To understand where I got my data, please look at pp. 166-169 'Noise Reduction Techniques in Electronic Systems' Henry W. Ott
Here, he describes ABSORPTION LOSS, and how thin aluminum shields are lousy at low frequencies. One way of improving it, with equal thickness would be to use copper, and that was just too expensive. The other approach, used by most people, is to use steel, but we did not want to, so our only alternative was maximum thickness, which 'came with the territory' of hi end audio. We didn't hurt anything, we only helped increase 'absorption loss' due to the increased aluminum thickness, and we made at the same time, a stable and essentially non-resonant platform to mount our pots, switches, connectors and electronics. It worked and it still works for us.
However, IF we could use a better grounding scheme that made the circuitry 'bulletproof' from any RFI that might come from the outside world, I would be open to it, so long as it did not compromise the audio frequency grounding (as it often does).
Here, he describes ABSORPTION LOSS, and how thin aluminum shields are lousy at low frequencies. One way of improving it, with equal thickness would be to use copper, and that was just too expensive. The other approach, used by most people, is to use steel, but we did not want to, so our only alternative was maximum thickness, which 'came with the territory' of hi end audio. We didn't hurt anything, we only helped increase 'absorption loss' due to the increased aluminum thickness, and we made at the same time, a stable and essentially non-resonant platform to mount our pots, switches, connectors and electronics. It worked and it still works for us.
However, IF we could use a better grounding scheme that made the circuitry 'bulletproof' from any RFI that might come from the outside world, I would be open to it, so long as it did not compromise the audio frequency grounding (as it often does).
A good example. For others: A coaxial cable has no external magnetic field, so the ferrite will never see signals which come in via the core, and return via the shield. Any signal which does not meet that will create a magnetic field that the ferrite will see. At audio frequencies, the ferrite will generally be useless at dissipation, but it works well at rf. That is why it rarely is adequate for ground loop current quashing, the ferrite doesn't absorb at 60 hz.
Well stated.
As for the use of small caps with physical leads as shunts to stop rf intrusion...at the point where the shunt is placed, there will be a mutual inductance from one side of the system to the other. This is a direct result of the lead inductance of the cap.
To get around that, use a feed through capacitor. It has extremely low mutual inductance. A descrete version using smd's can easily be designed.
jn
Thanks, looks like a page turner..
. Very nice, gonna read it through and through. That's why I said it's a beddy beddy bad idea..beddy bad. Don't do it..
jn
I think adding a cap for RF grounding (but audio isolation) where a cable goes through a screen is likely to be better than not adding it. The hole will couple the inside to the outside, so is there any advantage in connecting the cap ground on the outside of the box? My guess is not. (NB for non-RF people reading this: the inside and outside of a complete metal box are two different conductors for RF purposes, so if there were no hole it would definitely be better to ground the cap on the outside)
I would have to say outside. Inside means the current has to go inside just to get to the shunt element.
One point that's not been made yet here... Given a hole in a box, there will be a minimum frequency that can pass through. A 4 mm diameter will only allow frequencies above several Ghz to pass through.
Once a wire passes through that hole, the minimum frequency is zero.. DC to daylight.
Text in red is a direct quote from Tom Van Doren, page 114, Grounding and Shielding of Electronic Systems, 2005.
jn
In this case the hole is partly filled with the socket outer conductor, which if effectively grounded will reduce coupling. Anyway, the RCA/phono socket should ideally have its ground tag on the outside, together with an appropriate grounding cap.
My thought was that as the hole couples the inner and outer currents then grounding on the inside near the hole will be almost as good as, and much more convenient than, grounding on the outside. Is this the same issue as size of hole vs. incoming radiation frequency, or a different issue?
My thought was that as the hole couples the inner and outer currents then grounding on the inside near the hole will be almost as good as, and much more convenient than, grounding on the outside. Is this the same issue as size of hole vs. incoming radiation frequency, or a different issue?
Electric field coupling yes. The partially filled hole aspect doesn't contend with the current path, just the electric field lines. Since the shunt element is low impedance at the frequency of interest, the potential of the rca outer conductor will be low, so only the currents will be of concern.
The mere sight of a hole where an isolated rca jack is denotes dc to daylight. As such, it's best to keep the currents outside the enclosure. Unfortunately, those d##m electrons don't concern themselves with convenience, they gots a mind of their own they do..
jn
Hi,
Not necessarily. This only holds true if the physical enclosure is also the only RF barrier. But I think I'm giving too many hints.
Ciao T
Not necessarily. This only holds true if the physical enclosure is also the only RF barrier. But I think I'm giving too many hints.
Ciao T
There seems to be scope here for a genuine audio gizmo: RCA socket with external ground tag, small cap, solder tag, bolt, nut, shakeproof washer. Sell in red and white pairs for $99, or $499 for the deluxe set of 5 pairs and a 12V minidrill for making the holes for the bolts. I can't write the sort of blurb these would need to sell, but I'm sure we could frighten people with thoughts about 'RF soup', 'radiation energy' etc.
John (JN),
Glad to see you chime in.
There are some other folks who made good points.
A lossy capacitor will reduce shield/cap resonance. This is easily done by adding a resistor in parallel to the cap in the example under discussion.
In a power supply bypass the practice was to use an electrolytic in parallel with a combination of a small ceramic cap in series with a resistor to aim for the same goal.
J.C. mentions that picking specific capacitors of different types and ranges can get a uniform bypass. Very application specific and using off data sheet unguaranteed parameters. Very audiophile tweaky.
Wavey, has show the basic principle of a device that really does exist but is virtually unknown, although the name is well know. Yes there really is a device that is a flux capacitor! It is essentially a capacitor like device that stores energy in the magnetic flux not the electric field. It is different from an inductor in the phasing between voltage in current.
Now there has been some confusion about the difference between the source, the path and the receiver.
The Friendly Candy Company or some such (Ice cream guys also) set standards for how much noise devices can emit. Good practice says that audio designers should be aware of not only the external emissions but also the internal effects. This is an issue in almost all digital audio products with analog outputs.
The path is what is under discussion. This has been confused with the receiver issues.
As has been mentioned starved bipolar input stages are more susceptable to EMI (RFI). The example mentioned way back around here is the LM4562. JC designed a product using it. It was discovered that adding a small resistor in series with the input lowered the system noise! Exactly opposite what would have been expected. I found that I had to add a small coil to the input to get the specified performance in the real world. I asked S about this and he mention the starved input RFI issues. So EMI at levels below demodulation issues most likely affect performance.
Pavel, please stay in the spirit of things here and avoid showing actual measurements or mentioning results based on actual experience!
Glad to see you chime in.
There are some other folks who made good points.
A lossy capacitor will reduce shield/cap resonance. This is easily done by adding a resistor in parallel to the cap in the example under discussion.
In a power supply bypass the practice was to use an electrolytic in parallel with a combination of a small ceramic cap in series with a resistor to aim for the same goal.
J.C. mentions that picking specific capacitors of different types and ranges can get a uniform bypass. Very application specific and using off data sheet unguaranteed parameters. Very audiophile tweaky.
Wavey, has show the basic principle of a device that really does exist but is virtually unknown, although the name is well know. Yes there really is a device that is a flux capacitor! It is essentially a capacitor like device that stores energy in the magnetic flux not the electric field. It is different from an inductor in the phasing between voltage in current.
Now there has been some confusion about the difference between the source, the path and the receiver.
The Friendly Candy Company or some such (Ice cream guys also) set standards for how much noise devices can emit. Good practice says that audio designers should be aware of not only the external emissions but also the internal effects. This is an issue in almost all digital audio products with analog outputs.
The path is what is under discussion. This has been confused with the receiver issues.
As has been mentioned starved bipolar input stages are more susceptable to EMI (RFI). The example mentioned way back around here is the LM4562. JC designed a product using it. It was discovered that adding a small resistor in series with the input lowered the system noise! Exactly opposite what would have been expected. I found that I had to add a small coil to the input to get the specified performance in the real world. I asked S about this and he mention the starved input RFI issues. So EMI at levels below demodulation issues most likely affect performance.
Pavel, please stay in the spirit of things here and avoid showing actual measurements or mentioning results based on actual experience!
Hi,
Funky. So THIS is why I am having a 10R resistor there in parallel with the multiple SMD 100nF Cap's...
Ciao T
Funky. So THIS is why I am having a 10R resistor there in parallel with the multiple SMD 100nF Cap's...
Ciao T
Amusing as I worked my way through the last few pages of posts this morning (all in the space of one night) and before getting to this one, almost anticipated it verbatim!
Of course I know we've batted around the pros and cons of balanced lines versus the RCA-coax "standard". If done with great care balanced helps a whole lot, but at a substantial penalty in complexity and increase in active devices for the same device series ("voltage") noise performance.
I think JC's point is well-taken: IF the environment is relatively free of hideous amounts of RFI/EMI we can get away with a good deal less bypassing, shielding, common-mode choking, etc. Sadly, even since the one-day seminar on extrinsic noise ("Ground Noise" I think it was called) put on by AES some years back, the general level of congestion, even in residential areas, has gotten to be enormously worse. Between the noisy electronic ballasts of regular and compact fluorescents, noise on the mains from everywhere-switching supplies and power line comm gadgets, on up through 2.4GHz and beyond from the ubiquitous cellular phones and wireless LANs, and who knows now what else, our typical environments are abuzz as never before. And it sucks.
An anecdote: many years ago a physicist friend showed me a recently-taken scope photo that he'd taken at TRW. He was working on a chemical laser prototype, one which used banks of triggered spark gaps and substantial power overall.
The scope photo looked a bit like an X-Y display akin to a Lissajous pattern, except much "messier". He said What do you think was hooked up to the scope to generate this? It was as they say a trick question.
The answer was, the input was open. The scope was in X-Y mode all right. But the beam was only producing a stationary spot until the power supply in the next room operated for a brief period. Then the pulsed magnetic fields caused the pattern
Brad
My statement holds for a chassis with a hole..I've not alluded to further things beyond that.
I have elaborated on the functioning of the device shown. It can see only imbalance currents in the cable. When imbalance currents are there, the toroid will dissipate based on the frequency of the imbalance current only. It is not designed to store energy, just dissipate it. If the toroid used is high permeability rather than HF plus lossy, then it will add inductance to the ground loop current path without compromising signal integrity.
Flux capacitor indeed...I knew you were gonna say that..
cheers, jn
Hi,
Ahh, so what if there a is a further RF barrier...?
Then the hole in the outside does not cause so much harm... Right?
Ciao T
Ahh, so what if there a is a further RF barrier...?
Then the hole in the outside does not cause so much harm... Right?
Ciao T
I have elaborated on the functioning of the device shown. It can see only imbalance currents in the cable. When imbalance currents are there, the toroid will dissipate based on the frequency of the imbalance current only. It is not designed to store energy, just dissipate it. If the toroid used is high permeability rather than HF plus lossy, then it will add inductance to the ground loop current path without compromising signal integrity.
Flux capacitor indeed...I knew you were gonna say that..
cheers, jn
Apple's industrial design folk had a fit that they weren't fit to have when Harman told them that there had to be a common-mode choke in the power supply line from the AC adapter and one in the USB cable to meet agency emissions/conducted noise criteria. They had already made life incredibly more difficult by insisting that the innards of the subwoofer be exposed --- no shielding allowed! This was the Soundsticks system, with two multidriver tubular satellites and a "sub"woofer that looked a bit like an old beauty salon hair dryer.
There were a lot of "laws of physics" clashes with them. One computer had drivers with inverted domes and a optically precise curvature and shiny finish. The complaint was made that if you looked into them, you saw images that were inverted. Wasn't there something that could be done about this? ? Steve's not going to like it, and customers may complain
Brad
Hi,
Ahh, so what if there a is a further RF barrier...?
Then the hole in the outside does not cause so much harm... Right?
Ciao T
My discussion is only on the simple hole and hole with wire. Engineering beyond that limited scope is where others come in.. the value added thing.
jn
- Status
- Not open for further replies.