Excel spreadsheet for random vibration Grms calculations
Ryan Simmons
For anyone doing random vibration analysis, this spreadsheet may come in handy.
http://analyst.gsfc.nasa.gov/FEMCI/random/grms.xls
You can also find general random vibration information at this page.
http://analyst.gsfc.nasa.gov/FEMCI/random/
________
___ __ \_____ ________ ________
__ /_/ /__ / / /_ __ `/__ __ \ Ryan Simmons
_ _, _/ _ /_/ / / /_/ / _ / / / Ryan@ singollo.gsfc.nasa.gov
/_/ |_| _\__, / \__,_/ /_/ /_/ http://analyst.gsfc.nasa.gov/ryan/
/____/ (Oops, there's a space in my address)
Odilo Vazquez
Ryan thanks for the useful info, it will be much appreciated. I have been
triing to get people to clarify the PSD versues ASD for some time now with
no definitive results.
Thanks Odilo Vazquez
Ryan
In article <01bcef0d$47f0ada0$bccbaccf@odilo>, ovaz...@erols.com says...
>
>Ryan thanks for the useful info, it will be much appreciated. I have been
>triing to get people to clarify the PSD versues ASD for some time now with
>no definitive results.
Hey, that one's easy!
ASD is Acceleration Spectral Density, which is G^2/Hz. That's the actual
values used to determine the Grms value.
PSD is Power Spectral Density. That's literally the power that is measured
during a test. Accelerometers don't really measure acceleration, they
measure power output levels, you know volts or whatever (sorry, I'm not a
testing or electrical type). But since the power output is directly
proportional to the acceleration, the same plot can be used for both. The
equipment just needs to be calibrated properly so that the PSD output
corresponds to the appropriate ASD output.
I love engineering!
Ryan
Ryan.Simmons@ gsfc.nasa.gov
(Oops! How did that space
get in my e-mail address??)
Mike Porter
Ryan wrote:
> In article <01bcef0d$47f0ada0$bccbaccf@odilo>, ovaz...@erols.com says...
> >
> >Ryan thanks for the useful info, it will be much appreciated. I have been
> >triing to get people to clarify the PSD versues ASD for some time now with
> >no definitive results.
>
> Hey, that one's easy!
>
> ASD is Acceleration Spectral Density, which is G^2/Hz. That's the actual
> values used to determine the Grms value.
>
> PSD is Power Spectral Density. That's literally the power that is measured
> during a test. Accelerometers don't really measure acceleration, they
> measure power output levels, you know volts or whatever (sorry, I'm not a
> testing or electrical type). But since the power output is directly
> proportional to the acceleration, the same plot can be used for both. The
> equipment just needs to be calibrated properly so that the PSD output
> corresponds to the appropriate ASD output.
[picky mode on]
Actually, accelerometers output an electrical signal that is proportional to
the acceleration. The instrument that reads this signal squares the signal to
get G^2 which is proportional to the power. PSD is the power (G^2) divided by
the bandwidth of the measurement (Hz) leading to the normal units of PSD =
G^2/Hz. Instead of using G's as the unit, ASD uses the acceleration in what
ever unit system is being used (e.g. FT/Sec^2 or M/Sec^2).
[picky mode off]
:-)
--
Mike Porter
DYNAMIC ANALYSIS
Stress and Vibration Consulting
Services
FEA Consulting and Training
www.dynamicanalysis.com
(913) 341-3269
Mike Mitchell
I think accelerometers typically use a piezo-electric crystal. My understanding
is that the output from a piezo lattice is a voltage proportional to the strain
rate -- not strain. Therefore, I'm led to believe that such devices are really
reporting the integration of the change in acceleration not the acceleration
directly. The proof is that they read zero at rest in a gravitational field. Is
this true?
William Letendre
John wrote:
> >The accelerometers I'm familiar with use a capacitive principle.
> >Consider two plates that overlap. A capacitance exists between the
> >plates that is a function of the distance between them. One of the
> >plates is rigidly connected to the base, while the other is
> >(theoretically) free to move.. As the base accelerates, the distance
> >between the plates changes, as does the capacitance, which is measured.
> >The genius behind these accelerometers is that there is an added circuit
> >that adjusts the charge on the plates based on the capacitance. As the
> >plates move closer together, a charge builds up that repels them.
> >Therefore, the plates really don't move at all! These accelerometers
> >are micro-machined, and so are very small (the size of an average IC)
> >and actually quite cheap. (BTW, they don't use just two plates, but two
> >series of plates.)
> >
> >-Paul
>
> Do you know the manufacturer of any such device?
>
> You are describing something close to the way piezoelectric
> sensors work when coupled to charge amplifiers.
> Are you sure your facts are straight?
>
> Regards,
> John
Analog Devices makes this gadget on a silicon chip, along with preamps &
demodulators, using micromachining technology. Original application was sensor
to fire gas gnerator for automobile air bag. Since this unit was designed to
fire on impact, sensed fairly high accelerations (~ 50 G). AD got so many
requests from vibration analysis weenies for a more sensitive version that
they came out with a 5 g part a few years back. Part number, IIRC, is ADXC05.
W Letendre
Dir Eng
NEAT
John
Thank you for the lead. On the Analog Devices web site,
http://www.analog.com/search97cgi/s97_cgi,
I found a very informative data sheet for their ADXL05 and on
page 9 was the theory of operation which described clearly the
"force-balanced " system discussed in this thread.
Regards,
John
Dave Martin
Accelerometers for virbartion testing are typically piezoelectric due to thier small
mass, linearity and frequency response range. Typical PSD spectra for vibration
testing range for 20-2000 Hz, which is compatible with the useable badnwidth of
electrodynamic shakers, and the range of interest on most applications. So DC
response is not needed. For DC response (ie measuring sustained acceleration) a servo
accelerometer is needed, with attendant mass increase and supporting electroincs (see
any good avionics or inertial navigation text). Piezoelectric accels depend on a
variation in capacitance to produce a voltage proportional to the acceleration. The
capacitance variation is caused by straining of the piezoelectric material under the
foprce produced by a proof mass bonded to the material, being accelerated by the
object to which the accelerometer is attached. DUe to the impedance of the charge
amplifier the piezo accel and the resistance of the piezo material itself, the finite
charge across the piezo material will bleed off, but with a timeconstant many times
the max period of the accel;s minimum useable frequency. Endevco may still publish a
good set of technical papers on peizo accels, selction and usage.
Dave
Mike Mitchell wrote:
--
Dave Martin
dma...@fia.net
David Wuchinich
Mike Mitchell wrote:
>
> I think accelerometers typically use a piezo-electric crystal. My understanding
> is that the output from a piezo lattice is a voltage proportional to the strain
> rate -- not strain. Therefore, I'm led to believe that such devices are really
> reporting the integration of the change in acceleration not the acceleration
> directly. The proof is that they read zero at rest in a gravitational field. Is
> this true?
>
> Mike Porter wrote:
>
> > Actually, accelerometers output an electrical signal that is proportional to
> > the acceleration. The instrument that reads this signal squares the signal to
> > get G^2 which is proportional to the power. PSD is the power (G^2) divided by
> > the bandwidth of the measurement (Hz) leading to the normal units of PSD =
> > G^2/Hz. Instead of using G's as the unit, ASD uses the acceleration in what
> > ever unit system is being used (e.g. FT/Sec^2 or M/Sec^2).
therefore can not provide static output voltage analogs of strain. The
voltage they produce is indeed proportional to strain. But, under
static, or quasi static conditions, the strain can be obtained by
producing a voltage output as the integral of their charge displacement
(for example connecting the output to a charge integrator or op-amp).
The stain is then the division of the this response divided by the
thickness of the crystal. As the the crystal thickness is fixed the
voltage. produced by integration, which will be stable at least over
several minutes, providing a femtoampere op-amp is employed, is the
strain analog.