2
\$\begingroup\$

I have a bunch of devices that will sit on a two-wire bidirectional RS485 bus (DMX-like, but not quite DMX).

I'm trying to protect against an edge case, where a user mis-wires a ballast, and inadvertently connects mains voltage to the bus, causing the it to become live and potentially injuring anyone else working on it.

I will be installing an 'isolating box' between the bus and the user, essentially behind a wallplate, such that the bus itself can never be directly accessed. The connection to the bus itself would be through opto-isolators and associated circuitry (with a separate isolated supply), however I'd like to 'protect the protection circuitry' if possible, such that it won't catch fire and can gracefully recover.

Would a couple of PTCs, and a couple of TVSs (for redundancy) as below, be enough to reliably save any low voltage circuitry to the left of the circuit, when IO1 and IO2 are connected directly to 240V mains?

rough sketch

Normally, IO1 and IO2 would see ~22v at roughly 2mA.

Thank you kindly

\$\endgroup\$
4
  • 3
    \$\begingroup\$ You might want to start a step earlier and prevent miswiring on a physical level e.g. by using properly keyed connectors. \$\endgroup\$ – PlasmaHH Jan 30 '18 at 12:13
  • \$\begingroup\$ Unfortunately I'm not in control of this, as this is a manufacturing situation with bare wires and screw terminals (essentially a test bench) - And electricians. \$\endgroup\$ – BaudThief Jan 30 '18 at 12:15
  • \$\begingroup\$ can you redesign the test bench so that the wrong wires won't reach the wrong terminals? \$\endgroup\$ – Jasen Jan 30 '18 at 19:30
  • \$\begingroup\$ Sure, but there are only two terminals on the bench. The device under test has four screw terminals in close proximity and is attached to mains by a cord, which is where the problem is \$\endgroup\$ – BaudThief Jan 31 '18 at 2:27
2
\$\begingroup\$

In the real world you can never make things totally idiot proof.

If these things are designed to be installed and then left alone, it is foolhardy to add more components than are necessary to protect it from normal operational faults.

If it is meant to be routinely moved around, and or reconnected, then I'd say it maybe more prudent to invest in more protection. Use of TVS though is intended for transients. They will not protect your circuit from being connected to a multi-megawatt power supply. If I really had to do this I would be looking at some form of crow-bar circuit, perhaps using a triac, to detect over voltage, short the line to mains ground, and blow a fuse, possibly a resettable one.

Instead of spending development money on making it bullet-proof and reducing the reliability of the unit as a whole, invest time in preparing clear and concise installation manuals and trust that the electrician can read / do his job properly. If you are really paranoid, add paper tags to the low voltage connections indicating "NO MAINS HERE" or "LOW VOLTAGE CIRCUIT".

If you are getting lots of field returns then analysing what can be done to prevent that is prudent. However, fixing the connectors and improving the documentation is the better approach. Preventing the problem ALWAYS trumps protecting from it.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ I was asked to look into the feasibility of this by a client. Essentially they're testing lighting ballasts prior to shipment, and there's a chance the testers will mis-wire them during tests on the bench. Accidents happen, even out in the field by qualified installers (lots of blown up ballasts due to mis-wiring), I'm just looking for a way to minimise damage/exposure to faults in the lab. Given the feedback that you and others have provided (which I honestly really appreciate), I don't think this is something I'd want to pursue. Thank you for taking the time to comment. \$\endgroup\$ – BaudThief Jan 30 '18 at 14:21
  • \$\begingroup\$ @BaudThief if you are getting lots of field returns then analysing what can be done to prevent that is prudent. However, fixing the connectors and improving the documentation is the better approach. Preventing the problem trumps protecting from it. \$\endgroup\$ – Trevor_G Jan 30 '18 at 14:24
  • \$\begingroup\$ @BaudThief BTW: if I really had to do this I would be looking at some form of crow-bar circuit, perhaps using a triac, to detect over voltage, short the line, and blow a fuse, possibly a resettable one. \$\endgroup\$ – Trevor_G Jan 30 '18 at 14:43
  • 1
    \$\begingroup\$ Cheers for that. I looked into what you have suggested, and found SIDAC Thyristors (Diodes Inc) are pretty much purpose-built for this kind of situation, when used with a fuse. \$\endgroup\$ – BaudThief Jan 30 '18 at 15:07
1
\$\begingroup\$

I'm trying to protect against an edge case, where a user mis-wires a ballast, and inadvertently connects mains voltage to the bus, causing the it to become live and potentially injuring anyone else working on it

Use a GCFI or residual current device so that the AC becomes disconnected within a few milliseconds of an inappropriate connection. This will save lives but won't necessarily save components on the bus.

To save bus components the GFCI will need supplementing with some circuit that can handle "the overload" for up to 30 milli seconds before the GFCI kicks in.

You may get away with PTCs but without knowing which one you intend to use I cannot be sure and I'm not going searching for one that will.

\$\endgroup\$
10
  • \$\begingroup\$ GFCI is a good idea I think under the circumstances you probably need to sacrifice the bus components to the great god of the magic smoke. \$\endgroup\$ – RoyC Jan 30 '18 at 12:59
  • \$\begingroup\$ I really appreciate the answer, but each bench is already equipped with an RCD - Perhaps I should have left out the human safety aspect, and asked how I would go about protecting only the attached circuitry. This is more of a secondary measure in case existing measures fail \$\endgroup\$ – BaudThief Jan 30 '18 at 13:02
  • \$\begingroup\$ You should always consider questions thoughtfully. I can only read what you write and because safety trumps reliability this was what I concentrated on. Apart from anything else, you have given no detail of the protection device so how can anyone really help you? \$\endgroup\$ – Andy aka Jan 30 '18 at 13:20
  • \$\begingroup\$ Thank you for your feedback, unfortunately communication doesn't appear to be my strong point. I didn't want to overload the question with detail that wouldn't add any value and cause confusion. I was hoping for some general direction as to whether the TVS's and PTCs themselves would be sufficient protection /in general/ for low voltage circuitry on the left, or if there are alternatives/additions that I should consider (ie: crowbar and a fuse). In any case, thank you. \$\endgroup\$ – BaudThief Jan 30 '18 at 13:48
  • \$\begingroup\$ This is a question and answer site. You ask the question and the best questions are specific and link to data sheets etc. and have diagrams. If the detail in the diagrams looks wrong you will get help and alternative ideas but without the detail nobody can answer specifically on the parts you have drawn. \$\endgroup\$ – Andy aka Jan 30 '18 at 13:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.