2
\$\begingroup\$

I need help to design a circuit that will detect if one of several GND connections break.

I am working on a PCB for an industrial application where I am designing an add-on module for an external system.

The external system exposes both 24V and 48V, along with their respective grounds, which are neither common nor fully isolated from one another.

The add-on module does not require the 48V supply, however, it is crucial that there is common ground with both the 48V and 24V system. Until now, I have shorted the two ground interfaces in the module I am designing.

I have illustrated the scenario below, including the non-isolated nature of the two ground potentials, indicated by the impedance drawn between them. There are additional interfaces between the external system and the add-on, but I have omitted these for simplicity.

enter image description here

I have on several occasions experienced that either ground connection may fail, for instance by the wire breaking in the connector between my module and the external system.

As GND24 and GND48 are not isolated, the system may proceed to operate with only one of the two GND lines, but it may cause erratic behavior of the interaction between my module and the external system.

This must be avoided, either by detecting the issue, or shutting down the add-on module altogether, if either GND line breaks.

In essence, I need help to design a “GND health” monitor as illustrated below, that may either disconnect the add-on module, if either GND line breaks, or alternatively, report the issue to the internal logic such that an alternative action may be taken.

enter image description here

Any help would be greatly appreciated!

\$\endgroup\$
3
  • \$\begingroup\$ Why can't you just detect "erratic behaviour"? \$\endgroup\$ Commented Oct 6 at 12:06
  • \$\begingroup\$ Okay... this smells like an XY problem. A side effect of missing ground is that 48V probably wont be 48V anymore, can't you detect that? \$\endgroup\$ Commented Oct 6 at 12:21
  • \$\begingroup\$ Why is it important to reference 48V ground? Does your add-on send or receive signals from systems powered by the 48V supply? At minimum, do not short the two grounds in your design. That could have unintended consequences. \$\endgroup\$ Commented Oct 6 at 15:50

1 Answer 1

Reset to default
1
\$\begingroup\$

I would agree with previous comments that this seems a little cart before the horse. If I were in this position I'd be inclined to focus my efforts on avoiding the poor connection, and avoiding erratic behaviour in the event of a poor connection.

The former might involve alternative connectors or cables. The latter could involve adding a checksum to serial communications to detect data corruption, or a voltage monitor to the power supply to detect fluctuations.

If those options are out of your control, I suspect detecting the ground loss would still just be covering over an existing design weakness.

If the 48V supply is totally unused there should be no need to connect the 48V ground.

It seems likely what you have is some signals between external and add on which are 24V ground referenced and some which are 48V ground referenced, and you are handling this by forcing the two grounds together within the add-on.

If the two supplies are intended to be common ground and this is just to deal with some noise or voltage drop in the interconnect then that may be enough.

If the supplies are not intended to be common then that could be a risky move, and your add on could end up carrying a lot of current unintentionally or breaking a deliberate safety isolation. In this case, best to keep the grounds seperate with an isolated connection between the two halves for any signals required.

If you really, really dont want to change anything else, then I'd suggest measuring the supply voltages as a proxy for ground.

Measure the 24V and 48V supplies with respect to 24V ground.

If the 48V ground is lost the 48V supply will drift by the difference in grounds. If the 24V ground is lost then the 24v supply will appear to drift by the difference in grounds.

This wont detect anything if the grounds are externally well connected, though it sounds as if in that case you wont have a problem.

A more robust solution is likely to require measuring ground currents, or flow between grounds. This will involve adding some sense resistance, which may cause problems for your add-on.

\$\endgroup\$

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.