Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Can you build a mechanical memristor?
2 points by DavidHaerer 34 days ago | hide | past | favorite | 1 comment
I'm trying to understand the memristor and have an idea for how to build a mechanical system, that might exhibit the characteristics of a memristor. However, ChatGPT keeps telling me it wouldn't work, but I'm not convinced. As you might already conclude, I only have a basic understanding of electrical engineering, so I hope this question is not too stupid.

I've started with this [1] common diagram showing the relations of the four fundamental quantities.

Charge q: Experiences positive / negative force in an electric field.

Current i: Amount of charge passing through a surface.

Voltage v: Potential energy of charge.

Flux Linkage Φ: Amount of magnetic field passing through a surface.

dq = i dt: Definition of current as charge per time.

dv = R di: Ohms law, voltage drops as current flows through a resistor.

dq = C dv: Charge on a capacitor changes with a voltage change according to the capacitance.

dΦ = L di: An inductor creates a magnetic field when current flows through it according to the inductance.

dΦ = v dt: Faradays law of induction, change of the magnetic flux linkage in an inductor induces a voltage change across the inductor.

dΦ = M dq: Formula of the memristor. The device must change resistance M according to the amount of charge that has passed through it.

According to ChatGPT, the dΦ in the memristor is an abstract state variable and does not imply there is a magnetic field involved. But lets get to the mechanical system.

Imagine a coil on one side, and a rotatable permanent magnet on the other side. When charge flows through the coil, it creates a magnetic field, causing the permanent magnet to rotate to align field lines. The changing magnetic field of the permanent magnet causes a voltage change in the coil. Hence the measured resistance of the coil changes, when current flows through it. The formula dΦ = M dq seems very intuitive to me, now that it involves an real magnetic field that changes.

What do you think of the described mechanical system. Can it behave like a memristor? According to my understanding, the way to find out would be to build it and make measurements of voltage and current. Plotting it on the v-i plane must show a pinched hysteresis loop.

Also, isn't this very similar to how magnetic storage works using the magnetoresistance effect?

[1]: https://en.wikipedia.org/wiki/Memristor?useskin=vector#/media/File:Two-terminal_non-linear_circuit_elements.svg




description of one way to build a mechanical memristor : http://dr.ntu.edu.sg/bitstream/10356/156385/2/Mechanically%2...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: