NPN and PNP Is one safer?

So one maintenance guy in my plant advocates negative logic and supplying power to all Commons. The plant is switching to a new product line and the old modular machines are being phased out.

The maintenance man told everyone that the negative logic is tied to safety and that positive logic is not safe. He also goes into the concept of failure mode analysis about shorts and what he would rather see happen in that event. I would rather see a fused output or isolation relays on each output.

This concept mystifies me, but we are having trouble finding maintenance men that can follow negative logic.

My PLC designs and the main hardware EE are going to change the standard in the plant to sourcing outputs and sinking inputs. I just can't seem to find anything on safety and npn vs pnp. I need to make an informed decision.

Penny for your thoughts?

Somebody asked to Donald Bentley why he chose PNP transistor in his first designs and the answer was: because PNP transistors were cheaper than NPN ones.

All work must be done according to NEC 70 - national electrical code, NFPA 79 - electrical standard for electrical machinery, NEC70E arc flash, and any other codes that apply to your plant.

in regards to PNP vs. NPN.
I prefer power being applied to the device and using a sinking input, less power required and I believe its cheaper.

I would look at what is in the plant currently and stick with that philosophy !
the reason is simple, someone in maintenance will grab the wrong card and put it into the chassis and then apply power. lots of things go wrong at that point. It happens here with relays and sometimes with cards, , not often, but it happens.

james

Here is someone who argues that PNP is safer than NPN with regards to short circuits:
https://sensortech.wordpress.com/20...fundamentals-–-back-to-the-basics-npn-vs-pnp/ (go down to Henry Menkes comments).

I’ve heard some theories about the ease of trouble-shooting PNP vs. NPN control circuits. With PNP, if a load wire shorts to ground, the short can be found relatively easily because there may be visible damage due to the higher-than-normal current that would flow in the absence of a load impedance. An unprotected, unfused sensor may also have its output blown. If the PNP sensor has integrated thermal, latching, or pulsing short-circuit protection, the existence of a short circuit condition on the load wire will be apparent at the sensor.

In the case of a short to ground in an NPN circuit, there is no abnormal current flow; the load is simply pulled in regardless of the sensor state. This could prove to be somewhat difficult to detect, because if the sensor associated with the grounded load wire is examined, it will be found to have no fault.

Click to expand...

As for machine safety, PNP or NPN makes no difference. The sensors used must match the safety relays and safety circuits used.

edit: Maybe what the maint meant with "negative logic is tied to safety and that positive logic is not safe" was something about fail-safe versus not fail-safe. Again that has nothing to do with PNP vs NPN.

Yeah there is some mixing up in terminology here...NPN and PNP have nothing to do with negative logic or failsafe. If your maint man is arguing that failsafe logic is better it depends on the application. Take a high level switch used to stop a pump. It should be failsafe high....meaning the input should be on in the safe state of no level and go to 0 when level makes. That way in the case of a failed switch or broken wire the pump stops. But that has nothing to do with NPN or PNP. A high level failsafe switch could be either depending on how it's wired to an input module.

A broken wire for a NPN device (wired to a sourcing input point) would have a better chance to give a false positive to the input channel. If the wire breaks and grounds itself out, the input point would be pulled low, thus becoming true.

I always think sinking input points (that connect to PNP or sourcing devices) are easier to understand as voltage going INto the input point makes it true.
The same but opposite applies to outputs for me.
I prefer sourcing output points (that connect to a sinking device) so that the voltage comes OUT of the output point when true.

I always curse when I come across NPN or someone using the phase as common in a 1 phase control system. It just doesn't feel right, and quite frankly I don't know why they chose it? I just curse and say "it's those bloody electronics guys".

Then again this is what my boss also say so I just follow his lead!

I don't like NPN. It makes troubleshooting more difficult.

Let's say you have a sensor wired back to a panel through half a dozen junction boxes, and the on/off signal is not getting back to your MCC. If the sensor is PNP, at each j-box, you measure for voltage with reference to ground, until you find where it disappears. Easy. If the sensor is NPN, you have to measure for voltage with reference to 24V, which means you need a 24V reference at every point. You might have it, you might not.

And as others have said, if an NPN sensor shorts to ground, it will just appear to be "stuck on". This could be difficult to detect, especially if that sensor is meant to be on most of the time. What if that sensor is a NC high level switch? Sure, on wire break it will fail open and stop the process, but if instead of a wire break, a wire shorts to ground, the first you'll know about it is when your tank is overflowing. Whereas, if the sensor is PNP, then on wire break it will fail open and stop the process, and if it shorts to ground it will blow a fuse/trip a breaker/shut down the DC power supply/in some way or another, turn off the input and shut down the process.

The only coherent argument I've ever been given for NPN over PNP is that "if the sensor fails, you can just short the wire to ground to get the machine going again". To which my response was (after the initial blank look and facepalm), "well, if that's really a viable option, you could just short the PNP input to 24VDC back at the panel and achieve the same thing."

Another possible argument for it is that shorting 24VDC to ground in the event of a fault will cause sparks, whereas shorting 0V to ground in the event of a fault won't. In my mind, still a bulls*** argument, because (a) that only applies if it's the signal wire that shorts to ground - every sensor still has a 24VDC wire in there somewhere, (b) shorting a zero volt wire to ground CAN still create a spark, and (c) if causing a spark can create any sort of hazard, picking NPN over PNP, is NOT a suitable control measure. Hazardous area installations are expensive for a reason.

AS everyone here has said, there is no difference, safety-wise, between PNP and NPN. There is a difference, cost-wise, if your maintenance guy installs an NPN input card in your machine when it needs a PNP or if he configures a sensor for NPN operation when it should be PNP.

Or vice versa.

James is right. If you can, standardize your plant so you have one or the other. The easier you can make it for your maintenance guys to troubleshoot, the better.

Woo, harsh crowd!

As an electronics-before-automation guy, every panel I've designed since I've been building panels is sinking logic, both in and out.

I see one possible pro for sourcing is easier to find a fault throughout a series of j-boxes. That's great if your conduit is bonded to your DC 0v, but is anyone actually doing that? I feed +v and 0v from my panel straight to the load or sensor in every design, so this seems like a wash?

Brandon_K said:

Woo, harsh crowd!

As an electronics-before-automation guy, every panel I've designed since I've been building panels is sinking logic, both in and out.

I see one possible pro for sourcing is easier to find a fault throughout a series of j-boxes. That's great if your conduit is bonded to your DC 0v, but is anyone actually doing that? I feed +v and 0v from my panel straight to the load or sensor in every design, so this seems like a wash?

Click to expand...

Things are a little different across the pond...

I see a lot of american machines where all the wiring is single wires in conduits. That's just not done here - double-insulated cables for everything, and usually just on cable trays.

Typically, the 0VDC is grounded (although that's a whole 'nother argument, lets not start that one just now), and every piece of metal within cooee of anything electrical is earthed. So I can just stick my meter probe on whatever handrail, I-beam, machine guard, cable tray or other exposed piece of metal I can find, and 99.9% of the time, it's a valid reference.

I teach my coworkers to use sinking inputs in plc and sourcing outputs.
This is because I was taught in school with AC control, I prefer to leave the Neutro wired and interrupt the Hot, and testing for a short or damaged was very easy because ground was everywhere in the enclosure but not the hot. If i suspect about a sensor or a switch, i simply put one test lead into the output of the sensor and the other lead anywhere, even under my boots. Same for plc outputs.

and for safety, that doesnt have anything to do with NPN or PNP, that is more related to safety devices, fuses, monitoring relays, overload, and of course, the performance level, CAT & SIL

but, I remembered a guy that works with Virtual Instrumentation, he had a good theory about using NPN, but only when using opto/electronics devices such as DAQ cards. This is something related to the pull-up, pull-down in TTL circuits. things I used to know when I was electronic student

Have heard all these arguments and reasons before. If you do not ground the 24VDC negative from a switchmode power supply and allow the 24VDC to float it does not matter - there is no ground reference. NPN, PNP - does not matter. Have used both and still do. By the way in the early days NPN transistors were cheaper than PNP by a fair margin - I do remember. NPN outputs are still extremely common in Japan as are NPN inputs. From what I can see the Europeans seemed to drive the PNP output situation. There were many PLCs that did not even have NPN outputs available from Europe.

It is many years since I last used a PLC with NPN outputs (Omron as it happens) but one thing I do remember is that if you removed the power from the PLC while the power to the outputs was still on then all of the output transistors would conduct causing all of the outputs to turn on. This could definitely lead to safety issues if the system was not designed to accommodate it.