isolated grounding
- Thread starter kstanton5
- Start date
- Status
Pierre C Belarge
Senior Member
- Location
- Westchester County, New York
Ignorance. Either on the designer's lack of knowledge, or on my lack of knowing what the designer is thinking.
What are the receptacles supplying?
What are the receptacles supplying?
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
091122-1526 EST
Two sizes larger may not be enough. Maybe try 0000 copper.
The problem is where different pieces of equipment are spaced apart, maybe 20 ft or 200 ft, and the low level signal is referenced to equipment chassis (implying EGC), then any small current in the EGC becomes an error voltage added to the desired signal.
Suppose the audio signal peak is 1V and the dynamic range is 100,000/1, then a 10 microvolt drop on the EGC becomes a part of the background noise problem.
Rather than excessively large ground wires between the equipment a transformer might be a better solution to provide isolation. Another method is to use a balanced line driver and receiver to reduce common mode noise.
.
Two sizes larger may not be enough. Maybe try 0000 copper.
The problem is where different pieces of equipment are spaced apart, maybe 20 ft or 200 ft, and the low level signal is referenced to equipment chassis (implying EGC), then any small current in the EGC becomes an error voltage added to the desired signal.
Suppose the audio signal peak is 1V and the dynamic range is 100,000/1, then a 10 microvolt drop on the EGC becomes a part of the background noise problem.
Rather than excessively large ground wires between the equipment a transformer might be a better solution to provide isolation. Another method is to use a balanced line driver and receiver to reduce common mode noise.
.
ghostbuster
Senior Member
- Location
- Great White North
In our studio designs ,we will sometimes use flat insulated copper bar (1/8"X2") to electrically connect all the rack grounds together.It all depends on the studio performance equipment levels required.
It is complete ignorance on the designer/engineer part. The designer is tinking a larger conductor has a lower impedance than the smaller diameter wire, therefor reducing IR voltage drops along the length of the EGC. Well that thinking would be correct for low frequencies used for power and lighting, but makes very little difference to high frequencies.
There are two good solutions:
Beyond that for analog line level audio/video transport use balanced signals or optical.
There are two good solutions:
- A simple isolation transformer feeding dedicated circuits to the AV equipment.
- Balanced power which also uses a special isolation transformer defined under article 647.
Beyond that for analog line level audio/video transport use balanced signals or optical.
Last edited:
dbuckley
Senior Member
- Location
- Canterbury, New Zealand
If balanced audio isn't being used over any sort of distance then its amateur hour.
Video is more of an nuisance as it frequently is unbalanced signals, and frame refresh rates are similar enough to mains frequency and thus rolling bars can be a nuisance. But either digital video, or balanced analog video (usually over Cat5) solves that problem too.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
091022-2321 EST
It true that the impedance is of a ECG is frequency dependent. However, in many applications a larger EGC will reduce interference problems, and demonstrably does. Also may reduce probability of damage.
I favor isolation as the best solution.
My RS232 isolator will solve many problems. However, there are potential customers that will get, they think, adequate noise reduction with a large area EGC, or some other techniques that are not NEC compliant. Therefore, they do not buy isolation which would do a better job for them.
Noise is usually their biggest problem, and they do not seem to care about vulnerability to fault voltages from line to chassis shorts.
.
It true that the impedance is of a ECG is frequency dependent. However, in many applications a larger EGC will reduce interference problems, and demonstrably does. Also may reduce probability of damage.
I favor isolation as the best solution.
My RS232 isolator will solve many problems. However, there are potential customers that will get, they think, adequate noise reduction with a large area EGC, or some other techniques that are not NEC compliant. Therefore, they do not buy isolation which would do a better job for them.
Noise is usually their biggest problem, and they do not seem to care about vulnerability to fault voltages from line to chassis shorts.
.
Was common practice for years
Was common practice for years
This was common practice for years, the purpose being to lower the ground voltage differences between widely separated locations. This is the brute-force method, and is only moderately successful.
The isolated ground scheme is to avoid shared impedance, to avoid sharing a ground path with an electrically noisy, high leakage device not related to the AV system. In this it is very successful and should continue to be used.
Upsizing the ECG is a little more vain, when one looks at impedance vs frequency. Of far greater importance is to not generate voltage onto the EGC by the physical relationships between it and the CCCs. Loose singles in conduit is the worst, and twisting the CCCs together and running the isolated ground separately in the same conduit is best.There is a longer explanation in my post near the bottom of this thread
http://forums.mikeholt.com/showthread.php?t=119126
Was common practice for years
This was common practice for years, the purpose being to lower the ground voltage differences between widely separated locations. This is the brute-force method, and is only moderately successful.
The isolated ground scheme is to avoid shared impedance, to avoid sharing a ground path with an electrically noisy, high leakage device not related to the AV system. In this it is very successful and should continue to be used.
Upsizing the ECG is a little more vain, when one looks at impedance vs frequency. Of far greater importance is to not generate voltage onto the EGC by the physical relationships between it and the CCCs. Loose singles in conduit is the worst, and twisting the CCCs together and running the isolated ground separately in the same conduit is best.There is a longer explanation in my post near the bottom of this thread
http://forums.mikeholt.com/showthread.php?t=119126
The IG form of grounding is only intended to be used as a possible means of obtaining common-mode electrical-noise rejection on the circuit which it is used. It has no other purpose and its effects are variable. It cannot remove any noise from a circuit as it has no active or reactive components.
IG is not about what happens in the CCCs and can't affect noise on them. It is about leakage currents into the ground from devices not associated with the "sensitive" circuits. These leakage currents will return to the source via the ECG and any and all associated ground paths, including conduit, building steel, etc. Since these paths have finite impedance there will be voltage dropped along the way. The purpose of IG is to not have the ground path of the sensitive circuits share any of these impedance paths. Therefore a separate insulated grounding system is used that only connects to the G-N bond point, thus not sharing any of the "dirty" path. It is effective.
I understand that perfectly, I even wrote the section in the Emerald book that addresses IGR, but the very equipment affected and their L-N conductors are the cause/source of the of the leakage. It can certainly be compounded by daisy chaining with multiple drops, or running with other none sensitive equipment power and lighting circuits, but still the very equipment you are trying to isolate is guilty.
The leakage comes from one main component; the unbalanced architecture of the single phase 120 volt circuit with line being referenced to ground. This causes capacitive coupling of 60 Hz current into the EGC when ran with their circuit conductors. You cannot get around that, code will not allow it. The second factor lies in the equipment itself by placing MOV's and RFI device between L-G and N-G that also couple 60 Hz leakage into the EGC circuit. By using an isolated EGC only bonded only and the N-G bond instead of using the lower impedance ground plane of the conduit and building only mitigates the problem.
There are two excellent ways to prevent this.
One is using a completely different power architecture covered under NEC article 647 called 120/60 aka as balanced power systems used in recording studios and running standard dedicated SG circuits.
The second method used is just a good ole fashioned isolation transformer and dedicated SG circuits. Just your common dry types provide 90 db of cmr, and special purpose isolation transformers provide up to 160 db of cmr
However today it is becoming a moot point because IGR is antiquated because most all low level signal transmission is now being done with balanced signals, digital, and optical transmission methods. What is left can be taken care of with isolation devices and techniques like a transformer for anyone still operating in the stone ages using grounded referenced signals.
Last edited:
Common misconception. Do the math on the capacitive reactance and circuit impedance at 60 Hz and even at many KHz, and you'll see that wire to wire capacitance and even "filter" capacitors don't really enter the picture until you get up to RF.
I quite agree that the "protected" equipment is itself partly to blame for ground noise, mainly due to switch mode power supplies and such. However the mechanism by which it enters the ground is not just capacitive, but inductive coupling from the CCCs, as I've stated earlier.
The isolated EGC is more susceptible to ground voltage induction if the CCCs aren't oriented properly with respect to the EGC. The coaxial nature of conduit is both less susceptible to ground voltage induction and lower impedance when all fittings and connections are new and tight. However I would argue that over the lifetime of the system, the copper ground wire will maintain a better ground connection.
I agree about using an isolation transformer, and I spec them into quite a few of my systems. Balanced power, however, is a solution in search of a problem. When applied into a situation with many other problems it can sometimes mask them enough to present an impressive improvement. It's proponents seem to miss the fact that it is inherently an isolation transformer and would offer the same improvements if the output were unbalanced instead of balanced. Couple that with the GFI and other requirements, makes it less than ideal. It comes back again to the fact that it's really not the capacitive imbalance that's the problem.
I quite agree that in the future these issues will be minimized because of digital and optical signal transfer. However today it is still an issue and it's worse than a decade ago because high harmonic loads are everywhere. Wall Warts are all becoming switchers now, CFLs are particularly nasty, dimmers are everywhere and built way too cheap. It's all about current waveform risetime, and those big heavy coils of wire that slow things down are the first victims of cost cutting. So ground noise today is a lot more of high narrow spikes than it was years ago. Since electronically balanced signal interfaces reduce their usable signal headroom by whatever (peak) voltage they need to cancel, these ground noises are more troublesome than in the past.
And here I thought I was going to try to stay out of IG arguments.:roll:
ghostbuster
Senior Member
- Location
- Great White North
I agree with your idea of having the ground conductor "seeing" equally the magnetic field from both current carrying conductors.(the induced voltage on the ground conductor will be kept to a minimum)
On larger ampacity runs with single 3 phase insulated condutor runs (coreflex) we have just measured 16 amps on the ground wire compared to 100 amps on the phase conductors.This has been installed as per code.
In large studios,we have measured the ground current as equipment is gradually turned on.For test purposes the ground conductors were seperated by approx. 100 feet from the main current carrying conductors(as a result virtually zero magnetic field induction on this ground conductor).The current measured on this ground directly 100% correlated with the capacitive leakage current from each device located in the racks.
- Status