Shop OBEX P1 Docs P2 Docs Learn Events
What are the effects of gravity on electronics? — Parallax Forums

What are the effects of gravity on electronics?

bambinobambino Posts: 789
edited 2009-01-12 16:33 in General Discussion
Ok,·I am going to stop pulling my hair out for a minute and ask the forum for help. My current life's goal is building an impact recorder. There are plenty around, but this one is wireless and therfore the circuit resides on the device that is incurring the impact. The goal is to withstand 500 g's, but right know I would be happy to get anything over 50, which is what I can get with out a power on reset of my circuit.

I will upload some pictures and schematics of my device later, but for now I am curious if any one has some rules of thumb concerning circuits and gravity.

All thru the developement of this project I have been more concerned with sturdiness of the PCB, and not having components fall off during a drop of the device. Now that I am in the testing phase, I am not having a structural problem, such as I feared, but a power supply problem( or so I think ).


General description of my circuit:

2.5 inch diameter Pcb Disc
Propeller processor
1301 maxim ADAC
Sena Bluetooth Module
Dytran 0 to 500 g accelorometer.

The circuit is encased in a 20 lb steel cylinder and dropped a fixed distance of about 2 feet. The propeller collects the impact data and sends it via bluetooth to a PDA for viewing.

I have·no doubts it's possible due to the feed back gained from military missles which exceed 500 g's easily.
If they can get data at such great distances and velocity, why can't I at· 2 feet and no explosives?

I realize my last question can not be answered without pictures and schematics, but I am at work right now and will have to bring them from home later this week! So any input would be appreciated!
Thanks
«1

Comments

  • SRLMSRLM Posts: 5,045
    edited 2008-12-22 14:53
    There are a few threads around here about putting electronics in model rockets, and one of them discusses some effects here. Not quite as high as you're looking for, but a start.
  • bambinobambino Posts: 789
    edited 2008-12-22 15:09
    Indeed, I was hopeing one of those rocket specialist could chime in and tell me what they thought of my project.

    Of course there's not much they could say at this put without more detail. Even without that I'm sure they have some ground rules as to what works and what doesn't!
  • FranklinFranklin Posts: 4,747
    edited 2008-12-22 16:14
    Make sure all parts are soldered well and the whole thing is protected from vibration from the impact with potting or some other means. use devices that don't have coils to minimize detuning due to impact (might not be possible) or send the data after things calm down.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - Stephen
  • bambinobambino Posts: 789
    edited 2008-12-22 16:34
    Nope, No coils, That I am aware of. Unless it is internal to the bluetooth radio(not likely).
    I do know that the propeller itself is having a POR. The test program I have loaded currently is measuring the battery voltage at the same rate it would normally read the acceorometer, and sending out the average every second or so. that along with an incrementing number so I can detect the reset.

    In normal operation the radio would not be being used until after the drop, but both are resetting!

    I am using Two AnyCap Reulators from analog( one 3.3 being feed by one 5 v) and they get their juice from a 6 AAA Nimh Cell pack filtered by a diode. The only other components up stream is the caps(off the top of my head I think they ara 1uF's in front and behind each regulator).

    Off of the 5V supply is a constant current diode that feeds the accelorometer!

    ReSoldiering is Certainly a possibility, but I used fresh components on a just manufactured board. All clean parts and soldiering equipment.
  • Carl HayesCarl Hayes Posts: 841
    edited 2008-12-22 16:54
    If there are any unsoldered contacts in the system (such as battery holder contacts, mechanical on/off switches, etc.) they are likely to open momentarily on impact. That, of course, could cause a reset.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net
  • bambinobambino Posts: 789
    edited 2008-12-22 17:12
    Yes, The on/off switch is a mechanical switch, but when this problem surfaced in testing, I cross soldiered the switch to always on.
    Any drops I did over 50 g's failed.

    I guess I need to go back and look at my contacts!
  • Beau SchwabeBeau Schwabe Posts: 6,568
    edited 2008-12-22 17:34
    bambino,

    Just thinking out loud... I know that crystals are sensitive to gravity ;·you can change their ppm simply by flipping them over, but I wonder if 50 g's would be enough to warrant a failure of any kind.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Beau Schwabe

    IC Layout Engineer
    Parallax, Inc.

    Post Edited (Beau Schwabe (Parallax)) : 12/22/2008 5:55:48 PM GMT
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-12-22 18:02
    I think any mechanically-resonant frequency source (e.g. crystal, ceramic resonator, etc.) will have problems with shock and intense vibration. You could try a silicon oscillator instead. They're not terribly expensive and boast good accuracy.

    -Phil
  • bambinobambino Posts: 789
    edited 2008-12-22 21:04
    Beau, Phil Thanks. I have not tested for that scenario. Will Do!
  • bambinobambino Posts: 789
    edited 2008-12-22 21:31
    Let me back up here a second! I am asuming that the affects on a crystal would not be the same as for the internal fast clock on the propeller? If it is I would have no way to test for it!
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-22 21:36
    I agree with Phil that a crystal might be affected by a shock - the Piezoelectric Effect could give you unwanted pulses off the clock crystal, or maybe even the lack thereof. In fact, I would like to know if silicon in general would be susceptible to this kind of problem when shock waves are involved.

    Another thought: is shock necessarily the same thing as acceleration??? If I place a golf ball on the end of a steel bar and hit the other end of the steel bar with a hammer, the compression shock wave can travel through the bar and knock the golf ball across the room, yet the bar itself moves very little. The shock moves as a compression wave but the acceleration of the bar is.... what? In a case like this, I would question what it means for the accelerometer to measure up to 500 G's. Mentally, I can picture a compression wave traveling through a chip, yet the overall chip displacement might be next to nil.

    I'm no expert on this; I'm just thinking out loud. I know that in the design of some mechanical systems, engineers sometimes use the concept of "jerk" to evaluate automated movements. Jerk is not acceleration; it is the derivative of acceleration with respect to time.
  • bambinobambino Posts: 789
    edited 2008-12-22 21:41
    I am no guru on this either, although I do know two different kinds of accelerometer's are made and they call them static and dynamic. one for say acceloration and the other for say measureing earthquakes
  • sylvie369sylvie369 Posts: 1,622
    edited 2008-12-22 22:07
    bambino said...
    All thru the developement of this project I have been more concerned with sturdiness of the PCB, and not having components fall off during a drop of the device. Now that I am in the testing phase, I am not having a structural problem, such as I feared, but a power supply problem( or so I think ).

    One common "failure mode" in high power rocketry is the power on reset due to the ejection charges or other shocks (even just the acceleration from the motor) briefly disconnecting the batteries. What are you using for a power supply? How is it connected to the circuit? On the assumption that you're using batteries, how are they contained?

    We usually use 9v batteries, which are fastened to a board of some sort (often G10 fiberglass) with zip ties, one across the battery and one lengthwise. The battery is connected to the altimeter with a standard (but high quality) clip with wires coming out the end. The bare ends of the wires are fed into screw-down type terminals of the sort that you can see here:

    www.missileworks.com/rrc2-mini.html

    (you can also see the construction of a pretty normal altimeter here - notice the hold-down wire on the large capacitor).

    Regular spring-loaded battery terminals will _not_ work. The acceleration will inevitably compress the springs, allowing the battery to briefly move out of contact with the other terminal. Been there, done that, several times.

    Here is a pretty good description of a really cool project, with photos showing the electronics:

    www.vernk.net/Construction/AngelfireGpsPayload.htm

    Here's a much more typical setup, again, with good photos:

    www.rocketryplanet.com/content/view/2466/38/

    Attached is a graph of the acceleration data from a flight I made back in August, collected using a Sparkfun SD card datalogger. You can see it pulled about 23 Gs when the motor lit, then there's a quick shock from +5 to -5 Gs when the apogee ejection charge went off. So far so good. Then some jiggling back and forth at a couple of Gs under the drogue parachute. What you can't see is the effect of the main ejection charge and parachute deployment, because that pulled enough acceleration to knock the SD card out of the socket (as discussed in that thread linked to earlier). Everything after about 48 seconds into the flight was lost.

    The battery for the datalogger was a tiny LiPo cell, simply taped to the mounting board with a couple of wraps of masking tape. It was small enough that even under acceleration it didn't weigh much. But the altimeter battery, a standard Duracell 9V was held down by a metal clamp, which also held it under compression against a pair of terminals.

    By the way, it's also possible that the internal connections of your batteries aren't up to the task. There's quite a bit of variation in the construction of 9V batteries, and I wouldn't be surprised if other cells also have some problems. The Duracell 9Vs are known to be pretty solid.

    Post Edited (sylvie369) : 12/22/2008 10:19:27 PM GMT
    643 x 419 - 44K
  • bambinobambino Posts: 789
    edited 2008-12-22 22:14
    sylvie369,
    I will look into your suggestions tomarrow, I have to give up my internet connection for now>
    Will definately followup with you tuesday!
    Duty calls!
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-22 22:18
    bambino said...
    I am no guru on this either, although I do know two different kinds of accelerometer's are made and they call them static and dynamic. one for say acceloration and the other for say measureing earthquakes

    Just curious: which kind are you using? I'm guessing the "static" NON-seismic version is for relatively smooth movements (no mathematical discontinuities, no sharp spikes) so maybe the dynamic/seismic version would be better for sharply changing motions or shock??? Mine is a question, not an answer.

    One confusing thing about shock is that it is measured in G's, which is a unit of plain old smooth acceleration, too. This might be analogous to measuring air pressure in "inches of mercury". Except for a few industrial places in China and New Jersey, there isn't much mercury in the air, but using the unit of "inches of mercury" could get confusing if you don't understand what the unit really makes reference to.


    You might want to take a look at this:

    books.google.com/books?id=mQmGEVW_TaEC&pg=PT1494&lpg=PT1494&dq=units+of+shock+g+squared+hertz&source=web&ots=XPWjZGclKY&sig=VoPfgPW-vOEgRucwrg2oRFIzFSk&hl=en&sa=X&oi=book_result&resnum=7&ct=result#PPT423,M1

    Again, I'm no expert. I just thought I'd add some food for thought.

    Post Edited (ElectricAye) : 12/23/2008 3:44:36 PM GMT
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-22 22:27
    sylvie369 said...


    By the way, it's also possible that the internal connections of your batteries aren't up to the task.


    Another way to look at it: at 500 G's, imagine 500 batteries stacked on top of each other. Is the chemical inside the battery able to withstand that kind of compression without momentary detachment from the internal electrodes, etc?
  • bambinobambino Posts: 789
    edited 2008-12-23 16:19
    Sylvie369,
    Thanks for the response, I am using a battery clip from radio shack that holds 6 AAA rechargeable cells in series. They are laying flat and are heat shrinked in the package. Because they are laying flat I didn't think of them losing contact due to the springs, but I am know going to zip tie them as well. As for the internal battery contacts and crystal, I am going to load a routine this week that will use the internal clock and watch for a reset. My client mentioned this morning that we could attach an external power source to it to test for the battery failure.

    ElectricAye,
    Again, I am not a good source of knowledge on accelerometer's(I don't even spell it correctly half the time). I have spoken alot with the manufacturer's of them, telling them what my needs where and using what they suggested.
    PCB and Dytran are both very helpful
  • Dave HeinDave Hein Posts: 6,347
    edited 2008-12-23 18:41
    The G-force from impacts can be very high if there is no cushioning.· A 2-foot drop will result in a speed of 11.3 feet/second.· 50 G's of deceleration will occur if the object slows to rest over a distance of 0.48 inches.· This would take 7.1 milli-seconds to go from a speed of 11.3 fps to zero.· 500 G's will occur if the deceleration distance is 0.048 inches, and the deceleration time is 0.71 milli-seconds.

    If your·electronics does not have half an inch of cushioning, then it is experience more than 50 G's.· If there is·less than 1/20th of an inch, then it will see more than 500 G's!· Most small objects can survive fairly high G forces.· I did some experiments with an egg, and it was able to handle 100 G's.

    Spring loaded battery holders can lose contact even with a drop of a few inches.· 9-volt batteries work well because of the clip that is used.· A capacitor could be added to the circuit to maintain the voltage during the few milli-seconds that power may be lost.

    You mentioned that you do some filtering, and the result is read out once per second.· The deceleration time period is on the order of a milli-second.· You would need to sample at a rate of a few thousand times per second to capture the acceleration data.· Any filtering will reduce the value of the peak acceleration measurement.· I don't know the characteristics of your accelerometer, but many devices have a frequency response of less than a killohertz.

    Dave
  • sylvie369sylvie369 Posts: 1,622
    edited 2008-12-23 19:03
    bambino said...
    Sylvie369,
    Thanks for the response, I am using a battery clip from radio shack that holds 6 AAA rechargeable cells in series. They are laying flat and are heat shrinked in the package. Because they are laying flat I didn't think of them losing contact due to the springs, but I am know going to zip tie them as well. As for the internal battery contacts and crystal, I am going to load a routine this week that will use the internal clock and watch for a reset. My client mentioned this morning that we could attach an external power source to it to test for the battery failure.

    If the AAAs are held between two spring terminals (or one spring terminal and a hard terminal), that's an obvious culprit. I wouldn't expect heat shrink wrapping to help - under high G force, those tiny batteries weigh quite a bit. When I needed to use a couple of AAs for a radio receiver, I used these (attached) holders that I got from American Science and Surplus. I don't see them there anymore on the website, and I didn't notice them when I was at the retail store a couple of days ago - I suspect they're long gone, but you might find something of that sort elsewhere.

    Otherwise you could perhaps rig something to connect to the ends of the batteries that moves with the batteries (like a 9V clip moves with the battery that it's clipped to). That's the ideal solution. Anything set up so that the battery itself wants to move while the battery holder doesn't is going to be problematic.
    3264 x 2448 - 668K
    2448 x 3264 - 1M
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-23 19:16
    Bambino,

    I'm baffled by something. You stated that your Dytran accelerometer goes from 0 to 500 G's. But if you look at Dytran's SHOCK accelerometers, they start at Max = 2500 G's and run up to Max = 70,000 G's. Again, I wonder if somebody (i.e. your sales dude/dudette) hasn't become confused by the units being used. Just because shock happens to use the same units as linear acceleration, which just so happens to use the same units as gravitational attraction, does not necessarily mean they are all the same thing. Similarity of units does not always equate to an intersection of similar physical "realities".

    IMHO, that is.



    smile.gif

    Post Edited (ElectricAye) : 12/23/2008 8:48:37 PM GMT
  • bambinobambino Posts: 789
    edited 2008-12-23 23:14
    Thanks for the response's guys keep it coming.
    I am sorry I don't have the time today to answer everything. I am going to come in early in the morning, so until then!
    I'm not dodging you ElectricAye, My device from them actually isn't listed on their site. It is surplus from a batch made to be implanted in the head of oil rig drills to determine drill speed vs torque. Will elaborate more later.

    Dave, the once per second is just a test program to detect a reset. You are correct, the window I am measureing is 22 mS long at 57 Ksps.
  • Carl HayesCarl Hayes Posts: 841
    edited 2008-12-24 16:06
    ElectricAye said...
    Bambino,

    I'm baffled by something. You stated that your Dytran accelerometer goes from 0 to 500 G's. But if you look at Dytran's SHOCK accelerometers, they start at Max = 2500 G's and run up to Max = 70,000 G's. Again, I wonder if somebody (i.e. your sales dude/dudette) hasn't become confused by the units being used. Just because shock happens to use the same units as linear acceleration, which just so happens to use the same units as gravitational attraction, does not necessarily mean they are all the same thing. Similarity of units does not always equate to an intersection of similar physical "realities".

    IMHO, that is.



    smile.gif
    Dimensional similarity doesn't necessarily mean that quantities are the same (foot-pounds of torque are not the same as foot-pounds of energy, for example), but shock, acceleration, and gravity·certainly are the same.· Each is·rate of change (first derivative) of velocity.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net
  • bambinobambino Posts: 789
    edited 2008-12-24 16:20
    Ah Christmas eve, and the store is busy busy. I did take time last night to snap off a picture of the finished app. I will have to gut it out this weekend to reprogram, so I'll take another pic or two of the board and battery pack while I'm in there.
    1280 x 960 - 173K
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-24 17:55
    Carl Hayes said...

    Dimensional similarity doesn't necessarily mean that quantities are the same (foot-pounds of torque are not the same as foot-pounds of energy, for example), but shock, acceleration, and gravity certainly are the same. Each is rate of change (first derivative) of velocity.

    Carl,

    I'm looking at a rock sitting on my desk. Right now it is experiencing 1 G of gravitational "acceleration". Despite my best efforts, however, I'm having a hard time computing its first derivative of velocity. I can't figure out what it's velocity is and with respect to what. Any ideas?


    roll.gif

    Post Edited (ElectricAye) : 12/24/2008 6:01:00 PM GMT
  • Carl HayesCarl Hayes Posts: 841
    edited 2008-12-24 18:44
    Gravity is an acceleration. Gravity is accelerating your stone downwards, and the desk is pushing the other way, accelerating it upwards. The net is zero, assuming your desk isn't falling and isn't in a rocket. Where's the dilemma?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-12-24 19:01
    ElectricAye,

    What you seem to be running up against is Einstein's "equivalance principle", which states that gravitational and inertial (i.e. dv/dt) mass and acceleration are indistinquishable from each other. Consider his famous "gedanken experiment", in which you are placed in an elevator car and told that you are either sitting on the ground or accelerating upwards in outer space at a 1G rate. Would you be able to tell the difference? The answer is "no", there's no way to tell. Even a beam of light will be deflected just as much by gravity as it would appear to be deflected in a non-inertial frame of reference , such as the accelerating elevator car.

    The same holds for the mechanism of any accelerometer. There's no way for it to "tell" what is causing the acceleration it's experiencing, be it gravity or a change in velocity visible to an outside observer.

    -Phil

    Post Edited (Phil Pilgrim (PhiPi)) : 12/24/2008 7:08:52 PM GMT
  • Carl HayesCarl Hayes Posts: 841
    edited 2008-12-24 19:14
    Precisely so.· We are (mostly) talking in terms of classical physics -- a fixed inertial frame approximately equivalent to the surrounding countryside -- but ElectricAye wishes a formulation that would be the same in any reference, inertial or non-inertial (accelerating).· Einstein wanted that, too -- a way to describe the motion of objects that would be the same [noparse][[/noparse]invariant]·in any reference.· Einstein eventually succeeded in formulating such a description.· It's usually called General Relativity (but Einstein preferred to call it Theory of Invariants).·

    ElectricAye, I'm flattered that you think I can explain General Relativity in the space of a post on this Forum.· But even if I could, it wouldn't help the original poster solve his resetting problem, would it?· I think we've led his thread a little astray.· Apologies.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net
  • Nick McClickNick McClick Posts: 1,003
    edited 2008-12-24 19:41
    Dave's acceleration calculations are pretty helpful - What I bet they do with missles n' such is use an assembly (foam, spring, whatever) that lengthens deceleration by a known amount. It seems to me to be a better design to apply the minimum required force to the measurement equipment.

    If you're sure it's only 50G's, though, something is breaking contact. On Nintendo Wii controllers, they use a 3300 uF cap to provide a power supply when the user breaks the power connection from all the swinging.

    Hope it helps!!!

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Gadget Gangster - Share your Electronic Projects - Sign up as a Designer and get a free 4-pack of Project Boards!
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-24 20:03
    Phil Pilgrim (PhiPi) said...
    ElectricAye,

    Consider his famous "gedanken experiment", in which you are placed in an elevator car and told that you are either sitting on the ground or accelerating upwards in outer space at a 1G rate. Would you be able to tell the difference? The answer is "no", there's no way to tell.
    -Phil

    Since this is the Sandbox and nothing serious, allow me to play here a little more:

    The gravitational vectors in the unmoving elevator would not be parallel: they would all be pointing toward the center of earth. Imagine the elevator is 10000 miles across. At one side of the elevator, the acceleration vector of gravity would be pointing in a direction that is not parallel to the vectors seen on the elevator's opposite side. You could tell you are being pulled by a planet and not simply purely accelerating across "open" space because of this difference in vectors.

    This might sound like a trivial argument, but the fact is that Navy subs have been using gravity gradiometers for several decades. Gravity around the world is not the same; there are tiny variations due to geology, etc. The bottom of the ocean has been gravity mapped and the subs use the gravity gradiometers to help confirm their positions. I used to map gravity and the change in gravity from your left foot to your right foot is something I could measure without much difficulty.... and that was 25 years ago.

    Frankly, I'm not smart enough to know how Einstein's thought experiment panned out the way it did. The man had flashes of insight that he then worked backwards from using mathematics. His was a rare mix of intuition and mathematical abilities, but I'm not sure the elevator analogy is all that good when you really look at it closely. Furthermore, there are some other such shruggings I've seen in quantum mechanics that leave me feeling similarly uneasy. But... for the most part the theories give us something to work with.... at least, that is, until somebody comes up with something else.




    smile.gif
  • ElectricAyeElectricAye Posts: 4,561
    edited 2008-12-24 20:39
    Carl Hayes said...

    ElectricAye, I'm flattered that you think I can explain General Relativity in the space of a post on this Forum. But even if I could, it wouldn't help the original poster solve his resetting problem, would it? I think we've led his thread a little astray. Apologies.

    Well, I'm trying to illustrate how gravity is not really the same thing as acceleration, though the units are the same and in some cases you can use the same kind of instrumentation to measure both. And from there, I'm encouraging Bambino to ascertain that he is indeed measuring the thing he really intends on measuring. If his accelerometers are designed only for smooth accelerations such as seen by gravitational measurements, then he might be screwed if he's actually looking for sharp transitions as one might see in an impact. The fact that the Dytran shock accelerometers have maximums that start at 2500 G's suggests to me that he MIGHT be seeing data dropouts because his particular accelerometers aren't really built for shock, or jerks, or impacts, or whatever you wanna call it. For example, if his "drill bit" accelerometer was designed to measure merely the spin of a drill string, well, we're not talking about impacts are we?

    Probably, tho, it's something stupid... like a battery clip or a bad solder joint. But what's a Sandbox for if you can't play around with other ideas?
Sign In or Register to comment.