Talk:Ground loop (electricity)
This article is rated Start-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||
|
Archives (Index) |
This page is archived by ClueBot III.
|
Balance and differential
[edit]In this edit the section title was changed from "Balanced line" to "Differential signaling". In fact, the description of the advantages of the scheme (good common-mode rejection) is exactly the advantages of using balanced lines. Differential signalling helps neither CMR nor ground loop noise. If the line is not balanced and the more earthy leg of the line has ground loops induced then the receive end will pick this up since it is not common mode. SpinningSpark 17:06, 31 July 2021 (UTC)
- Spinningspark, thanks for reverting that. ~Kvng (talk) 14:04, 9 August 2021 (UTC)
Seemingly incorrect solutions and description of ground loops
[edit]Some of the descriptions of ground loops caused by induced emfs, and descriptions of mitigations for them, seem to contradict analysis and even experiments that I have done. See this post on stackexchange where I outline my analysis of the ground loop scenario. In the answer, I describe my experimental results. A brief summary is this:
The currently available Wikipedia article indicates that the problem with ground loops due to linked magnetic flux is the induced currents. This is not true. The issue is the induced voltage difference between the ground inputs to the source and destination amplifiers. While increasing the resistance of the signal shield will reduce the ground loop current, the increased resistance will more than compensate, and the voltage drop will actually go up (since a larger share of the induced loop voltage occurs across that larger resistance). Disconnecting the ground shield will just put that big loop (which is like the secondary of a transformer) in series with the signal. If you disconnect the ground shield, there is still a ground loop and induced currents will still flow, except only through the signal wire, rather than also through the ground shield. Having a low impedance ground shield however, will short this big loop out, decreasing the voltage drop between the two grounds and lessen induced noise seen at the load. So it seems to me like the reasoning for this kind of ground loop is incorrect, as well as the mitigations for it.
This article is also a little bit confusing in the ways that it conflates different kinds of ground loops. The section "Representative Circuit" describes a completely different kind of situation as described in the part about induced emfs. In the section "Representative Circuit", the issue has nothing to do with linked flux and the problem is because the resistance in ground is too high, causing voltage drops between device grounds in the mains wiring. I agree with this description and analysis (although I think I can improve the example circuit). However, in the former case, increased resistance in ground between the two devices causes them to couple together due to current flow through ground. In the case of induced emfs, increased resistance in the ground loop actually decreases the induced noise seen at the load (see my analysis). So for the resistance through the mains wiring, there are actually competing factors. However, it seems that in all cases, reducing the impedance of the ground shield of the signal cable connecting the two devices should reduce noise in that cable.
I can edit the article to fix these issues, but I am hesitant to do so since the internet seems to disagree on how to mitigate ground loops due to magnetic flux linking the loop (although my own experiments and analysis confirm it). This includes sources cited by the article, which seem incorrect to me. Does anyone here have any thoughts on my analysis? Can you find any errors in it? (I would be surprised, since it seems intuitively obvious.) And why the internet seems to think breaking the ground shield should reduce hum (when it actually increases it)?
Hddharvey (talk) 11:08, 12 September 2021 (UTC)
- "Put a small resistor of about 10 Ω in the cable shield conductor, at the load end. This is large enough to reduce magnetic-field-induced currents but small enough to keep the component grounded if the other ground path is removed. In high-frequency systems this solution leads to impedance mismatch and leakage of the signal onto the shield, where it can radiate to create RFI, or, symmetrically through the same mechanism, external signals or noise can be received by the shield and mixed into the desired signal."
- This is not the recommended solution. When you connect a resistor like this, the ground loop current flowing through the resistor places an additional noise voltage in series with the signal source, degrading the signal to noise ration. Placing a resistor in series with the signal return path has to be done in a very specific way in order not to degrade the SNR but to reduce the effects of cross channel ground loops (a separate type of ground loop and not considered the same as a classic AC ground loop) and is called a 'hum breaking resistor' or 'HBR'. Cross channel ground loops arise when mains related noise is induced into the ground connection between the two inputs inside an amplifier and then flows out to the source and back again. again, details in the presentation I linked to earlier. Harrison Thomas 14:22, 6 December 2022 (UTC) — Preceding unsigned comment added by Harison thomas (talk • contribs)
- Please see this presentation here https://hifisonix.com/wordpress/wp-content/uploads/2019/02/Ground-Loops.pdf
- I think the article could be more structured, starting with the different forms of 'ground loop' - classic AC power ground loop, cross channel ground loop, high frequency noise loops (typically a problem in double insulated consumer and professional gear) . Each of these has a specific set of solutions and all rest on a clear understanding of magnetic induction (major cause of ground loops), electrostatic coupling (aka capacitive coupling and an issue with HF noise sources like SMPS). There could also be an in-depth discussion on equipment layout etc to reduce these problems - again, see the linked to presentation.
- A discussion on unbalanced vs balanced equipment interconnects could cover the 'pin 1 problem'
- )
- Harrison Thomas 14:14, 6 December 2022 (UTC) — Preceding unsigned comment added by Harison thomas (talk • contribs)
Also note that in a linear circuit consisting of passive impedances, voltage sources, and current sources, increasing a resistance will always increase the magnitude of the voltage drop across that resistor. This can be verified by imagining taking that resistor out of the circuit and replacing the resulting two-port network with its Thevenin equivalent circuit. Putting the resistor back in reveals that it is simply in a voltage divider. This is why it is misleading for the article to talk about how increasing the ground shield resistance will reduce the current and thus reduce the noise. Yes, the current does reduce, but the voltage across it does not reduce!
Hddharvey (talk) 11:25, 12 September 2021 (UTC)
- "Representative circuit[edit source]
- Simplified circuit illustrating a ground loop.
- The circuit diagram illustrates a simple ground loop. Circuit 1 (left) and circuit 2 (right) share a common path to ground of resistance . Ideally, this ground conductor would have no resistance (), yielding no voltage drop across it (), keeping the connection point between the circuits at a constant ground potential. In that case, the output of circuit 2 is simply . . . . "
- What you are showing in your diagram is 'common impedance coupling' and not a ground loop. Its very important you separate out the different types of noise loop. You could then link out to a separate article on common impedance coupling. Harrison Thomas 14:56, 6 December 2022 (UTC) — Preceding unsigned comment added by Harison thomas (talk • contribs)
My suspicion is this: ground lifts do work, except not when applied to the signal cable of interest. They need to be applied to the mains connection of one of the devices (which potentially has safety issues). They could also work when applied signal wires if a different signal cable was inadvertently grounding a device that would otherwise be isolated from earth. This signal cable would cause other signal cables leading into or out of that device to form ground loops with other earthed devices. However, a ground lift will not stop noise from appearing on the signal cable it is applied to. This means that ground lifts on signal cables can sometimes work, but only indirectly. However, maybe a misconception has spread on the internet about why these ground lifts actually work. I've been searching for videos on YouTube about people successfully using ground lifts, and I haven't been able to find any yet (but I have found videos where people found ground lifts did not help).
Hddharvey (talk) 12:21, 12 September 2021 (UTC)
- I agree that the article is a mess. I would support giving you the latitude to substantially rewrite it, as long as you stick to Wikipedia principles. If there is unsourced handwaving, consider removing it. Even the first sentence is incorrect. A ground loop exists when a wire connects between two different ground points. The loop exists whether there is a potential difference between those points or not. There is usually not a problem when there is no potential difference. The physical loop, the circumstances under which the loop is a problem, and the nature of those problems are three distinct concepts.Constant314 (talk) 18:02, 12 September 2021 (UTC)
- Thanks, I'm considering the kinds of changes I could make. It's a bit hard to get sources for this (but I'll have a look) since there seem to be multiple misconceptions that have spread about the ways that ground lifts work and the kinds of problems they are able to solve (which means that I'd just be explicitly searching for sources that agree with me). The topic of ground loops is a funny topic, because it's probably more talked about by people with a more limited understanding of physics and electrical engineering, due to its prevalence in audio applications. Hddharvey (talk) 05:36, 13 September 2021 (UTC)
- Worst case scenario, the incorrect stuff such as the claim:
- "disconnecting one end of the ground shield of an unbalanced line will reduce/eliminate ground hum across that line"
- can be removed. I could replace that with correct facts such as
- "minimizing the impedance of the ground shield of an unbalanced line will minimize the voltage difference across the ground shield and thus the ground hum" (I'd almost go as far as to say that this is intuitively obvious in general, with the exception of exotic active circuits that can create effective negative resistance)
- or
- "balanced lines with ground lifts can be used to break ground loops in other unbalanced lines without introducing noise (since the ground hum is common mode and cancelled at the receiver of the balanced line)"
- if I can find sources for them (although it could be supported by analysis, but I'm sure plonking a bunch of derivations into an expandable box would not be considered a substitute for a citation). The last claim above is probably the reason the internet has this misconception about the efficacy of ground lifts for signal cables - it appears that they would be useful in balanced cables to prevent ground hum in other signal cables, since the noise voltage across the ground lift will be cancelled at the differential receiver of the balanced line. A ground lift on any unbalanced line would remove ground hum across other cables in the loop only by taking all of the hum for itself. Hddharvey (talk) 05:57, 13 September 2021 (UTC)
- Worst case scenario, the incorrect stuff such as the claim:
- As far as circuit analysis of the "induced ground loop" case goes, there's an argument to avoid the term "breaking a ground loop". You never really break the ground loop. The loop in space always exists. Faraday's law of induction does not care whether there is a break in the conductors that form the loop. The same induced electric fields will be circulating around that loop whether you break it or not. All that happens when you break the loop is that the electric field redistributes itself so that all of the "voltage" drop from the noise occurs across the gap. This means that "breaking" the loop could be a good thing if it causes all the noise to be concentrated across a signal line that is able to reject it (such as a balanced line with a differential receiver), however it would be a bad thing if it caused the noise to be concentrated across an unbalanced line with a single-ended receiver (which is where the current article gets it wrong). Hddharvey (talk) 06:14, 13 September 2021 (UTC)
- Ground noise and grounding problems can be incredibly difficult to fix and diagnose. Sometimes it is the potential difference even when the currents are nanoamps. Sometimes it’s the current even when the potential difference is millivolts. I’ve seen cases where breaking the shield fixes the problem. I’ve seen cases where adding an extra low impedance path fixes the problem. There is no general fix. Sometimes the problem isn’t the potential difference in ground at the inputs; it is caused by the current entering the chassis and getting into some other circuit. But I think the principle that we should apply is WP:NOTHOWTO. It is not the job of Wikipedia to prescribe how to fix ground loop related problems. I would remove all the stuff about remediation but leave in the design stuff like differential signaling.Constant314 (talk) 06:17, 13 September 2021 (UTC)
- "Sometimes it’s the current even when the potential difference is millivolts." - Out of curiosity, what does this exactly refer to? Regarding the general complexity of diagnosing and solving ground hum - noted. Especially since there are multiple kinds of ground loops that can be present in the same system. Hddharvey (talk) 06:52, 13 September 2021 (UTC)
- Techniques for avoiding or mitigating ground loops are an important part of the subject so I would not gut this as Constant314 is suggesting. I don't think the article is wrong in saying that the induced current is the source of the problem. The induced current do go on to cause secondary issues such as IR voltage and secondary induction that may be more relevant to system performance. To my knowledge, what Hddharvey says is not wrong but it needs to be backed by reliable sources, not original research. ~Kvng (talk) 16:02, 16 September 2021 (UTC)
- I agree that it would be nice to mention techniques for mitigation in the article, otherwise you're ignoring the elephant in the room. Ground lifts do not necessarily have to be described in full detail, since there is a "Main article" for that: Ground lift (and maybe that article could be improved too if there are good sources out there, but that's a separate issue). How much can be said would be dependent on being able to find good sources. However we would also have to ignore a lot of the sources out there, and that would be on the basis of "original research" (some analysis that I did and a single experiment that confirmed it). My main conclusion is that ground lifts on unbalanced lines don't really make sense (for the reduction of noise voltage, but there are other issues, like current ratings). They only make sense on balanced lines because they redistribute the noise voltage drop to the balanced line which is able to reject it.
- My interpretation of WP:NOTHOWTO is not that procedures to do things cannot be described in Wikipedia, it's just that the way that those procedures are described should not be in the format of instructions or a manual and should try to use grammar that is descriptive not prescriptive.
- Regarding your discussion about current being the problem due to IR drops. In *most* circuits, the IR drop goes up not down when you increase the resistance, despite the fact that current goes down (excluding circuits with dependent sources or more exotic nonlinear/active elements that allow real-valued negative resistances to be constructed). This is why I don't like to say current is the problem (for noise at least - current ratings being exceeded for weak ground wires intended for signals is another possible issue). The noise can get *worse* even when the ground loop current goes down, e.g., if you lift one end of the ground shield of an unbalanced line that would otherwise be part of a ground loop or if you insert a resistor in series with the shield (recommended by many sources on the internet and in the current article, and the diagrams imply the line is unbalanced). This applies similarly to the self-inductance of the line.
- Regarding effects of inductive coupling of the ground loop to other things due to induced ground loop currents, I agree that it could theoretically be an issue. However it's not clear that it would be a significant factor in normal circumstances because other parts of the system should have negligible loop areas compared to the ground loop itself. And it's also not obvious that the induced fields due to ground loop currents wouldn't actually work to reduce the noise coupled into other parts of the system. From a conservation of energy standpoint, ground currents should be taking energy out of the environmental electromagnetic field and converting it to heat.
- I am OK with mentioning fixes and causes, as long as the article is written in way that does not suggest that the list of fixes is exclusive or that the analysis applies in all cases.Constant314 (talk) 16:05, 18 September 2021 (UTC)
Draft rewrite
[edit]@Constant314: @Kvng: @Chetvorno: FYI anyone who is interested - I have a draft of a rewrite of this page that I have had sitting around for months now. I only recently put it into a sandbox page and tried to add some citations to it. If anyone is interested in improving this page, hopefully this can provide a good starting point (User:Hddharvey/GroundLoopDraft). I also have created some updated diagrams there:
- The diagram showing the magnetic fields producing current - I changed this to show electric fields too. Hopefully this makes it clearer why current starts flowing, and that the electric force will still be there even if no current can flow.
- There's a diagram illustrating how noise can behave like a parasitic voltage source. This prevents readers from reaching dodgy conclusions about how current flow is the only source of the problem. Don't get me wrong, current isn't not a problem - it's just clearly more complicated than "current flowing creates ground hum" when it's possible to have the current reduced to zero and yet the hum go up.
- There's a diagram illustrating common impedance coupling which is inspired by the existing "Representative circuit" in the article atm. Have had trouble finding good sources discussing this though.
- There's some diagrams that illustrate correct scenarios where you would actually want a ground lift.
The draft is not perfect (needs more citations and some sections need a little content). However, I do not believe it is far off from being an improvement over the current article, which has many issues including:
- Long extended discussions with no citation.
- The citation to the midimagic website is unreliable and its core points and reasoning are incorrect.
- While not completely wrong, the reasoning about "it's because of the current" is misleading and leads to wrong conclusions, such as the idea that open circuiting a cable will reduce the noise across it. (It's possible that it could reduce the effect of noise in the system overall, but it will not reduce the noise added across the cable - I discuss this in my stack exchange answer here.)
In my opinion large sections of the article should be heavily trimmed or outright removed. I might remove some of these dodgy parts myself. If anyone wants to improve the article beyond that using my draft as a starting point, feel free. At the moment, I don't have the time or motivation to work on this article much further myself. If anyone wants to improve this article, I'm sure that my diagrams will be useful at the very least (credits to Chetvorno who I believe created the original set of diagrams that many of mine are based off). Hddharvey (talk) 08:21, 1 November 2021 (UTC)
- I have reverted your most recent edit. Readers and future editors need to be able to see where we got the information from. If you think the source is not up to scratch, then you can mark it as unreliable. But if the text is left in place, then that (unreliable) source remains the source for it and everyone should know that.
- The particular thing you seem to have found dubious (as you marked it so) is the practice of grounding shield conductors at one end only. Now I know for a fact that this was established practice in major broadcasting organisations when I worked in the industry for the shields of audio twisted pair cable. I have heard that the practice has fallen out of favour in some places since I retired, but there are still many fairly recent published books verifying the fact [1][2][3]. This book [4] makes the point that this is only advantageous at audio and RF requires multipoint grounding to be effective. SpinningSpark 09:06, 1 November 2021 (UTC)
- I agree. The issue is:
- The article's examples imply this is true for single-ended lines/unbalanced lines too (it's unlikely but theoretically possible that open-circuiting a ground shield on a single-ended line could improve the noise, but not beyond a certain point - see here, also see my draft where I provide an example of breaking the ground shield being a good thing).
- The article implies that "if you stop the current, you fix the issue". This is not remotely true.
- I had also contacted the author of the source I deleted. There was some back and forth, but he ended up saying that the topic was beyond his understanding. The only source he was able to provide that somewhat implied disconnecting part of a single-ended line could be a could idea was this one. Maybe marking those statements dubious and deleting the source was not the most appropriate way to indicate the errors in the article, but that is all I currently have time for. Hddharvey (talk) 09:46, 1 November 2021 (UTC)
- A solution I have come across for analogue video coax cables is stop coils (that probably should not be a redlink). I'm talking about cables that have runs of miles rather than feet across a studio floor. Such cables suffer from mains hum induced in the shield. Although a full video with sound sub-carrier occupies a 6 MHz bandwidth it also goes down to very low frequencies. A high pass filter will not do the trick because that is going to block the frame pulses as well and cause issues on fade to black scenes for instance. Stop coils are inductors placed in series with the shield to present a high impedance to the mains hum. Essentially, it is a ground lift at low frequency but not at high frequency. SpinningSpark 10:57, 1 November 2021 (UTC)
- Here's a source [5] for that (just in case I imagined it which ins't out of the question at my age). This is relevant as well, but a very old source and only talking about test procedures. SpinningSpark 11:09, 1 November 2021 (UTC)
- What I've been able to find on these 'stop coils' (and including diagrams in the sources you provided) seem to imply that they're more like a common-mode choke than an inductor that is purely in series with only one conductor. It would surprise me if the latter was used on unbalanced cables. As I've said above, increasing impedance (at some frequency) will - for any "reasonable" circuit - increase the voltage drop while decreasing the current. It's conceivable that this could improve things overall if the current caused bad problems elsewhere, but it's still not a very good solution since it does nothing to reject noise induced across the cable. Hddharvey (talk) 11:27, 1 November 2021 (UTC)
- I don't know what diagram you are looking at, I'm not seeing anything that supports what you say, the only relevant diagram included shows just the opposite. Stop coils are not the same as common-mode chokes. SpinningSpark 11:56, 1 November 2021 (UTC)
- After reading this I'm not so convinced I'm right. That source calls them a trasnformer which ties in with common-mode choke. Since they are still available for sale I guess we could ask a manufacturer for a circuit diagram. SpinningSpark 12:09, 1 November 2021 (UTC)
- The diagram you cited shows the stop coil as a two-port device. Anything could be in the box. I may have spoken too strongly when I said your source implied it was more like a common-mode choke. I should have said: The source you provided gave no indication as to the nature of this coil, and mentions of 'longitudinal stop coil' on the internet implied that it was a device applied to both lines, such as an isolation transformer or a common-mode choke rather than just an impedance in series with the ground shield only. I saw that vendor too, the datasheet seemed to imply that the 'hum eliminator' was just a common-mode choke. It said: "to the unwanted ripple, it appears to be a choke and hence provides attenuation". Hddharvey (talk) 12:19, 1 November 2021 (UTC)
- I'll add that this 'hum eliminator' that I mentioned was listed here under the category "longitudinal stop coils" - so it's not just some random products. That also appears to list isolation transformers. Hddharvey (talk) 12:21, 1 November 2021 (UTC)
- My interpretation of the diagram (admittedly very ambiguous) was that the stroke across the bottom between the two earth connections implied a break in the ground continuity, but not in the signal conductor. SpinningSpark 12:34, 1 November 2021 (UTC)
- I thought that at first too. But I think it makes more sense that the label just ran over the top of the dotted line and is referring to the two-port box (see the line linking the two). Hddharvey (talk) 12:40, 1 November 2021 (UTC)
- In fact this is pretty clearly the case if you scroll back a few pages and look at similar diagrams, where the label is directly on top of the two-port box. Hddharvey (talk) 12:41, 1 November 2021 (UTC)
- My interpretation of the diagram (admittedly very ambiguous) was that the stroke across the bottom between the two earth connections implied a break in the ground continuity, but not in the signal conductor. SpinningSpark 12:34, 1 November 2021 (UTC)
- What I've been able to find on these 'stop coils' (and including diagrams in the sources you provided) seem to imply that they're more like a common-mode choke than an inductor that is purely in series with only one conductor. It would surprise me if the latter was used on unbalanced cables. As I've said above, increasing impedance (at some frequency) will - for any "reasonable" circuit - increase the voltage drop while decreasing the current. It's conceivable that this could improve things overall if the current caused bad problems elsewhere, but it's still not a very good solution since it does nothing to reject noise induced across the cable. Hddharvey (talk) 11:27, 1 November 2021 (UTC)
- Here's a source [5] for that (just in case I imagined it which ins't out of the question at my age). This is relevant as well, but a very old source and only talking about test procedures. SpinningSpark 11:09, 1 November 2021 (UTC)
- A solution I have come across for analogue video coax cables is stop coils (that probably should not be a redlink). I'm talking about cables that have runs of miles rather than feet across a studio floor. Such cables suffer from mains hum induced in the shield. Although a full video with sound sub-carrier occupies a 6 MHz bandwidth it also goes down to very low frequencies. A high pass filter will not do the trick because that is going to block the frame pulses as well and cause issues on fade to black scenes for instance. Stop coils are inductors placed in series with the shield to present a high impedance to the mains hum. Essentially, it is a ground lift at low frequency but not at high frequency. SpinningSpark 10:57, 1 November 2021 (UTC)
- I agree. The issue is:
Just adding my two cents worth. Breaking the shield is not dubious. It used to be the first thing that we tried because it was easy. It was common knowledge. Sometimes it works and sometimes it does not. There are two (maybe more) problems. 1. The difference in ground voltage adds to the signal. Breaking the shield does not help this problem. 2. The shield provides a low impedance path between the two points that allows high current to flow between them. Breaking the shield does mitigate this problem. Although it has no effect on the additive voltage noise at the input, the current flowing from the equipment chassis at point A into the equipment chassis at point B is causing a problem. The problem can even be in an apparently unrelated piece of equipment in the same rack. The ground current is going somewhere that it didn't used to go and that can cause a problem. Constant314 (talk) 17:03, 1 November 2021 (UTC)
- As I've said, breaking the shield of a balanced line makes perfect sense and it's possible that breaking the shield of an unbalanced line could improve things overall - however the latter case does not actually fix the problem of induced noise. This is why I believe I've had such a hard time finding reputable sources suggesting that this is a good solution for unbalanced lines (although I have found some, unfortunately also unreputable, sources that imply the opposite - see here). The article currently says that current flow is the problem and that physically interrupting the ground loop to stop the current will prevent this interference when in reality:
- Breaking the shield of an unbalanced line will never be able to fully reject the noise - at best it redistributes it to somewhere that's less sensitive to it.
- It's more likely to make the noise worse. For this not to be true, you'd need very high gains and specific (bad) circuit topologies in the transmitter or receiver that cause a smaller noise drop within the circuit to have a worse effect than the full noise drop across the line.
- As I've conceded, marking those statements dubious does not capture all of these subtleties and may not have been an appropriate way to indicate the errors in the article. Hddharvey (talk) 22:21, 1 November 2021 (UTC)→
- Let me gently suggest that this is an argument from incredulity. It cannot be imagined, therefore it must be dubious. You are only focusing on the noise seen by the input. Current traveling from chassis to chassis over the shield has other ways of causing mischief. Current is the problem sometimes. You may be thinking milliamps. It doesn’t take a lot of gain if there are a lot of amps. If there are a couple of volts of ground difference and the shield is 0.1 ohms, you can get many amps going where it should not.Constant314 (talk) 23:48, 1 November 2021 (UTC)
- I'm not sure how this differs from what I've already described. I'm not saying that internal voltage drops due to currents are not a problem, I'm saying that in comparison to the voltage across the line itself, it's likely that the voltage drop across the line itself is significant enough a factor that opening the shield worsens the performance or provides extremely limited benefits at best. Even if a reliable source showed that I was wrong about how likely these scenarios would be, the article does not discuss any of these subtleties at all. It implies the problem is the IR drop across the line itself, which opening the shield will not improve. Hddharvey (talk) 00:26, 2 November 2021 (UTC)
- I did not see any mention of IR drops in this Wikipedia article. Did you mean some other article? Here on the talk page, I took ~Kvng’s comments as referring to IR drop in the ground system and not in the cable from point A to point B.Constant314 (talk) 00:41, 2 November 2021 (UTC)
- Ground loop (electricity)#Ground currents on signal cables Hddharvey (talk) 00:43, 2 November 2021 (UTC)
- The article says in the above referenced section that the problem is IR drops across the ground shield due to ground currents, the following section then goes on to describe sources of these ground currents, and subsequently the next section says "The solution to ground loop noise is to break the ground loop, or otherwise prevent the current from flowing." - this is all quite misleading and inaccurate in full context - especially with the examples showing unbalanced lines (which as I have explained, I am highly skeptical that this is a good solution for unbalanced lines, and I have had great difficulty finding reliable sources showing that this solution is a good idea for unbalanced lines). Hddharvey (talk) 00:45, 2 November 2021 (UTC)
- I see that now. In the case illustrated, I agree that breaking the shield or ground conductor would not be helpful. There are other cases. Some solutions only apply in some situations. Breaking the shield is a solution that applies in some situations, although the analysis in the article may indeed be dubious. MidiMagic is not a reliable source and should be expunged from the article.Constant314 (talk) 00:58, 2 November 2021 (UTC)
- I agree. Ideally the article should show the shield being broken in a balanced line since that is a much more common scenario and makes much more sense (and it will be much easier to find a source suggesting this for balanced lines than unbalanced lines). I have created diagrams showing much more realistic scenarios for ground lifts in my draft which anyone is free to use as a starting point if they are interested in improving this article. In my draft, I tried to make it clear that - while current can itself be regarded as a problem (there's a section devoted to it) - it's much more complicated than that and just "stopping the current" is not guaranteed to fix the problem. Hddharvey (talk) 01:04, 2 November 2021 (UTC)
- I see that now. In the case illustrated, I agree that breaking the shield or ground conductor would not be helpful. There are other cases. Some solutions only apply in some situations. Breaking the shield is a solution that applies in some situations, although the analysis in the article may indeed be dubious. MidiMagic is not a reliable source and should be expunged from the article.Constant314 (talk) 00:58, 2 November 2021 (UTC)
- I did not see any mention of IR drops in this Wikipedia article. Did you mean some other article? Here on the talk page, I took ~Kvng’s comments as referring to IR drop in the ground system and not in the cable from point A to point B.Constant314 (talk) 00:41, 2 November 2021 (UTC)
- I'm not sure how this differs from what I've already described. I'm not saying that internal voltage drops due to currents are not a problem, I'm saying that in comparison to the voltage across the line itself, it's likely that the voltage drop across the line itself is significant enough a factor that opening the shield worsens the performance or provides extremely limited benefits at best. Even if a reliable source showed that I was wrong about how likely these scenarios would be, the article does not discuss any of these subtleties at all. It implies the problem is the IR drop across the line itself, which opening the shield will not improve. Hddharvey (talk) 00:26, 2 November 2021 (UTC)
- Let me gently suggest that this is an argument from incredulity. It cannot be imagined, therefore it must be dubious. You are only focusing on the noise seen by the input. Current traveling from chassis to chassis over the shield has other ways of causing mischief. Current is the problem sometimes. You may be thinking milliamps. It doesn’t take a lot of gain if there are a lot of amps. If there are a couple of volts of ground difference and the shield is 0.1 ohms, you can get many amps going where it should not.Constant314 (talk) 23:48, 1 November 2021 (UTC)
Turntable grounding
[edit]Ground_loop_(electricity)#In_low_frequency_audio_and_instrumentation_systems presents an example and then gives reason why it is a bad example (I have moved this to a note for now). Turntables do generally have a separate ground wire and if you do not connect it, you typically do get a hum. I have tried to research why the separate ground is necessary and whether a ground loop is involved and the information on the internet is swamped by hand waving from vinyl enthusiasts. Can anyone offer an EE treatment of this example? ~Kvng (talk) 20:14, 26 September 2022 (UTC)
- Isn't the "separate ground" from the pickup cable shield? If so, it's fairly self-explanatory why not connecting it causes a problem. I agree that that is not a ground loop problem. SpinningSpark 15:57, 28 September 2022 (UTC)
- I assume this ground connects to the chassis of the turntable. If you have a magnetic cartridge, it should be a balanced system. I guess the problem is the preamp is single-ended. ~Kvng (talk) 13:48, 29 September 2022 (UTC)