MEMS accelerometer measurement system

[pbarrett]

Howdy,

TLDR: Can a 9205, 9202, or similar DAQ be successful and useful at measuring DC MEMS accelerometers? Why/why not? What is the recommended set up?

Scenario: need an accelerometer and DAQ system to measure +/- 2g at 0-~600 Hz, the 0Hz being the important spec. NI has out of the box piezoelectric accelerometer DAQ solutions, but MEMs doesn't appear to have an NI turn key. I am fine adding in my own power supply excitation so the module doesn't need to provide that.

So we're left with what DAQ to choose, since DC MEMs and external excitation rules out IEPE.

Accelerometers

Both are single ended

1. 4030-002-120 (te.com)
   1. 1000mV/g sensitivity --> +/- 2V swing (nominal)
   2. 2.5V DC offset bias
2. 7533A1-XX_ds.pdf (dytran.com)
   1. 420mV/g sensitivity --> +/- 840mV swing (nominal, 950mV published)
   2. ~1.5V DC offset bias

Sensor	Type	Range (g)	sensitivity (V/g)	exc. current	exc voltage	Range FS
7533A1	DC Mems	2	0.42	350 uA	4 to 28	0.84
4030	DC Mems	2	1	4 mA	5 to 30	2

DAQ

Back in the day, you'd just go with a DSA/sound and vibration DAQ. But is that valid for DC MEMS since these DAQ modules favor IEPE and simultaneous sampling?

1. How should one use the DC offset bias in their DAQ selection?
   1. For example, a 9205 set to +/- 1V range would offer good resolution for the 4030, but the DC offset is outside this range.
   2. Do DSA modules null this out in hardware? In the DAQmx driver?
   3. Am I going to violate common mode or isolation requirements of non DSA modules?

I'm inclined to go with a 9202 w/ external excitation. Why? Channel count, cost, and resolution. Will this work?

DAQ model	Cost	Sampling	Channels	Sample Rate	Low	High	bits	range	V/bit	7533A1	4030	Cost/channel	Note
9234	$ 2,553.00	Simultaneous	4	51.2k	-5	5	24	10	596.0E-9	1.4E-6	596.0E-9	$        638.25
9239	$ 1,538.00	Simultaneous	4	50k	-10	10	24	20	1.2E-6	2.8E-6	1.2E-6	$        384.50
9202	$ 1,795.00	Multiplexed	16	10k	-10	10	24	20	1.2E-6	2.8E-6	1.2E-6	$        112.19	Top contender
9230	$    755.00	Simultaneous	3	12.8k	-30	30	24	60	3.6E-6	8.5E-6	3.6E-6	$        251.67
9229	$ 2,048.00	Simultaneous	4	50k	-60	60	24	120	7.2E-6	17.0E-6	7.2E-6	$        512.00
9205	$ 1,152.00	Multiplexed	32	250k	-1	1	16	2	30.5E-6	72.7E-6	N/A	$          36.00	Outside the DC offset bias
9215	$    772.00	Simultaneous	4	100k	-10	10	16	20	305.2E-6	726.6E-6	305.2E-6	$        193.00
9205	$ 1,152.00	Multiplexed	32	250k	-10	10	16	20	305.2E-6	726.6E-6	305.2E-6	$          36.00

I'll loop back to this once I get something working since I've seen a bunch of other forum posts that ask this but haven't closed loop.

[santo_13]

First of all, kudos for providing a good level of detail and sharing your exploration so far. Such detailed posts are hard to find in the forum.

Based on your requirements (0 Hz & external excitation), I would also agree that 9202 seems prospective. I will try to spend some time to check for any argument why not 9202 for your application.

The natural selection of DSA cards is for their excellent frequency retention and large dynamic range. 9202 provides a similarly wide dynamic range but is limited in input voltage ranges.

[Henrik_Volkers]

My comments:

If you look at the MEMS noise spec all bits beyond 13 (or less) are noise (consider full range) so 16 bit is OK

The output impedance of the MEMS is not specified , if you don't need simultanious sampling the multiplexed DAQ channels look cheap, but have a look at settletime error in the spec (prone to channel crosstalk). I always try to avoid muxed inputs.

The DC offset is sligthly different between units.. if DC is of concern .. so you will need an individual calibration (always recommended and easy done with a good estination of g_local) 

If you need a connector / supply board anyway, you can add one OP per channel that remove the offset and add gain if needed (see noise 😉 )  , again a calibration in earth gravity for each channel ...

@santo_13:  If you offer kudos .. there is a button to really provide them 😄

[pbarrett]

Thanks Henrik for the thoughts!

We've selected the 7533AX accelerometers family which have an output impedance of 100Ohm. Sampling rate would be anywhere from 1.2kHz to 6kHz. I will look out for this when I begin experiments.

I'm not certain adding in additional circuitry to the front end is necessary for this application. While maximizing the dynamic range is always nice, upon discussion with our engineer partners, we're well within the resolution of interest.