Basic (I think) signal conversion question


Thread Starter


I am shopping for a linear position sensor with 112" travel and needs to be accurate and repeatable to 1/32" at very slow speeds. I've shopped for drawstring potentiometers: not quite accurate enough. I've also looked at magneto-restrictive (piston or magnet-rail) types and they seem to be overkill as far as precision and price. On the middle end, I'm now looking at magnetic coded strips.

Here's the problem. I know nothing about signal conversion and/or controls except this. I want a 4-20mA signal for the entire range of 112"(not counted like an encoder). I can only get the magnet strips with rs422 output (I gather this is essentially digital/square wave?)

Note: this is the sensor I'm looking at

Q: Can I get a converter for rs422 to 4-20mA output? What resolution loss can I expect? Ballpark price? Remember, I'm very new to this so please keep jargon or spec info to a minimum or explain.


Bob Peterson

there are encoders that have an Ethernet output.

I don't know of any rs422 to 4-20mA converters. It would be sort of like trying to convert an apple into a cow. It just would not make much sense.

RS422 is usually some kind of ASCII signal. You would read it directly from a serial port using an RS422 to RS232 converter. You would likely get an ASCII string back that has a sequence of characters that would tell you what you want to know directly.

It is hard to give you good advice about this kind of thing without knowing more about your application and what you are trying to control and what you are using as the controller.

You can get a small PLC that an encoder can feed directly for a few hundred dollars.


Curt Wuollet

You might keep looking, that range is often used in machine tools and there are some low cost solutions. The conversion shouldn't be much of a problem. Without knowing the protocol it's hard to be too specific but for example you can get a PLC that has the ability to receive RS422 traffic and has an analog output. Get a reading, scale, send to the analog out. For encoded solutions the same PLC could read a quadrature or gray code input. Not absolute beginners stuff but not beyond a careful study of the manuals. I'm thinking of a system that uses an optical encoder and a tape scale of any length but I can't remember the name. Accuracy wasn't good enough for machine tools but well within your needs.


the first question is (according to this last mail): what is the protocol used by RS422?

The most popular in industry is modbus (with different media as rs232, rs422, rs485, ethernet etc.) but we find profibus, etc.
You can find example of solution here:

If this protocol is not a standard one (a proprietary protocol), I agree with previous comments. it will be difficult to find a solution without a small PLC with serial interface and analog output.
And this is an other way...


Curt Wuollet

Even though I am big on DIY and clean solutions, my suggestion was based on the fact that it's easy to spend a lot more than for a $100 PLC and it's probably the easiest way. Personally, I might use a cheap microcontroller simply because it's much easier and less clumsy to fake protocols in C. But if a low end PLC will do what you need, program it and forget it.

> I can only get the magnet strips with rs422 output (I gather this is essentially digital/square wave?)

I think in this case, the encoder is using RS-422 line drivers to output a quadrature waveform. It is probably not encoded as ASCII or any other protocol. If a voltage signal is an option, this converter might work:

This might be a case where trying to save money will end up costing more, however. I would suggest buying a linear position sensor with analog outputs. This way, you only have one phone call to make if you need assistance.
Hi all,

What you wanna do is converting incremental encoder signal into absolute position. Answer is simple - there's no "converter" for this. That is because your device outputs two square pulse trains based on the displacement. You can tell from this the displacement and the movement direction but not the position. So if you don't know position, you can't generate any signal based on it. End of story. It seems the device is not a device you're looking for.

However there are methods to obtain absolute position based on incremental signal. But for this you must have posiblilty to move the head of the sensor to well defined position after each power on anda device which can count the pulses in order to generate the signal you want. The cheapest solution off the shelf I know of is simple PLC with high speed counter feature. But this is still few hundred $. Also, as someone mention this already, you can built some hardware for this task based on any microcontroller. But this is no longer motion control topic I think.

Jacek Dobrowolski