Safety Isolation - Potentiometer Input

[MMussio]

I'm designing a motor controller for a client which uses a bunch of inputs such as digital inputs and a potentiometer. It has to be safety isolated. For the digital signals, no worries. For the potentiometer I thought of the HCR200 family but it is a little bit expensive for the estimated BOM.  I was just wondering if for a potentiometer i really need a high linearity coupler just for safety reasons or if there is another possible cheaper solutions.
The Board will receive 24V from a external power supply which will also supply voltage regulators, but that I don't really think is a issue if the user interface is isolated.
Thanks in advance.

[BrokenYugo]

Mount a regular pot back from the panel and connect with an insulated coupler and/or shaft?

[Gyro]

You can buy potentiometers with plastic shafts and mounting bushes, together with push-on plastic knobs, specifically for this application. They are normally not significantly different in price from the metal bush/shaft ones.

[PCB.Wiz]

I'm designing a motor controller for a client which uses a bunch of inputs such as digital inputs and a potentiometer. It has to be safety isolated. For the digital signals, no worries. For the potentiometer I thought of the HCR200 family but it is a little bit expensive for the estimated BOM.

Did you mean the Analog optocouplers HCNR200 ?  Those are well stocked, at not bad prices at lcsc  ~ 98c/100 ?
An alternative could be the ISOM81xx series, from TI, if you do not need high precision.
Or, if you are ok with calibration yourself, you could 'roll your own' using dual or quad opto couplers.

[Zero999]

Why not use a digital isolator? Cheaper and faster.

How is the potentiometer being read?

I wouldn't bother with analogue isolation.

If it's being read by a microcontroller, use another microcontroller on the other side of the isolator to convert it to a digital signal, such as SPI.

Another option is use PWM, then filter it or again use an MCU.

Here's a simple 555 PWM circuit.

[coppercone2]

I think you should use some kind of plastic coupler, alot of my hv equipment has modded pots that look like it would be cheap to remake but very safe

[Zero999]

We don't know where the potentiometer is, or if the original poster even has control of it. The customer might want to connect their own potentiometer for example.

[TimNJ]

If the analog potentiometer would get digitized and read by an MCU anyway, then a digital pot sounds fine, especially if you already have an isolated Vcc available for the digital inputs.

Still, a mechanically isolated pot may still be cheaper and simpler. I went through a similar exercise recently, for an analog control circuit. I thought about linear optocouplers and digital pots, but ultimately decided to just put a 10mm plastic extension shaft on the trim pot.

[Jeroen3]

Can you re-use the shunt analog isolators? Probably an AMC1200 or something similar.

[MMussio]

As we send the board for the user to install in a electrical panel of one of his machines, I don't have any control of the type of potentiometer. The safety concerns are in the board, not in periphericals.

The machine is sold to final users, which will eventually have to change components and have to work in the board. The potentiometer and digital inputs are conected via terminal blocks. The idea is: the final user has no risk of electrical shock.

[MMussio]

The problem is not in the potentiometer itself. The board is open in the electrical panel and eventually the final user has access. I want to make sure in the user interface area (inputs) he has no risk of electrical shock. (There is reports of other similars boards doing that).

[jbb]

Well that is classic motor control. As others have said before; if the pot does some analog control, consider analog isolation. If the pot just goes into a microcontroller, consider putting some kind of ADC right next to the pot and isolating the digital signals.

If I recall correctly, it’s quite common to see a 0-10V input for speed control signalling on commodity Variable Speed Drives (VSDs).

Something to consider; when designing your control interface, have a think about what should happen if a wire breaks off. Thus you probably want a gentle pull down on your speed control signal.

[Terry Bites]

Yep, that's the best way.

[Zero999]

As we send the board for the user to install in a electrical panel of one of his machines, I don't have any control of the type of potentiometer. The safety concerns are in the board, not in periphericals.

The machine is sold to final users, which will eventually have to change components and have to work in the board. The potentiometer and digital inputs are conected via terminal blocks. The idea is: the final user has no risk of electrical shock.

The problem is not in the potentiometer itself. The board is open in the electrical panel and eventually the final user has access. I want to make sure in the user interface area (inputs) he has no risk of electrical shock. (There is reports of other similars boards doing that).

You've not said whether it needs to output an analogue voltage or if digital will do?

And how many channels are required?

If digital is okay, then as I've said, along with many others, use a microcontroller with an ADC and transmit the signal via an opto-coupler or digital isolator.

If you need analogue, then the HCNR200 you mentioned at the start of the thread will do. There's also the AMC1350.
https://www.ti.com/lit/ds/symlink/amc1350.pdf

PWM can also be converted back to analogue with a low pass filter. The circuit I posted is better if you know the value of the potentiometer, but will work with other values, just with a different output frequency. There are other ways to generate PWM, which output a constant frequency, irrespective of the potentiometer value, or you could just use a microcontroller to generate the PWM.from a potentiometer connected to an ADC input. The PWM can be a low frequency, tens to hundreds of Hz and the low pass filter can have a cut-off of a couple of Hz.