblog]

Absolutely. It's all about acceleration. Could get a good 5 seconds or so of wireless datalogging on the way down...