Isolation Transformer Grounding

[AlWired]

I want to install an isolation transformer in a test laboratory to improve the performance of a sensitive piece of electronic equipment. One of the concerns is that the ground also be isolated. Is this the way an isolation transformer is constructed? If so, how is the ground re-established on the secondary side of the transformer? This transformer is not in a hospital, so it seems that Article 517 would not apply. Should the secondary be grounded as a Separately Derived Source per Article 250.20(D)? If so, how should the grounding be done? Doesn't seem like grounding to building steel would apply for this application. This isolation transformer is about the size of a shoe box, will be plugged into a receptacle and will be sitting on the floor near the equipment to be protected.

Any advice will be greatly appreciated.

 

[gar]

080627-0552 EST

What is the sensitive electronic equipment?
How is it known that an "isolation transformer" will solve whatever problem exists?
What is the nature of the problem to be solved?
What is their or your definition of an isolation transformer?
To what is the equipment sensitive?

Most normal non-autotransformers have a high DC resistance between primary and secondary, but no capacitive shield (Faraday shield).
Most normal transformers do not eliminate harmonics or noise that exist in the primary waveform (normal mode noise), but do reduce input common mode noise.
http://en.wikipedia.org/wiki/Faraday_cage
http://blackmagic.com/ses/bruceg/EMC/isolatrans.html

I do not know that code has much to do with your problem. You may not be able to adhere to code and solve the problem. For example: I have an oscilloscope and I want to monitor the difference voltage between two signals that are referenced to earth ground and using the scope difference amplifier mode is unsatisfactory. Then I need to float the scope chassis from ground.

If earth ground voltages/currents are a problem to this equipment, then at a minimum you may want a normal 1 to 1 transformer and on the secondary tie one line to a grounding point relative to the equipment's signal input.

Based on my above questions at the beginning you need to get more information.

If input voltage regulation is a problem consider a Sola constant voltage transformer of the low harmonic type, a battery supplied inverter, a motor-generator, or other regulated source.

If something has to be floated, then advise the customer of the safety problems.

.

 

[jim dungar]

Should the secondary be grounded as a Separately Derived Source per Article 250.20(D)?

Yes.

But, without knowing what problem you are trying to solve, it is hard to give you direction. What do you want this transformer to do for you?

 

[ike5547]

The isolation transformer is not intended to isolate the ground so it is grounded and bonded as usual. It's purpose is to isolate the primary and secondary systems from transmitting noise.

 

[gar]

080627-2119 EST

ike:

I see nothing in the question that indicates what problem is to be solved. What does ground mean in the context of the question? As I previously said an isolation transformer primarily reduces incoming common mode noise. Other noise passes right thru the transformer. Connecting the secondary to the incoming EGC does nothing to solve EGC path noise problems for interconnected equipment on each side of the transformer.

If the isolation transformer is just outside a screen room the transformer shield might connect to the incoming EGC, and the secondary might connect to the screen room shield.

What needs to be done depends upon the problem. Maybe the transformer is not needed. Maybe just a filter, or electrical isolation between different pieces of equipment.

.

 

[don_resqcapt19]

080627-2119 EST

Connecting the secondary to the incoming EGC does nothing to solve EGC path noise problems for interconnected equipment on each side of the transformer.

If the secondary voltage is one that is required to be grounded by 250.20, then there is no code compliant way to hook-up the transformer without a connection between the primary and secondary EGCs (with some rare exceptions, for example the isolated systems permitted in Article 517).

 

[gar]

080628-1228 EST

don:

In a research type application you may not pay any attention to code if you are trying to accomplish something such that following code would produce incorrect results.

However, that is not the point.

The important question is why is an isolation transformer needed, what is the problem that needs a solution, and will an isolation transformer solve that problem? We also do not know what the end user believes an "isolation transformer" is.

What is the end user's definition of ground and how does this relate to the problem?

.

 

[jim dungar]

In a research type application you may not pay any attention to code if you are trying to accomplish something such that following code would produce incorrect results.

Would you please provide a citation that supports this statement. Are you saying that it is acceptable to have unsafe installations as long as they are "reasearch labs" and not part of the test itself?

If the installation is subject to a code then that code must be followed. The NEC applies to installations of premise wiring systems. Is the use of this transfromer part of the premise's wiring?

 

[dereckbc]

This isolation transformer is about the size of a shoe box, will be plugged into a receptacle and will be sitting on the floor near the equipment to be protected.

If it is plugged into a wall receptacle there is no requirement for it to be grounded to anything as it uses the EGC to ground the secondary and not covered by the NEC. I suspect this is something like a Triplette isolation transformer with a copper static shield to eliminate Common Mode Noise.

If you plug the tranformer into a recept, and the equipment plugs into the transformer recepts; call it good, safe, and done .

 

[iwire]

Dereck how does that jive with 250.114?

Make sure you look at it's exception to (4). :smile:

 

[dereckbc]

Bob I assume from the description the OP gave this is a plug connected isolation transformer that has the output recepts built into it. Just like the ones you would see people use on high-end AV equipment. If in fact that is what it is, and it sounds like it, the EGC is grounding the case and secondary XO.

I assume it is something like this from Tripp-Lite

If it is, then there are no requirements for any additional grounding and is out of scope of the NEC IMO. Or put another way is not part of the premises wiring

 

[iwire]

Dereck we will have to remain in disagreement.

250.114 directly address's cord and plug connected equipment. I don't see a break for cord and plug connected equipment that is supplied through another piece of cord and plug connected equipment.

If a cord and plug connected appliance is found in the lists of 250.114 the NEC applies.

I also suspect that every piece of UL listed equipment that is manufactured with a grounding type plug is in fact somewhere to be found in 250.114.

For instance 250.114 makes mention of information technology equipment. Here is just a partial list of what UL calls information technology equipment.

Office appliances: Adding machines, bursters, calculators, collators, dictation
and transcribing machines, electric typewriters, erasers, folding,
embossing and sealing machines, label printers, microfilm readers, motoroperated
file cabinets, overhead projectors, paper cutters, paper shredders,
pencil sharpeners, sorters, stackers, staplers.

Printers/Reproduction equipment: Copiers, duplicating machines, microfilm
printers, mimeograph machines, plotters, printers.
Mailing, banking and currency-handling equipment: Cash registers, coin
counters, feeders and dispensers, accounting machines, check-writing-,
-assigning, -dating, -inserting, -mailing, -numbering and -stamping
machines, point-of-sale terminals.

Multimedia equipment/accessories: Cable modems, digital cameras, DLP
projectors, LCD projectors, microphones, set top boxes, speakers, video conferencing
systems.

Network equipment: Baluns, bridges, fiber optic transceivers, hubs,
nodes, Power over Ethernet (PoE) equipment (e.g., power source equipment
[PSE] and powered devices [PD]), repeaters, routers, switches, transceivers.
Wireless (RF, infrared) transceiving equipment: RF modems, hand-held
computers with integral transceivers

I am sure a piece of lab testing equipment is considered some sort of electric tool.

 

[dereckbc]

Dereck we will have to remain in disagreement.

OK :grin: But I fail to see how the transformer is not properly grounded if it is using a 3-wire plug with ground and its only interface with premises wiring is via the plug.

 

[iwire]

OK :grin: But I fail to see how the transformer is not properly grounded if it is using a 3-wire plug with ground and its only interface with premises wiring is via the plug.

The transformer may well be grounded.

The equipment supplied from it will also very likely be required to be grounded.

The exception to grounding I pointed out references the use of an isolation transformer but says the ungrounded secondary must operate not over 50 volts.



FWIW I understand there is often a difference between what the NEC requires and what is really happening.:smile:

 

[dereckbc]

Bob , I am not going dispute your interpetation. I have something simular to this in my home. So my last cooment to the OP is to consult the manufactures instructions assuming again it is UL listed. If the instruction require and external ground connection, then bond it to the nearest building steel to comply with NEC 250.30. However if it is optional as the unit I have is, then leave it alone as there is no requirement. Anyway that is all I got.

Here is the unit I have a 3RQ. Scroll down to the bottom for the backplane connections.

 

[iwire]

So on that unit, if I was to run a continuity check from it's output receptacles grounding terminals and the premise EGC there will be no connection?

If the answer is no continuity to the premise EGC then IMO you can not connect equipment to that unit that is covered in 250.114. :smile:

 

[dereckbc]

Bob there is continuity from the EGC plug to the recepts ground, same for the Tripp-Lite transformers I linked too. Only exception might be the Medical ones from Tripp-Lite

 

[gar]

080628-1843 EST

To several of your comments:

Look at the original question and part of the question was isolation of the ground. Truely I do not know what they want but I can guess. I will assume this was the reason for the consideration of an isolation transformer.

One can work safely and not adhere to the code. Just adhering to the code does not mean you will work safely. For many years I worked with electrical and electronic devices that had no ground connection. There were no 3 prong plugs. I still have some of this equipment, a Simpson VTVM for example. I have used 1 KW HF transmitters with no connection of any kind to ground. The 3 prong plug was a mid 1950s creation. And in fact a problem at times. It is not uncommon to isolate electronic equipment from the EGC whether it meets code or not. The person doing this really does not care if it is code compliant. Generally I would say this would be for short time periods.

I am not saying that if you do research that you ignore the code. What I am saying is that if there is no other way to solve your problem you may have to ignore parts of the code.

If ground path noise is the major problem to be solved, then maybe the best solution is to go battery powered and completely break any sort of ground path or conducted noise from the AC line.

On my primary interest, RS232 communication between different pieces of equipment, mostly computer to CNC, this subject of ground path noise is of great importance. We make a product that provides electrical isolation to +/-2000 V between the two ends, has baud rate capability to 115.2 kbaud at up to 4000 ft. Our system solves problems and allows users to adhere to the NEC requirements.

I find many people making suggestions to users in the CNC area that total violate the NEC and good safety rules and I constant point this out and the associated problems.

One bad common suggestion is to isolate the computer from the EGC connection. This means the computer now derives its reference to ground from the CNC and however it might be grounded thru a #22 wire of maybe 100 or more feet. What happens to the chassis of the computer if someone shorts a 240 hot wire to the CNC chassis.

Another bad suggestion is to disconnect the CNC from the EGC to it and in the place of the EGC connect the machine chassis to a local ground rod at the machine. This is an extremely bad suggestion. On a hot to chassis short this will almost certainly never trip the breaker. I illustrate the problem with a calculation that illustrates why, and this is obviously the reason that the NEC is written as it is relative to the EGC.

If I really had to measure the current in a hot 120 V wire with a wideband oscilloscope and a shunt (nothing else would do), then with great caution I would try floating the scope off of ground and probably at least one isolation transformer to supply power. Obviously the scope would be insulated and one would not touch it while the hot line was hot. A battery power scope would be better, but it still has the same safety problems.

.

 

[iwire]

I am not saying that if you do research that you ignore the code. What I am saying is that if there is no other way to solve your problem you may have to ignore parts of the code.

So what your saying is if it is difficult to follow the code then don't. :grin:

Spoken like a true DIY.

I fully understand that almost anyone with a test bench breaks code rules on a regular basis.

As 70E becomes more prevalent that will stop for people that are not self employed or hobbyists.

Basically your employer will have to put a stop to unsafe work practices or face fines. In some companies this is already happening due to corporate policies that cover all employees not just the ones on the production floor.

 

[dereckbc]

If ground path noise is the major problem to be solved, then maybe the best solution is to go battery powered and completely break any sort of ground path or conducted noise from the AC line.

On my primary interest, RS232 communication between different pieces of equipment, mostly computer to CNC, this subject of ground path noise is of great importance. We make a product that provides electrical isolation to +/-2000 V between the two ends, has baud rate capability to 115.2 kbaud at up to 4000 ft. Our system solves problems and allows users to adhere to the NEC requirements. .

And the problem is very easy to fix in a code compliant manner. I work in the telecom field and we have been using a optical isolation RS-232 for more than 25 years, so there is nothing new or special, just knowledge of how to deal with the problem.

In my career I have been an RF engineer, RF Technician, before becommin a liscensed engineer, and I hold a HAM liscense and very familar with oscilloscopes and isolation techniques. There is no reason to ever float a scope. It is a matter of using the right equipment for the job. One i am sure you are aware of is simple using a differential volatage probe that will work on any make or model oscope. Another approach is to use a scope made for the job like a tektronic THS720P with the appropiate test probes that allow for dual input channels, individually isolated from the oscilloscope's chassis as well as from each other. You can measure dual-trace waveform comparisons to be made, with each of the two channels referenced to its own common.

As for making current measurements with a shunt can be done as above, but again the right tool for the job in this case is a simple Hall-Effect Clamp-On probe.

As a young and inexperienced engineering technician I admit to bending the rules because I thought I knew everthing. Well I learned better from mistakes, experience, and some fried equipment. There is always a code compliant way to get things done. I have dedicated my life to power quality, special purpose grounding systems, and there is no reason not to be able to run, power, operate, test, or repair anything that is safe and code compliant manner. You just have to learn and understand what you are dealing with, and when you don't know or have the answer, someone does, you just got to search and be open minded...

Dereck C, PE