Sinking or Sourcing


Thread Starter


I have recently moved into the controls engineering department of my company. In the past we have always ran a hot or positive wire to a limit switch and used that for an input to the PLC. As a field tech this has been a cause of fustration due to blowing fuses and such. I would prefer to run the common to the switch and use that for an input. Are there any negatives to this approach that I should know about before I present my case?


Rich TECH411

When you switch a hot if there is a short in the wire or an open you know right away... when you use a common a short can look like contact or an open can look like contact.

Bad things if your controlling or protecting.
Not really, Safety circuits usually use hot wires in case they short to machine ground they will trip to safeguard, but normal inputs I think its just preference.


William Sturm

That is commonly done, but it has advantages and

advantage: easier to connect sensors in parallel, especially if they have different power sources.

disadvantage: if the wire short to ground, it will appear that the sensor is on. This is only an issue if the DC supply is ground referenced.

disadvantage: in the US, the big name PLC's tend to use sinking input cards. (switching the + side)

advantage: many drives and japanese PLC brands use the sourcing input cards. (switching the - side)

I'm sure there are many others.

Yes, while a fairly common, this practice poses one serious hazard. It's much easier to generate a false signal with an unintentionally grounded circuit than an unintentionally hot circuit, because there are more sources of ground to come in contact with (conduit, box walls, etc.). So if someone runs into a conduit with a forklift, the associated motor starts when it shouldn't. I'd rather deal with the blown fuse.

Steve Myres, PE
Automation Solutions
(480) 813-1145
I agree with the other observations and thoughts so far, but unclear why using a 'hot' line to the switch or sensor and a sinking input card to read the signal is causing blown fuses.

In my experience this occurs with input circuits primarily when a wire short circuits to ground in a crushed/mangled conduit, cabling saws into a grounded object, or the like. If that is the case, then it would be preferable to attack the root cause; reroute conduit/cabling to take them out of harm's way, fabricate protective guards, add strain reliefs, and so on.

William Sturm

The issue is how the control responds in the event of a broken conduit, chafed cable... A sourcing input may view the broken wire as a valid true signal if it becomes grounded. Motors could start, presses could cycle... A sinking input circuit would be much more likely to have a broken circuit appear as a false input signal. A wire shorted to ground could not be misconstrued as a logically true signal. For a false true signal, the wire would have to be shorted to a voltage source. This is much less likely to happen.

Steve Myres, PE

Excellent point. Just because your machine wiring is screwed up in some way is no sign you should screw up the design to match. Just fix the machine.
I agree with the other observations and thoughts so far... and with everything you've said, too.

My point was that, instead of using a sourcing input to prevent fuses from blowing (but which would trade that problem with even more difficult-to-diagnose, and potentially dangerous 'machines gone wild' symptom), to find out why the fuses are blowing, and take appropriate
Thanks for the comments. I had not considered the false signal. I suppose I should have said the application is Liquid Natrual Gas and CO2 plants. A blown fuse would probably be better than a blown up plant.



marc sinclair

Hi, If the signal is critical or safety related, then all switches should be monitored and_or duplicated. A limit switch which consists of NO and NC contacts and requires changeover to signal, will produce ambiguous signals with a cable break and would need two different faults to produce a false signal. I have worked on natural gas control systems on oil platforms, and it is not unusual to have five and sometimes seven levels of redundancy.

Marc Sinclair
Be careful of relying on blown fuses as a safety device.

I know of two plants in which a blown fuse started a cascade effect that resulted in a near catastrophic event. Only quick response saved the day.

Always rely on an holistic approach to fail safe.

Bob Pawley
You're not going to tell me I have to light candles, put on some new age music and massage the machine, are you? ;-)
Some of this is more a matter of programming philosophy as to whether the sourcing or sinking is really a safety problem, and the environment you work in.

If you program your controller so that events in a sequence are properly interlocked with internal bits (or words) i.e. event A always happens before Event B, regardless of the input status, it will minimize the catastrophe from a false true input status. Many folks are still relying on parallel, non sequenced ladder logic to trigger events, and, to me this is the root cause of many problems.

For instance, who here knows of a machine where you walk by a photocell and some ejector arm or similar mechanism just starts actuating, even if the machine has not loaded a part yet?

If you want to safeguard your machine you will sequence things properly, and if it *really* matters (something could jam and break something), you will check your limit switch to be in an "off" state before you actuate soemthing that will make it in the "on" state, just to make sure it is not shorted to ground *or* +24V.

Of course, none of this speaks of safety circuits, which IMHO should be hard wired always, and these should in fact be such that a short to ground causes the safeguard condition to be true.

I agree totally. Get the book "Cascading logic" from ISA press.

I constantly argue with people about "anyone can program" , the secret is exception handling etc. I make a great living cleaning up quick, dirty, outsourced, downsized, employed for a year and move on.............programmers code.

Just batting clean-up in the lineup..........


Steve Myres, PE

I agree with your main point that having inputs circuits properly designed is not a substitute for proper software design. However, being a belt and suspenders kind of guy, I feel the converse is also true: good software is no excuse for doing the I/O in what I feel is the more hazardous of two choices of equal complexity and cost.

Having said that, I freely admit to having done sinking systems when connecting to components that only offered negative logic I/O. I'm just trying to stand by my original point that the choice of I/O flavor does matter.
Sourcing inputs and sinking outputs are generally more popular and CE (Conformity European). These saves a PLC I/O card from smoking when shorts occur. However, this can be a trade off for more or less reliability depending on your goal. For example, some auto manufactures standardize on ungrounded circuits with sourcing inputs/sinking outputs. If there is a short it won't stop the line and maintenance can troubleshoot the problem without sacrificing lost production. On the other hand, and a lot of controls people will tell you a grounded circuit with sinking inputs/sourcing outputs is the best. Should any circuits short to ground this will trip or fry something. The resulting affect may be a signaled warning or white smoke. This may be seen as a safer alternative, but can blow more devices.

I don't believe this was a part of the original question, but some of the responses referred to N.O./N.C. switch option. The question of which is safer will depend on the application. Usually a high status bit indicates an OK signal. This would a more reliable indication for sinking inputs (a shorted wire will not cause a high bit).
As not to confuse anyone. Sinking inputs and sourcing outputs are generally more popular and CE (Conformity European). This is a correction from the previously mentioned reply above.

marc sinclair


CE has nothing to do with Sourcing or Sinking, If you are interested in understanding what CE is all about
is a good starting point.

You say more popular, where? NPN sensors (or ones configured for use with sourcing inputs) are rare in Europe, in 30 years of work in the food industry, I've not seen one single system with sourcing inputs and sinking

If a sinbgle fault has the capability to 'smoke' your PLC I/O card, then your design is at fault. I design control systems, one of the considerations is that the inputs and outputs may short circuit. This is not a problem at low
voltages as power supplies will fold back at a current limit. Higher voltages should be fused at above the normal running current and below the level at which damage can occur, this is very basic stuff.

Marc Sinclair