Effect of transformer grounding on shielding

fxb80 · Aug 8, 2010

“Do not ground the secondary of the 24 V power supply in
the control system. This will create a secondary current
path and negate the protection of any shielding.”

This is from a Honeywell document. Anyone care to confirm, deny, or modify this statement, or direct me to a thread discussing this?

Thanks

s2sam · Aug 8, 2010

fxb80 said:
“Do not ground the secondary of the 24 V power supply in
the control system. This will create a secondary current
path and negate the protection of any shielding.”

This is from a Honeywell document. Anyone care to confirm, deny, or modify this statement, or direct me to a thread discussing this?

Thanks

Good day fxb80,

Although I cannot comment directly about Honeywell's intentions, shields are meant to minimize induced noise on the inner conductors within the shielded cable where most of this noise will be from local AC power. So if you power your equipment from a one-leg grounded transformer and have a shield (assuming it is grounded) you are creating a potential current path... With current leads to magnetic fields which can couple to your shielded inner conductors (Faraday's law of Induction)... This coupled energy will result in electrical noise on your signals which may cause erroneous values on them...

Secondly, grounding one leg of the transformer also exposes your system of potential harmful (and lethal) ground currents that can occur. Using this approach creates a ground path for high voltage signals to use which is not good if you happen to have an electrical fault in some secondary equipment. Further, because ground potentials can vary (because of lightning strikes, etc) you can get ground currents flowing between differing connected equipment which is also a bad thing.

Remember transformers provide galvanic isolation through magnetic coupling and so keeping both secondary lines from ground provides maximum protection, etc.

Cheers,

Sam

klrogers · Aug 8, 2010

Not directly related to your question, but National Electric code (2008 250.2(A)(1) requires that if the transformer is supplied (primary side) with more than 150v to ground, then the secondary side must be grounded. Just something to watch for.

Kevin

crab master · Aug 8, 2010

Personally I prefer the secondary side grounded. Makes troubleshooting easier. Also as long as the leg is only grounded at the transformer, if you were to take a hit and it blew out the 'common' leg before it got back to the transformer you'd lose your ground reference, short of it shorting the common wire to ground. So only grounding at the transformer should 'never' allow for current to flow to ground, right?

So if the condition of the shield and the secondary grounded is "creating a potential current path" what would be recommended - a diode on the shield?

Even without the diode flow would only be one direction, to ground and isn't that the purpose of the shield, to make any inductance put on the line travel back to ground to minimize/eliminate noise on the cable? I don't see how a shield grounded at one end and a ground reference on a transformer would allow anything to try to travel back up the shield. Isn't noise due to inductance and signals trying to travel multiple directions or not being able to travel out of the wires? Is the document in reference a CYA for those not installing shields correctly?

s2sam said:
Remember transformers provide galvanic isolation through magnetic coupling and so keeping both secondary lines from ground provides maximum protection, etc.

Will you explain this more?

wolfdog · Aug 8, 2010

klrogers said:
Not directly related to your question, but National Electric code (2008 250.2(A)(1) requires that if the transformer is supplied (primary side) with more than 150v to ground, then the secondary side must be grounded. Just something to watch for.

Kevin

You might want to double check that.

Artrose · Aug 8, 2010

Secondary grounding...depends on the control system. Obviously, if they specify no ground, don't do it. Grounding the secondary can cause all sorts of issues with some systems (even blue smoke), and particualr attention must be taken when wiring components into a system that has it's secondary grounded, and yes, this opens the door into a discussion about the difference between bonding and grounding, the whys, and the why nots....
Simplified.....if we're looking at the issue of releasing interference or noise that often accumulates in communication wiring, you ground (drain) one end of the cable. If you were to ground both ends of the cable, you are essentially creating an antenna.

klrogers · Aug 8, 2010

wolfdog said:
You might want to double check that.

Well I got the ref. wrong its 250.20(A)(1), but other than that I stand by the statement, I do not know of any article in section 725 (Class 2 wiring is in this section) that allows exemption from this requirement.
Of course any manufactured assembly that includes such a transformer does not need to comply if it is suitably UL listed.

Kevin

crab master · Aug 8, 2010

Artrose said:
Secondary grounding...depends on the control system. Obviously, if they specify no ground, don't do it. Grounding the secondary can cause all sorts of issues with some systems (even blue smoke), and particualr attention must be taken when wiring components into a system that has it's secondary grounded, and yes, this opens the door into a discussion about the difference between bonding and grounding, the whys, and the why nots....
Simplified.....if we're looking at the issue of releasing interference or noise that often accumulates in communication wiring, you ground (drain) one end of the cable. If you were to ground both ends of the cable, you are essentially creating an antenna.

Agreed, but the statement in reference is mentioning the affect groudning has on shields. I can understand in full wave and half wave rectified devices, but not simply due to 'potential noise' on the shield.

fxb80 · Aug 9, 2010

Much of my work is 480V. To use ungrounded power I will need to quit using 480/24 grounded control transformers and use 480/120 stepdown transformers feeding 120/24 ungrounded control transformers. This is, in fact, what I see Trane do in their equipment where an automation controller is factory installed.

If an actuator is fed with grounded power and shielded cable, it seems there would be the potential for interference to be induced by external sources within the actuator itself if the shield is connected to earth at the controller end, since the shield wouldn't be connected to the actuator. Grounding the actuator enclosure and grounding the shield at the actuator end would seem to create the potential for current flow in the opposite direction if the power is grounded, plus opening a can of worms about where to land which shield. So why not eliminate the possibility altogether by not grounding the transformer?

s2sam,
I also would like to learn more about galvanic isolation.

s2sam · Aug 9, 2010

crab master said:
<snip>

Remember transformers provide galvanic isolation through magnetic coupling and so keeping both secondary lines from ground provides maximum protection, etc.

Click to expand...

Will you explain this more?

Good day CrabMaster,

My comment was directed to a couple of items that I grouped together:

1. Having a "floating" (i.e. non Earth grounded) power supply to a device ensures that you do not have a potentially dangerous current path if an electrical fault should occur in the powered equipment or perhaps other devices connected to this powered device. If the equipment has no Earth ground, then no current can flow which is good!

These electrical faults can also be the result of a person touching various pieces of equipment that may be improperly wired and/or Earth grounded... If the Ground of the powered equipment is not exactly the same location as the other equipment, then the person themselves can provide the current path... which can be dangerous and potentially lethal depending upon the circumstances.

2. Secondly, any time you power a device that has an Earth ground as part of the device's power supply you effectively couple the 60 Hz AC noise into your system. Ideally one would want to minimize this in order to minimize this noise on your communication and other sensor inputs. I say this in a general sense, as at times multiple devices can be interconnected and who knows how they are powered and connected and so electrical noise can result. If you were to correctly wire and power everything appropriately then the 60 Hz AC noise can be minimized substantially.

3. Thirdly, a lot of people do not realize that Earth ground is not necessarily the same (electrical) potential unless the grounds are attached at exactly the same physical location. Lightning strikes, high current surges, etc can occur in physically different areas and so the Earth ground potential at location A can be different than at location B... and so if you have two different potentials, current will flow.

Cheers,

Sam

s2sam · Aug 9, 2010

fxb80 said:
<snip>

s2sam,
I also would like to learn more about galvanic isolation.

Good day fxb80

Effectively galvanic isolation is a mechanism that ensures that two devices do not have an electrical connection between then (i.e. electrons do not flow between them) and yet signal information can be passed between them. There are several methods of providing this isolation and are dependent upon the application. For example, you can have optoisolators where the interface is via an optic coupling, or magnetic such as a transformer or its variants, or even mechanical (Relay, etc). Here the actual "information" is being transferred via these coupling mechanisms and yet there is not a physical current flow between them.

Ideally isolation is highly desirable but is not used all the time because of cost (added electrical components and circuit complexity) or perhaps the limitation of the coupling modality (i.e. signal characteristics, voltage isolation limits, speed, etc).

Since the HVAC World can include be a number of interconnected devices (i.e. communication buses, sensors, etc) where currents can flow between these devices and if these currents are unexpected and undesirable havoc can occur. In an ideal World all of the equipment would be properly designed and so field issues can be minimized. However, in the real World a number of field devices are not suitable designed with these interconnections in mind and so we have problems. Also, some equipment even if designed correctly may also be installed and powered incorrectly which the manufacturer has little control over... and so again problems result... anyway I digress... I hope I answered your question.

Cheers,

Sam

s2sam · Aug 9, 2010

crab master said:
<Snip>

So if the condition of the shield and the secondary grounded is "creating a potential current path" what would be recommended - a diode on the shield?

Even without the diode flow would only be one direction, to ground and isn't that the purpose of the shield, to make any inductance put on the line travel back to ground to minimize/eliminate noise on the cable? I don't see how a shield grounded at one end and a ground reference on a transformer would allow anything to try to travel back up the shield. Isn't noise due to inductance and signals trying to travel multiple directions or not being able to travel out of the wires? Is the document in reference a CYA for those not installing shields correctly?

Will you explain this more?

Good day Crab Master,

If we were to make some assumptions then effects of the grounded leg of the power supply can be minimized.

1. You do not have any other interconnected devices (i.e. communication buses, sensors from other equipment, etc)

2. The shield ground location is exactly the same point as the power supply ground.

I reality your equipment may not have both of the above... or may change in the future... i.e. additional devices added, wiring changes, etc and so problems can result.

Remember the manufacturer gives generalizations and suggestions and has little control one how the equipment is installed and potentially modified in the future. Could the document be a FYA? Absolutely, how can they troubleshoot their equipment on a site with who knows what for wiring? Secondly, installers vary with experience, knowledge, qualifications, and understanding and so this too factors into their motivations.

Cheers,

Sam

crab master · Aug 9, 2010

Thanks for the descriptions thus far.

If I basically understand what you are saying, then I might as well have my installers pull a 'neutral' to my control panels and in turn 'ground' all my 'grounded' transformers to the 'neutral' and take all my shields to ground. Then I know I should have a single path back to the common ground point.

The part that confuses me here is trying to simply design around 'what if's'. If my control system takes a hit by a lightning then I expect to have a number of fixes to do. If my customer wants things designed around lightning strike potential then he needs to get an EE involved and get it designed as such. I can't reasonably design every control panel around the consideration of a lightning strike, in my mind that's the duty of the electrician, short of me creating a high point - weather station, OA Press. reference, but in turn all such devices (in most) have been tied into the lightning strike grid, by the electrician that put together the grounding system.

If my control panels are designed with a 'properly' sized ground and the ground is only used as a reference and not a neutral will electricity not travel in the path of least resistance? I understand that there may be resistive paths that vary slightly between control panels due to longer wiring runs, etc. but I would think even at that the overall signal interference would be minimal and likely not even noticed. Maybe it is just my ignorance on the subject, but for me a better understanding would be great so I can make sure I am designing my control systems properly. Personally I don't design/install floating secondaries unless forced to do so by equipment spec sheets. It's just way too much fun trying to find where you lost your 'common' in a floating system, whereas a grounded reference makes it very easy.

wolfdog · Aug 9, 2010

You cannot design for the unexpected.
Friday night I was called by a customer because the conduit to the air cooled chiller had blown out. It was run in conduit with a prorly sized ground conductor. I know because I installed it 7 years ago. The cable TV people had grounded their box to the 2" chiller conduit with a clamp...that was not tight. A power company feeder cable went down in the alley and hit the cable feed. The cable feed went hot - don't know what voltage but I suspect 13,000- and went to ground thru the clamp on the chiller. It arced and heated up, melting the insulation of the 1/0 conductors inside. Then the fireball of a phase to phase short blew out a whole in the conduit and in the brick. It took out the feeder fuses to the chiller. I was 1:30 in the AM getting temporary power restored and the chiller back on line. All because of a loose ground clamp.
You cannot design for the unexpected.

Artrose · Aug 10, 2010

In our area lightning strikes to building control systems are not normally a major issue. If there's concern, we use lightning arrestors on the communication runs between buildings. I'm not trying to be a nanny, but, a word of caution....follow the control system manufacturers installation recommendations. There's lots of short cuts and reinventing the wheel that you might pull off, especially on smaller installations......but, if you don't follow manufacturers guidelines, kiss your factory tech support and warranties goodbye. I'm sure I'm not the only one here who has spent many frustrating, wasted hours on some kind of intermittent issue created by a poor or reinvented installation. It can be extremely expensive for both you and your customer, and honestly, can ruin reputatiuons. Sometimes it might cost more in time and money, but there's good reason behind why controls systems and component manufacturers invest so much time and money providing proper installation documentation.

s2sam · Aug 11, 2010

crab master said:
Thanks for the descriptions thus far.

If I basically understand what you are saying, then I might as well have my installers pull a 'neutral' to my control panels and in turn 'ground' all my 'grounded' transformers to the 'neutral' and take all my shields to ground. Then I know I should have a single path back to the common ground point.

The part that confuses me here is trying to simply design around 'what if's'. If my control system takes a hit by a lightning then I expect to have a number of fixes to do. If my customer wants things designed around lightning strike potential then he needs to get an EE involved and get it designed as such. I can't reasonably design every control panel around the consideration of a lightning strike, in my mind that's the duty of the electrician, short of me creating a high point - weather station, OA Press. reference, but in turn all such devices (in most) have been tied into the lightning strike grid, by the electrician that put together the grounding system.

If my control panels are designed with a 'properly' sized ground and the ground is only used as a reference and not a neutral will electricity not travel in the path of least resistance? I understand that there may be resistive paths that vary slightly between control panels due to longer wiring runs, etc. but I would think even at that the overall signal interference would be minimal and likely not even noticed. Maybe it is just my ignorance on the subject, but for me a better understanding would be great so I can make sure I am designing my control systems properly. Personally I don't design/install floating secondaries unless forced to do so by equipment spec sheets. It's just way too much fun trying to find where you lost your 'common' in a floating system, whereas a grounded reference makes it very easy.

Good day Crab Master,

My points made are generalizations and are to provide some insight as to what can cause problems and/or what are the trade offs doing it one way or another. As Artrose mentioned, one must follow the manufacturer's directions or face voided warranties, etc. They (my comments) are not meant to be the defacto standard, but merely to provide another perspective as to what things should be considered.

As for "what if's" or fault tolerance... Since I am an Engineer I always design for the worst case scenario as it is much easier for me to account for this in my designs than to deal with it in the field. It is a lot less costly to add more effort ($$$ and time) into the design than it is to replace a device under warranty and/or support calls. Further, at times we see issues resulting from site problems, etc that have nothing to do with our equipment and yet we spend a lot of time assisting our customers with sorting out the problem. Given the benefit of these experiences, we design our equipment accordingly. That being said, I have seen a number of issues related to large ground currents and line voltage faults which escalates them from a rare event to one that must be considered (at least by me and our firm). So, if the upfront effort is rather small (in terms of cost and effort) and I protect the system from these events then all of my customers are very happy with that. As an example I have one system (comprised of over 16 interconnected elevator interfaces) that has been installed since 1996 that runs 24/7/365 and has yet to have one failure in that time... was the design overkill? You bet, but the customer is very, very pleased, as there has been 0 support calls for this equipment given the nasty environment it is in (600V DC, large inductive loads, large AC noise, etc)... and which make us and our equipment look excellent. We are by no means perfect, but we examine and analyze all field issues and listen to customers to see if there are any design elements we can add to our devices to improve them...

Cheers,

Sam

crab master · Aug 11, 2010

Sam,
Although I agree with most of what you are saying in your last post, short of you still can't design around everything, (whoops I didn't mean to put 120 on the comm circuit or cutting them there wires caused a problem?)

I still don't believe I understand the reasoning behind the explanations given to the OP's statement.

fxb80 said:
“Do not ground the secondary of the 24 V power supply in the control system. This will create a secondary current path and negate the protection of any shielding.”

I can understand if you start taking various ground references for your secondary, other than just at the transformer itself. For example at the compressor contactor one adds a connection to the panel base for a ground reference in parallel to the common back to the transformer and then later someone/something causes a disconnect on the common reference at the transformer itself, and now your ground reference at the contactor just became a neutral. Now if we were add to the scenario that the shields are not terminated properly, one end is connected to ground, but the other end touching ground cause it didn't get taped off, and the panel with the compressor contactor (who's common was taken to the panel base) doesn't have a solid ground actually ran to it and so you end up with current a loop, now I can understand a major mess and potential for noise issues. Had a similar scenario years ago with transformers in parallel....

Here's the thing, most controllers I work with allow for a ground referenced secondary so I install them with a ground referenced secondary. While I try to only run shields in areas I suspect may be noisy, I have yet to come across a problem as described in the OP's statement. I've gone back many times and re-pulled shielded wire/terminated the existing shielded wire correctly to fix noise issues. 99.9% of these were in grounded secondary systems, so I can't believe a grounded secondary will cause a problem with a properly installed shielded wire system. I see more shields that really aren't installed correctly, some are close and many are far from it and I see that as the bigger problem.

Also I do agree that systems that want a floating secondary install it as such. Learned the hard way by mixing full-wave and half wave rectified devices and smoked a device/two (that's all I'll admit to anyway

)

s2sam · Aug 11, 2010

crab master said:
Sam,
Although I agree with most of what you are saying in your last post, short of you still can't design around everything, (whoops I didn't mean to put 120 on the comm circuit or cutting them there wires caused a problem?) I still don't believe I understand the reasoning behind the explanations given to the OP's statement.

I can understand if you start taking various ground references for your secondary, other than just at the transformer itself. For example at the compressor contactor one adds a connection to the panel base for a ground reference in parallel to the common back to the transformer and then later someone/something causes a disconnect on the common reference at the transformer itself, and now your ground reference at the contactor just became a neutral. Now if we were add to the scenario that the shields are not terminated properly, one end is connected to ground, but the other end touching ground cause it didn't get taped off, and the panel with the compressor contactor (who's common was taken to the panel base) doesn't have a solid ground actually ran to it and so you end up with current a loop, now I can understand a major mess and potential for noise issues. Had a similar scenario years ago with transformers in parallel....

Here's the thing, most controllers I work with allow for a ground referenced secondary so I install them with a ground referenced secondary. While I try to only run shields in areas I suspect may be noisy, I have yet to come across a problem as described in the OP's statement. I've gone back many times and re-pulled shielded wire/terminated the existing shielded wire correctly to fix noise issues. 99.9% of these were in grounded secondary systems, so I can't believe a grounded secondary will cause a problem with a properly installed shielded wire system. I see more shields that really aren't installed correctly, some are close and many are far from it and I see that as the bigger problem.

Also I do agree that systems that want a floating secondary install it as such. Learned the hard way by mixing full-wave and half wave rectified devices and smoked a device/two (that's all I'll admit to anyway )

Good day Crab Master,

Indeed, you are right... one cannot design an equipment for every possible failure... with human error being the most difficult... However I do try but mostly for my own blunders

I think the original comment was a generalization and should be taken as such. Some equipment, especially interconnected devices from different manufactures, are perhaps not designed appropriately with a grounded power leg. Secondly, not all installers have the same knowledge and skill and so perhaps it is easier to make a statement like this. Remember your your frame of reference is yourself where your knowledge and skill is excellent. However, the manufacturer's reference may be entirely different in that they may have had customers doing who knows what and so they make a blanket statement. Hard to say what the motivations are.

From my case, I much prefer ground isolated power supplies because of the havoc I have seen. That being said my frame of reference is that of a manufacturer that has devices that are interconnected with different manufacturers... some manufacturers who has some really poorly designed interfaces. So in my case AC induced noise, ground currents, etc are a frequent occurrence and cause numerous issues.

Cheers,

Sam

Effect of transformer grounding on shielding

fxb80

s2sam

klrogers

crab master

wolfdog

Artrose

klrogers

crab master

fxb80

s2sam

s2sam

s2sam

crab master

wolfdog

Artrose

s2sam

crab master

s2sam

Top Contributors this Month

Recommended Communities