I’m hoping to get some assistance in piecing together a kit for a special use case.
I need a lightweight vibrational sensor ( < 5g ideally ), that will record from 400 hz - 12 khz vibrational data (single axis should be fine) and then a way to get that data into a computer using usb (ideally with the computer seeing the interface as an audio interface).
My intentional is to analyze the data using a ML algo for additional processing in real-time.
Is there a kit that will meet (or come close to meeting) these specs?
Welcome to the Technical Forum. The 5g is low. Though I did find part number 3461-SVT200-A-ND that would do the range. 5g would be included. Here is the link:
In the data sheet you would have to use this part with this gateway 3461-GU200-ND. Here is the link:
This one does list the USB interface. I just did not see anything else for that 5g . You can review this to see if it will work.
I suspect that the “5g” reference in the original post was a reference to mass rather than acceleration, and that the limiting acceleration values have not been defined. Getting an estimate of such would be quite useful in picking a device with appropriately-scaled sensitivity.
In any event, there are many accelerometers available that are quite light, many of which are available on evaluation boards that can simplify connection. The upper frequency limit requested is quite high, and that will limit the options significantly.
I’d offer the suggestion that it may be convenient to use an analog-output device (ADXL1002 for example) and capture the signal using a mic/line input, either built-in or external/USB connected. But again, understanding the required acceleration range would be important in order to pick a device and scale the signals appropriately.
You’re correct in 5g is in reference to mass (lighter is better).
This sensor will be attached to the underside of a drum cymbal, so it will need to be able to withstand a decent impact, although it won’t be a direct impact. Preference would be for the sensor to be screwed on to the cymbal.
So it’s possible to take the output directly from something like an ADXL1002 into a microphone input on a recording interface? I assumed the sensors required a special power source or something. If this is the case, is there an adapter designed for this purpose → XLR or 1/4 TRS ?
Why not? Analog signals are analog signals… As mentioned one has to pay attention to scaling things properly, and yes a power supply is required. The output of a device like the ADXL1002 is approximately half the supply voltage at zero acceleration, swinging to zero and the supply when positive and negative full-scale acceleration are applied. Standard cords aren’t available since it’s an atypical use case, but it shouldn’t be rocket surgery to put something together.
A person might also give something like a piezo bender a try if one isn’t so much interested in accurate measurements as looking at frequency spectra.
Thanks for the input. I’ll take a look at Piezo Benders as an option as well.
Regarding the ADXL1002, is there a product that Analog or somebody else sells, that will provide power and amplification appropriate to take the signal into a mic input?
Am very new to these type of devices, and have limited knowledge of the components involved in the chain from source to output. If you have any pointers to an educational resource that would provide a foundation on these types of sensors, I’d love to take a look.
As an off-the-shelf product, not that I’m aware of. It’s not the typical use case for such products…
Re. power, any stable source within the rated range will work. (3) triple A’s or a small LiPo cell should do, else a USB source could probably work also.
If anything, you’ll probably need some signal attenuation, as the output can range from zero to the supply voltage (about 3-5V) while mic signals are typically at mV scale. A typical line input will accept something on the 100mV-1V scale, and would require less attenuation. The most typical way to achieve this would be a “voltage divider” for which explanations can be found at wikipedia, khan academy, johnny wannabe’s electrawnik skewl, and everything in between.
A piezo transducer would be a self-powered source, and I’d not be surprised if it would be content driving a mic input directly. That’s a very different thing however, as it’d be detecting flexure of the surface to which it’s attached, rather than actual acceleration. Might work, might not–depends on what precisely it is you desire to measure.