Recently PLC Power Supply Modules, Communications Modules and CPU got damaged. These module are under warranty period but when we check with the OEM, they have said the damage due to High voltage Surge/Improper Grounding. The PLC Panel is receiving 220V DC Supply from DCDB and inside the panel 220V DC-24V DC Convertor is present.

Inside the panel, we have Instrumentation Earth bar and Protective earth bar. To protective earth bar, we have earthed all panel doors (Including side doors) and also earthed 230V AC (Non-UPS Supply). we have not connected Mounting plate to Protective earth. We have not earthed the above DC_DC Convertor and we have not earthed the above PLC Modules.

Now, if the Modules failures are due to High Voltage surge, how is it possible that DC-DC Converter working fine and its load item PLC Power supply modules, Communication Module & CPU module damaged due to Surge?

if the Failure is due to grounding, what we need to correct to come out of the problem.

 

This is kinda tough without being able to see the big picture, but if they are correct, here are the most likely possibilities:

The DC converter is not "just fine" and passes spikes from the input or generates it's own.

The floating (ungrounded) 24V power supply is occasionally lifted to
destructive voltages.

The power supply is staying near ground and something attached to the
PLC is lifting it to destructive voltages.

In my view, the low side of all the DC power supplies should be grounded unless there is a really good reason why not. The backplane should be routinely bonded to the cabinet and doors and grounded, that's a requirement in this part of the world. The PLC should be grounded with the terminal provided.

I like for these to all be connected to the same point (star grounding). If that's not convenient for some reason I have a large single ground lug for the house power and another single grounding point for the electronics, both securely bolted or screwed to the backplane. You can handle instrument ground as needed. I use separate grounding only if it's shown to be needed. If you plan for it it should be only removing a link.

The reason I do this, after long experience, is so that I never have the problems that you are having. And if I did, there are rational points to connect MOVs or snubbers, etc. There are a few situations where the rule must be bent, but I'll generally use a relay or optoisolator to isolate things that cannot be reasonably grounded. It greatly simplifies troubleshooting and has safety implications too.

I can't suggest how to fix your situation as the relationship between the various power and signal system is undefined. But I would be studying the prints on how to apply my system so they have a definite relationship. With all that alleged ungrounded circuitry, overstress is much more likely.

Others may disagree, but I have spent a lot less time chasing bizarre voltages and strange glitches since doggedly following these rules. The ground terminals are there for a reason. And if you aren't sure exactly what to do with them, find out. Your days will go much smoother.

Generally, when a vendor attacks your grounding, they are guessing because they don't have any more ideas. But that's _only_ if your
grounding is correct. It's a great catch all because it's so poorly understood that they get the benefit of doubt.

I think all automation types should some time with high speed analog ICs. Those circuits simply will not work without proper grounding. You learn ho important it is.

Regards
cww

 

A few thoughts and questions which might be useful.

You mention "Instrumentation Earth bar" - so is this a process application where a mixture of field analog and digital I/O are transmitted from the plant over longish distance to a control room? If not, what is the application?

Is all the I/O instrumentation powered directly from the DC/DC convertor mentioned? If so, the implication is that this common 0v fans out far and wide to the I/O and is possibly vulnerable to EMI. Is this 0v tied to your "Instrumentation Earth" that you mention?

Has the vendor inspected the damaged DC/DC convertor and reported on where the surge entered the unit?

Is there still any construction work and/or welding going on, on site?

Had any thunderstorms lately?

If this is a process application, I'd expect to see a separate DC power supply for the field I/O. Whether this has a grounded output depends on whether analog signals are isolated - which is often preferred.

And the most basic question - are you sure that your site ground quality is good?

 

I would suggest this. There are some common lies in the automation industry. They often involve statements that use terms like:

good grounding
bad grounding
noise
surges
transients
shielding

They are often used when someone cannot explain why something undesirable happened.

Personally, I think you are better off with a common ground as opposed to isolated DC power supplies. Tie all the DC power supplies to ground. Usually this involves attaching a conductor from the DC common directly to a known ground. It is not necessary to run it back to the service ground point. Usually the equipment ground is quite adequate. This will make sure that all the DC voltages have the same reference point.

Most DC power supplies have a fair amount of isolation from input to output so there is not all that much chance for so called surges or transients to pass through them.

 

"Personally, I think you are better off with a common ground as opposed to isolated DC power supplies. Tie all the DC power supplies to ground. Usually this involves attaching a conductor from the DC common directly to a known ground. It is not necessary to run it back to the service ground point. Usually the equipment ground is quite adequate. This will make sure that all the DC voltages have the same reference point."

Bob - this has advantages in a factory / machinery / control network environment, where the dc power supplies and I/O are physically local to the controller, but less so (here I believe) in process systems where some of the I/O signaling may be to/from remote equipment and dc power supplies may be remote too.

If the 0vdc side of the (remote) field equipment power supply is locally grounded, it's likely that, in time, ground loops will interfere with analog signals, which is one reason why analog isolators are widely used.

Your comment about DC supplies input / output isolation is one thing, but what the output DC is connected to can be a larger vulnerability for the system.

 

Thanks:Curt Wuollet,

Sir in your view you have mentioned " the low side of all the DC power supplies should be grounded unless there is a really good reason why not. The backplane should be routinely bonded to the cabinet and doors and grounded, that's a requirement in this part of the world. The PLC should be grounded with the terminal provided."

As mentioned in my query, Since all panels Doors & Side panels are connected by earth wire to Protective Earth (Non-Isolated) and assume this will mean that My Mounting plate was also earthed to protective earthed as it mechanically fixed inside the panel. Since the back plane is mounted Directly on the Mounted panel(Without any Insulators) which in turn mean that my backplane was also grounded without use of ground wire (green/Yellow) to protect earth. And Since Modules Rack Power Supplies and PLC Modules are installed in the back plane. I assume that these are also grounded to protective earth.

That is the reason we have not grounded. In view of the above do i need to bond the backplane to cabinet Door/frame with some ground wire and also connect all the lower side of all the DC power Supplies to Protective earth.

 

Thank you Mr Alan

Hope the following points would give better insight to my query;

The Industry is Hydro Power plan and plant capacity is around 8 MW.
The Instrumentation earth bar is basically Isolated with separate earth pit called electronic earth pit with Earth resistance less than 1 Ohm.

My 220V DC/24V DC Converter is connected to a Bus bar inside a panel. from that bus through fused TB's we have used for Field Interrogation Purpose & as well as for PLC Rack Power Supply Modules. DC-DC Converter is perfectly healthy and working with-out any Problem.

The Site is Under construction and as such welding other works are going on.

Analog Signals are non-Isolated.

As Such Since the Power Plant Capacity is very Low, We usually go with Common DC-DC Converter for PLC Power Supply Modules as well as Filed I/O Interrogation.

Grounding scheme adapted in the plant is not clear as it is in the Scope of EPC Vendor and We are only PLC Panel Supplier with Commissioning of PLC Being in Our Scope. So they are claiming grounding scheme from their end is correct and they are suspecting the problem is from Our Panel only.

 

In my world view, yes. Especially the power supplies. The door grounding is a code requirement and not necessarily relevant to you problem, but it's a real good idea anyway. The power supply grounding is important because floating systems can easily, through leakage, triboelectric (static) charging be elevated to voltages outside the common mode voltage range of attached devices which can cause exactly the kind of failures you are experiencing.

Let's take a simple example: You have a transistor radio operated from an isolated 9 volt battery. Lets take a 1000 v supply with the negative side grounded and hook to either terminal of the battery. As long as the radio remains insulated it will still work just fine, I'd be careful tuning it though. But, now if any connected part happens to touch ground, the radio would most likely be toast. It works the other way with ungrounded supplies. It's very likely something on the other end of the wire _is_ referenced to ground or a different voltage. I'm not sure that's the most effective explanation, but it's been a long day.

If all the parts are referenced to ground it's far less likely that they will see stressful voltages.

Regards
cww

 

Dear satyasrinivas1

Thanks for your additional info.

If you have welding going on, your I/O is non-isolated, and it's all fed from the same DC/DC convertor, it seems possible that someone like a welder has at some point grounded his machine onto a structure which has one of your sensors/output also grounded to it. Bursts of EMI, which occur when welding machines discharge to ground, could then have got back to the output of the DC/DC convertor and from there on to the PLC. If possible, you may want to disconnect the I/O, or output of the DC/DC convertor (to the I/O) during the plant construction phase.

With a Hydro Power installation - although I don't know details - I wouldn't expect there to be many outside sensors or actuators, if any. But if there is any analog I/O outside, and they have non-isolated signals, I strongly recommend that you at least fit isolators, - and surge protection too if the site location is in an area where thunderstorms are common.

If I may, I'd advise you to get familiar/clarity with regard to the grounding associated with your PLC I/O and its signalling/shielding.

If you'd like to communicate direct on this matter, you are welcome to email me on albalcombe [@] gmail [.] com.