I'm currently recording data points from an analog accelerometer (PCB 353B33) using a TDS3014B oscilloscope which has a record length of 10,000 points at a sample rate of 1.25 GSa/s.

Since the oscilloscope is a triggered data logging device, it's not capturing the whole cycle of the experiment.

I'm trying to look into analog data loggers that would help in seamless data logging for a minute or more with a good sampling rate that would help me capture data for at least one cycle. Is analog data logger the only option remaining?

I'm not sure if an analog data logger would be a good solution for my problem, so if anything else would work for my case please suggest that as well.

  • \$\begingroup\$ This is why data aquisition systems (DAQs) exist. They, however, qualify as lab equipment and therefore have lab equipment prices. Maybe you can borrow or rent one. \$\endgroup\$
    – DKNguyen
    Commented Apr 3, 2021 at 22:12
  • \$\begingroup\$ What sample rate do you need? \$\endgroup\$
    – Mattman944
    Commented Apr 3, 2021 at 22:18
  • \$\begingroup\$ You want to log analog data. Isn't anything that does the job an "analog data logger"? \$\endgroup\$
    – hobbs
    Commented Apr 3, 2021 at 22:27
  • \$\begingroup\$ If price is no object and 200 kS/s is sufficient, I used these for years, they are really nice: tm.astronovainc.com/products/data-acquisition-systems/… \$\endgroup\$
    – Mattman944
    Commented Apr 3, 2021 at 22:36
  • \$\begingroup\$ PCIe data acquisition cards are readily available that can record indefinitely, or at least until your hard drive fills up. If you want to stream data at >1 GSa/s you will need relatively fast (NVMe) storage. \$\endgroup\$ Commented Apr 4, 2021 at 0:22

1 Answer 1


You need a data logger or Digital Acquisition (DAQ) system. I have used National Instruments and AstroNova products. I worked for a large company and price was not the most important consideration.

I would get vendors to help me. Ask the accel manufacture what they recommend. If price is not important, contact National Instruments and AstroNova. Otherwise, google Data Logger and contact some other vendors. You need to balance time and cost. Is it better to just get something that you know will work (high price), or spend a lot of time researching to get a system that has the best cost/performance that you need?

As for the sample rate required, the sensor frequency range (bandwidth) is a starting point. EEs normally specify bandwidth as the frequency at which the output is down 3 decibels (dB). For a voltage, this is -30%.

An accelerometer is a little more complicated because it has a resonance above the bandwidth frequency. It is important to know if you might be exciting the resonance. I can't help you there, you need to look at the mechanical input to your system. For example, a shock test might have frequency components at the resonance.

If your scope has an FFT option, sample at 1 MHz and/or 10 MHz and see what range of frequencies you get. If you have significant frequency content at the resonance, you will need a higher sample rate than if you don't have content at the resonance.

According to Nyquist, you only need to sample at 2X the maximum frequency of interest. But, this assumes that there is no higher frequency content that could alias. I would sample at 10X to 20X the highest frequency that has a significant amount present.

enter image description here



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.