IEPE Accelerometer Compatibility with USB-6353

[runjmoney]

For my Senior Design project we're using an NI USB-6353 since the school lent us one.  Two of our inputs are accelerometers, currently Wilcoxon 786F accelerometers. Are these IEPE sensors compatible with this DAQ? DAQ assist and DAQmx both kept telling me this device doesn't have any channels that support accelerometers so I'm currently confused. My background is in mechanical engr so I don't know too much about IEPE sensor wiring or TEDS. 

Thanks

[santo_13]

Just by itself, USB-6353 does not support IEPE sensors. You need an external amplifier/power injector that can provide the IEPE excitation and output a Voltage that can be interfaced with USB-6353.

In simple terms, IEPE sensors carry power and signal along the same wire, an additional circuit is required to inject power and extract signal from the same wire.

Are you limited to USB-6353? if you are open to purchasing new equipment, there are better instruments. Or you need to get an external IEPE amplifier such as https://buy.wilcoxon.com/it150-vibration-transmitter.html

[Henrik_Volkers]

You don't really need to care about TEDS  unless you want to read the actual calibration values (the sensitivity)

If you have the calibration reports you should find the actual values (around the 100mV/g nominal) 

The signal conditioner and power supply for the sensor could be as simple as as current source and a decoupling capacitor . 

Maybe this unit (s) are in the same drawer where the sensors have been?

To build it by your own you need a 18-30V powersupply , a current diode (or a JFET and a ~330 Ohm resistor (value depends on the JFET ) ) for a 3-10 mA current limiting ) and a 100µF (or 470µF) capacitor with a 36V rating (or 25V if you use two 9V batteries as power supply) . Current diode and cap is needed for every IEPE accelerometer. Schematic can be provided if you don't find it and want to build it.

If you can read the voltage, the conversion can be done in voltage input setup or in your software .. and maybe you want to convert into SI units (m/s²) just in case  you want to do something practical with the results 😉 

Your DAQ is using a multiplexed input (one ADC is shared to multiple channels and switched) and these units are prone to inter channel crosstalk due to settle time errors. If you see 'funny' signals on other channels correlated to prior sampled channels (you will not see if you sample  only one channel), there are  some tricks to minimize it.

Good luck, have fun , and if you run into problems ask here . (Since you need a documentation of your project anyway, it helps if you provide relevant information with your question . like you did with your sensor datasheet. 🙂 )

[runjmoney]

Wilcoxon has similar spec sensors that are 4-20mA sensors, could the daq read these without the external amplifier? 

[santo_13]

---

@runjmoney wrote:

Wilcoxon has similar spec sensors that are 4-20mA sensors, could the daq read these without the external amplifier? 

---

Yes, with some external components (resistors), this 4-20mA signal can be converted to a voltage signal that is understood by 6353 AI channels. As Henrik pointed out, 6353 is a multiplexed ADC, the signal captured will be affected by the signal on other AI channels.

The best choice of simplest instrument for an Accelerometer is one of these bundles - https://www.ni.com/en-us/shop/category/compactdaq-chassis.html?productId=412268

[Henrik_Volkers]

If you already have the sensors of your first post , here is the simplest (but rude) way to use them:

3x 9V batterie provides 27V (bit more on fresh ones)

R1 = 3300 Ohm is the current limiter (current is always below 10mA , a value between 3000 and 3500 should do)

C1 = 470µF decouple the 12V bias , value could be lower but I wouldn't go lower than 10µF.

R2 = 1Meg Ohm is for the BIAS currents of the DAQ and (after some seconds) bring the cap down to ~0V (will never be exactly zero due to the bias current of the DAQ and leakage of C1)

the more elegant way would be to use a constant current limiter instead of R1 ( aka a ~4mA current diode )

[runjmoney]

I appreciate that, only problem is I don't have enough confidence in the EE on our project to make that.  I found a somewhat cheap signal conditioner that looks like it would work to me, do you agree? Here is the link to it:

https://www.picotech.com/accessories/bnc-cables-and-adaptors/iepe-signal-conditioner

Thanks.

[santo_13]

---

@runjmoney wrote:

I appreciate that, only problem is I don't have enough confidence in the EE on our project to make that.  I found a somewhat cheap signal conditioner that looks like it would work to me, do you agree? Here is the link to it:

https://www.picotech.com/accessories/bnc-cables-and-adaptors/iepe-signal-conditioner

Thanks.

---

Yeah, this looks to be a good generic IEPE conditioner.