I'm a navigation engineer, so I often have to deal with gyroscopes and their measurements. Out of curiosity, I've decided to build one myself and will give it a try with a Fiber Optics Gyroscope (FOG), which is a Sagnac Interferometer that consists of a laser, optical fiber coil, a beam splitter and a detector.

The basic idea is that the beam from the laser is split into two parts that enter the coil from the opposite sides. They travel through the coil until they meet and interfere in the splitter and then reach the detector. Since the interference ranges from destructive (zero output) to constructive (double the amplitude of a half beam), a large dynamic range detector is needed. Also, being able to sample it at least 10Hz (preferably even more) is desired.

I've looked into sensors like:


which could be an easy option, since it does amplification and AD conversion internally. Just its max frequency is 10Hz, which is on the lower side of what I want. Also the sensor itself is very small, thus the alignment between the laser and the sensor could be problematic. Some laboratory sensors like this have much larger sensing diameters (5-7 mm), thus would like to find something bigger as well. The laser I am using has 635nm wavelength, thus silicon based sensors is not a problem.

So, I am seeking for an advice on what kind of sensor suits the most here for this type of application taking into account the high dynamic range, sensitivity and preferably linearity (although this will probably depend on the measurement circuit).

Thanks for help! (and sorry if smth is wrong - this is my first post :))

  • \$\begingroup\$ One thing you could try is to use a LED as a photo-sensor. When a LED is hit with light near it's emission frequency, it generates an analog voltage signal. It has a bonus feature that It is only responsive to the frequencies it emits. There's not enough evidence for this to be a full answer. \$\endgroup\$
    – tuskiomi
    Commented Mar 25, 2020 at 19:09
  • \$\begingroup\$ I wish I had an FOG just lying around. o.0 \$\endgroup\$
    – DKNguyen
    Commented Mar 25, 2020 at 19:25

1 Answer 1


You will need a photodiode as the detector element.

You will also need a transimpedance amplifier circuit is to use it. 10Hz sample rate is nothing for optics. But sample rate is perhaps not as important as bandwidth.

But 1MHz is easy to achieve: 1 MHz, Single-Supply, Photodiode Amplifier Reference Design

If you end up needing a logarithmic output then you need more circuitry.

You say sample, but that is not the same as bandwidth. It is unclear what you actually mean. Are you going to be displaying this signal on an oscilloscope or what?

Do note that laser diodes produce coherent BUT uncollimated light. I don't know enough about fiber optic gyros to know what type of light you need to inject but do not expect a parallel beam (or even a circular beam unless it is a VCSEL laser diode) from a laser diode without optics.

  • \$\begingroup\$ 1. There many different designs of FOGs, some of them they modulate the phase of the light, etc., but in my case I just want to measure the strength of the light, which should be constant at constant rotation rates. If everything is perfect, this would only depend on the rotation rate, thus the bandwidth, the way I understand it, does not play a role. 2. I have a laser which output is already with optical fiber FC/APC connector, so I assume that the injection is done correctly there, although not sure. \$\endgroup\$
    – bronius
    Commented Mar 25, 2020 at 21:07
  • \$\begingroup\$ @bronius Bandwidth always plays a role. Because you can sample a sensor's output at 1000Hz, but if it has zero bandwidth, the waveform you get from those samples would just be a DC signal that has existed since the beginning of time, so I am not sure what you mean when you say bandwidth does not play a role. Bandwidth is reflective of how quickly a signal can change (and how it can be detected without lagging too far behind said change). So I am not sure what you mean when you say bandwidth is not an issue. \$\endgroup\$
    – DKNguyen
    Commented Mar 25, 2020 at 21:12
  • \$\begingroup\$ @bronius Are you referring to spectral bandwidth? I am referring to temporal bandwidth (i.e. the beat frequency of your interference pattern). \$\endgroup\$
    – DKNguyen
    Commented Mar 25, 2020 at 21:15
  • \$\begingroup\$ Correct me if I am wrong, or misunderstood you, but when you have two sine waves of the same frequency but shifted in phase relative to each other, the frequency component does not change, it only the amplitude that becomes a function of the phase shift, so I do not see here a beat frequency, and indeed, what I want to measure is the amplitude, which for a constant phase shift (constant angular rate), will stay constant. \$\endgroup\$
    – bronius
    Commented Mar 26, 2020 at 15:26

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.