encoder : shielded cable termination

Depends on the circumstances....the vender for the equipment should provide the proper wiring diagram recommendations...If you are unavoidably close to power lines or your equipment is very sensitive to high frequencies, more creative solutions may be necessary...but I would try it grounded on both ends first then if there is interference, try one end grounded from the power source end...If that doesn't work then try some suggestions from the link provided or the manufacturer of the equipment you are using...

https://instrumentationtools.com/where-should-we-terminate-the-shield-and-why/

Sometimes the vendor or manufacturer makes a mistake in their instruction manuals.

In the company I used to work in, we had a spark detector system. Sensors were layed out in a chamber to detect flames/sparks. The manual said, "Do no connect the shield of the detectors to the panel."

I noted that the detectors were isolated from ground so not connecting the shields would have rendered them ineffectual. I told my supervisor and manager at the time but they said, "follow the manual." So I did. What followed was years of intermittent false alarms and investigations and solutions that never worked. I was a lowly technician at the time.

Years later, I had a new manager and I had become a supervisor. I told him to let me try wiring in the shields but he told me to consult the manufacturer first. Their reply was a nonchalant, "go ahead and try." So, I did.

Oh, but the false alarms disappeared to the point where we wondered if the system was working. Tests confirmed the system was functioning, however.

Following the instruction manual, schematic diagrams usually work but knowing your stuff is equally important.

regards,

Vulcan

Are the connections labeled GND protective earth or power supply return? Or perhaps both where the power supply return is tied to protective earth.

In my experience, Diagram A generally works better where the noise sources of concern are in the 10kHz range and up, e.g switching power supplies, motor inverters, etc.

It seems the old 60 Hz ground loop issues (where single-ended shield terminations are used) are not as prevalent as they might have been in the old NTSC video days. Anybody remember seeing 60 Hz hum bars?

A ferrite core with a couple of wraps around it just before the encoder receiver works far more wonders than shielding ever thought off, check your signal out with a ‘scope, before & after. Differential transmitters & receivers also are a far better technology for getting a good usable clean encoder signal than what you have displayed. This if you are in an industrial environment. In a lab, you won’t be able to tell the difference either way.

As far as I can see from the 2 diagrams the GRND terminal is not a ground but the negative for the 5V rail and the signal rail. Therefore to correctly screen the encoder the case of the encoder, assuming it is metal or has a ground terminal should be connected to the screen and the other end connected to the power supply earth not the negative rail.

The encoder supply should not be connected to the screen as all manner of interference will be injected. Having worked in mining, oil drilling, power stations and refineries I can say that connecting a shield to the negative supply was a total no no.

As others have said

Obviously you only need the ground connection at the encoder end if the encoder has an identifiable ground.

That sort of a connection has been accused of creating interference in the signal due to circulating currents due to ground potential differences. That is the reason the shield for instruments i.e. 4-20mA, HART and even 0-10Vdc is grounded only in one place, by convention, the side that supplies the power for the signal.

You're right: I was assuming that the encoder was isolated; I should have said:-

Obviously you only need the ground connection at the encoder end if the encoder has an identifiable ground connection, but is not grounded anywhere else.

If one of them is floating (not grounded or is insulated from ground), A will work.
If both of them are grounded, B is the only one that will work.

What you want to avoid is a ground loop, which is what you will have if you use A but with both instruments grounded. You will get noise galore, particularly a 60Hz signal riding on your signal. That will sometimes give you strange things happening to your controls.

Good practice, however, it to use B.

If you use A and your encoder suddenly gets grounded (either accidentally or incidentally), you'll have a ground loop and all the problems a spurious signal will create.

Follow #8⇑.

There are many ways to connect a shield, and there are even driven shields in very sensitive instruments.

I have seen this used in Magnetic Flow Meters.

Example:https://en.wikipedia.org/wiki/Driven_guard

A simple method to reduce noise is to put a few twists in the signal wires. This will reduce noise by a couple of DB in some cases. Maintaining the twist rate is very important especially in Cat6 cables.