Converter: 24V from PLC to TTL devices

Use a resistor, a 4.7V zener and a 100nf capacitor (capacitor parallel to the zener). Schmidt trigger inputs are advisable (74hc14).

On a sunny day (Fri, 25 Feb 2011 03:56:14 -0800 (PST)) it happened jodleren wrote in :

optoAndormeda deep space network CRC error best.

Yes.

Not much point in an optocoupler if there is a common ground.

In environments where PLCs are frequently used, there's no such thing.

"Common ground" is just an approximation in most complex electrical systems.

In the OP's case, things might work well if the TTL side is truly floating (battery powered or isolated power supply) and there are no other connections from the TTL side.

For instance, connecting the "common ground" inside the TTL system (or in a third device connected to the TTL system) to the (mains) ground can cause various currents, such as relay coil return currents flowing through the TTL system ground plane, causing voltage drops in the ground plane, reducing the noise margin _inside_ the TTL system.

Even if each subsystem is well designed with appropriate grounding arrangements and the whole system is designed with proper grounding hierarchy,things can still go wrong. An outsourced, not so well trained installer, will note that the drawing requires separate grounding wires to the specified grounding points, will also notice that there are several adjacent screw terminals with "ground" and just use short jumpers between them, destroying the grounding hierarchy:-(.

In industrial control systems, in which PLCs are typically use, you should use optoisolators at the TTL systems, not just to protect the actual input pin, but also keeping any unwanted ground currents out of your TTL system ground plane.

Thanks for all the comments.

I have common ground, as I plan to feed my item with the same power - that either by 7805 (max 0,8W) or a 24/5 swich power block. Either way, I can get common ground. I think that the solution with a Zener is the one I will go for.

Nothing and that's the way I would do it. Cheaper than a zener clamp and works for any input voltage up to the diode breakdown voltage.

Why make it more complicated than it needs to be etc ?...

Regards,

Chris

On a sunny day (Sun, 27 Feb 2011 15:59:24 +0000) it happened ChrisQ wrote in :

Optos are the best solution. Then the diode solution has the disadvantage that it can lift the supply, and you need 2 diodes to protect against negative (for example mains shorts) too. The zener limits negative, AND positive at or just below the supply voltage. Zeners (or transzorbs) have a higher capacitance, and reduce some transients. but really use an opto, ground faults will else get you sooner or later in a big way in industrial equipment. Also the opto gives a very good isolation for any incoming RF.

That I did not understand it at first - I am thinking of that now, due to the font here :) Later when I noticed that and copied into notepad, I could see it. It is simple and probably cheaper too, so I will try that.

Sonnich

too.

big way

Optos are a good solution for a very noisy environment, or where they may be ground loops, but the op has already said that this is not the case. You could also take a look at isolation devices like the Si Labs Si8420/50/60 series devices, which i'm currently looking at to build a jtag isolator for a debug adapter.

btw, John's solution has the diode facing *outwards*, with a pullup at the other side, so doesn't lift the supply at all up to the diode breakdown voltage...

Regards,

Chris

On a sunny day (Mon, 28 Feb 2011 09:10:52 -0600) it happened John Fields wrote in :

too.

I was talking about the common case of somebody wiring the mains to an input designed for 24 V. That happens in industrial, sure there are cases where one wires it all up onself, and even then... I remember the case where a technician connected the 24V PLC supply to the 230 V AC, several thousand dollars of damage, basically all new cards in the Siemens PLC. In industrial ANYTHING can happen and does. Sometimes the cable numbering is wrong, I remember a technician coming to me with " This design does not work'. What it turned out to be was that he was using the wrong signal cables :-) I mean those went through the whole factory, he lost track of the right numbers somewhere. So opto is usually safer, especially as you have huge ground differences, sometimes when big machines start up etc.

too.

big way

i.e. typical industrial environment.

the

Why would anyone believe this ?

Aw man, you're letting reality intrude.

Unless it's on a lab bench -- use an opto.

I make a series of a board for a customer, both opto and non opto versions. They wire the grounds of both sides of the opto together. They can't see the point of running more wires than necessary. The non opto version actually works better.

Forget the theory, how will it be wired?