Accelerometer Advice Needed

[carl]

Hi All,

I wonder if any of you can steer me in the right direction regarding a new project I will be working in shortly.

The project involves integrating an accelerometer (not I.C. based, but an already manufactured unit) to a PIC (16 or 18).

The problem is:

High operating voltage - typical 18-30V
High output bias voltage - typical 8-12V

Some examples of what I mean are here (voltage output types):

https://www.ctconline.com/__intrinsically_safe_accelerometers.aspx?qcwg=116_365&qman=

http://www.wilcoxon.com/vi_index.cfm?PD_ID=12

http://www.hansfordsensors.com/products/accelerometers/ac/top-entry/

Does anyone have experince or suggestions how these type of accelerometers can be interfaced with a PIC?

The main problem is the high bias voltage and dynamic range of the output (can be +/-5V or more) - i can't see an easy way to interface this output to a PIC?

Thanks
Carl

[asmboy]

the first link is to a standard accelerometer ,
which is meant to be used with a CURRENT amplifier - which in turn can be scaled for any range of supply voltage you care to use. say 5V ??
there simply is no 'high voltage' issue with IT at all.


the second link is a self powered unit,
and in that case SCALING the output voltage OR capacitor coupling with DC restoration ( my choice) will make it very easily PIC compatible, whatever 'DC bias' may be present. the analog circuitry to do all this is not very complicated. been there - done that for vibration measurement systems.

You do understand however that acceleromters are at their heart AC output devices- and that with no acceleration - ie steady state - there is no output change.

there should be no issue other than sampling frequency vs output delta V
to deal with really.

you have never used one before, correct??

[carl]

Correct never used one.

Been looking at datasheets of various types today.
Understand that mainly the output is AC centered around a bias voltage which varies proportionally to the vibration measured.

The application will be for assessing rotating bearings - and as you said dealing with vibration.

So basically the accelerometer types posted WILL HAVE to have some analogue circuitry on the front end to change it into a suitable output signal for the pic. Could you possibly eleborate on this a bit more or comment on which type (current or voltage) in your opinion would be easier to implement.

EDIT: The two types that I am considering:

Current: http://www.pruftechnik.com/condition-monitoring/products/sensors/product/industrial-accelerometers.html?cHash=2bc4b419e1

This comes with a ‘Current Line drive, 3.5 mA closed current with superimposed AC signal’


Voltage: http://www.wilcoxon.com/vi_index.cfm?PD_ID=12


Thanks
Carl

[asmboy]

the voltage mode unit - 2nd link- should be easily DC decoupled - once you choose a MINIMUM frequency of interest.

but as to the PIC - even pics with a `12bit A/D may not give excellent resolution owing to the very wide dynamic range of possible output amplitude.

my experience with a bearing monitor device was that there could be
EASILY a 40db dynamic range to resolve - 60db in the product i delivered.

the PIC A/D is nowhere near able to resolve that very well - and i ended up using an external PARALLEL connected FAST 16 bit A/D to sample for the pic.

BTW: if not a db thinker -in voltage land 40 db is 100:1 60 db is 1000:1 ratio, so with only 10 or 12 bits and a wide frequency range - the PIC a/d surely wont cut the mustard.

ALSO - are you planning to use rectified RMS data over fixed time constant - or will you be doing frequency analysis, attempting FFT etc?