Isolation/coupling: basics?


I feel I need some education in this regard, and I guess I'm not alone... I read most of the discussions about it, but I couldn't find the basics: why?
Could anyone who understands the physics behind all this explain why those vibrations, resonances, and energies are that bad, especially for components without moving parts, such as amps?
dmitrydr
Read Bad Vibes by Shannon Dickson from November 1995 Stereophile.

See: http://www.stereophile.com/showarchives.cgi?52

PS: Have your enginerring & thinking hat on!
Take a high powered telescope. Look through it first on a tripod and then while holding it in your (trembling) hands - that's basically analogous to what vibrations do to an electronic signal.
OK...here is one point of view. (I'm sure you will get the other one too).

Phono pickups develop their signal by vibrations of the stylus caused by the record groove. Obviously any other source of vibration will also cause the phono pickup to generate a spurious signal. Virtually all phono playback systems have this problem to some degree. In my system mechanical vibration effect is (as far as I can tell) zero, but I still have acoustic vibration (sound) which makes the vinyl LP vibrate like a drum skin. Various record clamps have been developed to minimize this problem. The best solution is a vacuum system that draws the LP into solid contact with the turntable, so it isn't flexible any more.

Vacuum tubes have various internal pieces, the exact spacing of which determines parameters of the tube, like gain. Vibration causes these pieces to move, and that generates a spurious signal. This effect is called microphonics, because the tube acts like a microphone. This problem can be almost completely eliminated by good equipment design and reasonable placement. High gain tubes, in preamps and the input stages of power amps are most likely to exhibit audible microphonics.

It is claimed that transistors, capacitors, resistors, and even wires can exhibit microphonic effects. With the possible exception of some types of capacitors exposed to severe vibration I know of no scientific basis for the claims. Few inductors are used in today's audio circuits, but a defective one, with loose wire, could be microphonic.

It is claimed that vibration degrades performance of digital equipment such as CD players. The mechanism for this degradation is supposedly errors in reading the digital data. Indeed, vibration might defocus the optics and degrade the signal so that a "one" is misinterpreted as a "zero" or vice versa. However, some errors like this are expected, and the digital information on the CD is encoded for error correction. Unless the error rate exceeds what the error correction algorithms are designed to deal with the information is recovered without errors, as if no transmission errors had occurred. (An analogy would be my spelling mistakes in typing this post. After I run it through my spell-checker, can you tell whether I made any errors when I typed it?).

Speakers present an interesting situation: some feel that they should be isolated, for example by hanging them from the ceiling on chains, while others strive to "couple" them to a solid floor with various devices, often cones. (I'm in the isolation camp). The notion that a few ounces of vibrating driver cone could move a suspended fifty pound speaker enclosure is absurd. What does happen is that flat panels of the enclosure can vibrate because of the pressure changes internal to the enclosure, but this problem relates to internal bracing of the enclosure and is unaffected by how it is mounted. Although the enclosure panels move much less than the speaker cone, they have a lot more area than the cone, and can radiate a significant part of the sound emitted by the speaker system. If you choose to couple the speaker to the floor, three (not four) cones would be best to avoid rattles.

I bet this post will get some of our friends all shook up!
El: Not only can individual components and circuits be microphonic, you can literally hear the sounds eminating out of transistors that they are amplifying during certain conditions. In some RF amplifiers, this can be heard quite plainly and is referred to in the industry as "transistor talk". The only way that one could hear this is if the conductors inside the transistor casing were vibrating enough to resonate the outer case in time with the signal that was being amplified. If those vibrations can be transferred from inside the transistor case to the outside case, and be done in a fashion loud enough to be easily heard and understood, what is stopping external vibrations from taking a reverse path from the outside of the transistor case and modulating the conductors inside the case ??? My guess is that the same thing occurs in audio gear, but with audio gear, you typically have the system running in the same room. The sound from the speakers will smother the sound of the transistors "talking".

I've seen / heard / could not believe this, but one can talk into an inductor that is coupled to a high gain audio circuit and hear your voice come out of the speaker. It is not clear, but it is plainly audible. The inductor that i heard this demonstrated on was wound on a plastic former and dipped in wax to try and minimize microphony. Even with all of this, it was still quite sensitive to outside stimuli. Inductors are commonly found in the majority of amplifiers that are currently on the market.

As far as other "passive" devices being microphonic, much of this comes from the wire leads or legs attached to the component vibrating. As such, it is best to keep the leads on each component as short as possible in order to avoid this. Many higher quality capacitors actually come with a sleeve over the legs so as to damp any external vibrations from being transferred into the body of the cap. Sean
>
Sean...Thanks for your input. I take your comments under consideration and will respond later.
Sean...When you lift your coffee cup to take a sip, you change the mass distribution of planet Earth, and change the distance from the earth to the Moon. This is absolutely certain, and the exact change of the distance could be calculated, but who cares? Effects such as those mentioned by sean do exist at some level, but is that level one that we should care about?

Inductors and capacitors take their values from the geometric dimensions of their construction. Unless an inductor is fully potted (filled and encapsulated in plastic) it is much more prone to dimensional changes due to vibration than is a rolled mylar capacitor. I have no doubt that one could actually create an inductive microphone (capacitative ones are common) by deliberately increasing the design features leading to microphonics, and placing the device in a circuit designed to be sensitive to inductance value.

I forgot the output transformers of tube amps. Power transformers all hum, so I suppose that audio transformers must sing. These effects are not necessarily reversible, so I don't know if an audio output transformer is microphonic. If so, I expect it is a lot less so than the tubes associated with it, and therefore of secondary importance.

It is necessary to consider HOW the electronic component is used in the circuit. I have not built or repaired any amplifiers for a few years so I am not familiar with recent circuits. As I recall, inductors were used in the signal path only at the amp output, to limit bandwidth to 200KHz or thereabouts so as to preclude possibly damaging oscillation. If such an inductor were to change value the effect would be to vary the HF rolloff slightly around the 200KHz value, which would have no practical impact on a 20KHz signal.

Do transistors "talk"? Well yes...in fact most electronic circuitry makes various buzzing noises that can be heard if speakers are not swamping it out. But do transistors "listen"? I don't accept this without evidence.

Such evidence would be surprisingly easy to gather. I suggest that sean (in his spare time) should take some of his numerous spare audio equipments, and throw together a Preamp/Interconnect/Power amp/Speaker wire system, and put a dummy load (resistor) at the end of the speaker wires. This rig should be located in the listening room and, with the preamp input shorted, turn it on, and then crank up the regular audio system to about 90 dB. The output of the test rig will now exhibit its total microphonic effect from all causes (except from a turntable/phono pickup that we all agree is vibration-sensitive). You could look at this output on a scope, listen to it with headphones, or record it on a cassette (and send it to skeptics like me). Measure the microphonicly induced output. Also measure the output when the main system is not playing, as this is the background noise level. Measure the output of the power amp that is driving the speakers: this is what the test rig would be doing if it were connected to the speakers. Now you can calculate how many dB down is the system noise level, with and without microphonic contribution. I suspect that a tube preamp will measure significantly worse than a transistorized one.

"Worse"?? Well that's another issue. Some have suggested that a little microphonics is good. Adds "body" or some such thing.

Are we having fun yet?
Eldartford, do you mean that for SS gear this issue is not really an issue? Do you use any kind of isolation/coupling yourself?

In regards to CD players, I just don't get it. Optical reading mechanism reads data, and it doesn't really know what kind of data is it: audio or data files. Processing of that data actually determines it. And optical systems of CD drives seem to be tolerant enough to vibrations to be able to read the data with no errors even in computers, where drives vibrate like crazy. The only explanation I could throw is that in spite of the fact that vibration defocuses the optics and causes read errors, a computer, reading data in X speed, is able to read, check CRC, re-read again and again until he gets the perfect read. Audio CD player has only one attempt, and if it isn't lucky there will be a error. Do I miss anything?
Any number of people have testified to the positive effect that Audiopoints/Sistrum products have had on their systems and I in no way doubt their results. However, the explanation offered as to why these products works seems incomplete or even flawed. The idea that cone type points can control the flow of energy in one direction only is false. Coupling a component to a rack and then coupling the rack to the floor allows for any vibration to flow in both directions. As a practical matter if you live in a house with a poured concrete floor, then coupling will probably work very well. If you live in a wood frame house, then isolation might be preferred.
Dmitrydr... My source electronics are located in a massively constructed alcove in my listening room. If I have any vibration at that source equipment it is acoustic. The phono turntable has vibration-absorbing feet.
My power amps are in the cellar.

IMHO vibration is not a significant problem with SS equipment. By "not significant" I mean that a problem would be hard to induce, and easily corrected. Others disagree. I am sure we all (well most) agree that tube gear is more problematical than solid state.

What you describe as the error correction method is not how it is usually done these days. You assume that the data receiver DETECTS that there is an error, and requests retransmission. That works in some cases, however CD's (and most other devices) use something called a Reed-Soloman code. The transmission includes redundant information, basically each bit of the data word is spread across a number of bits in the transmission, so that if one or more transmitted bits are screwed up there is still enough information to calculate what all the data bits should be. No time-consuming retransmission is necessary.

Reed-Soloman error correction is vital for applications such as sending video back to earth from a space probe around Jupiter. Requesting retransmission would be impractical with a round trip communication time like 40 minutes. The Reed-Soloman code can be implemented with different degrees of redundancy in the transmission, deprending on what the bit error rate is expected to be. I seem to remember that for space probe applications it is not unusual to transmit 150 bits or more so as to be sure that one bit is correct.
And don't be forgettin' sesimic vibration, due to continuous motion of Earth's crust, peak energy around 1-2 Hz.
Geoffkait...And then there are land tides. Yes, the earth's surface rises and falls several inches each day. This is 0.0000116Hz. Put that in your subwoofer and smoke it.
Eldartford, just to make sure, are you sure Reed-Soloman error correction is implemented in regular CD players? It seems that CDP and long distance signal transmission are very different applicaitions; this method seems to be good when you need to recover the most possible info when retransmission is not possible, but not when you need to get absolutely 100% of information or to report a error. So, I'm not really sure this method is being used in data storage techniques, when re-read attempt doesn't cost much. When CDROM reads a data file (I'm not sure if the data is retrieved using Reed-Soloman error correction or not) and detects CRC error, it does perform re-read on lower speed. Just because of that it is possible to get a bit-perfect copy of CD using digital audio extraction, which is generally impossible using audio CD player.

Whoever is right, it leads us to conclusion that CDP is error-protected enough to be immune to vibration too. But it conflicts with a practical experience: people claim, at least, changes in dynamics when isolation/coupling device used...
I don't personally make audio CDs, but audio CDs are often cited as an every day application of Reed-Soloman error correction. CD ROM protocol on your PC may very well be different.

One thing that I learned recently is that in current engineering practice the purpose of error correction is not to correct errors. Rather, error correction is used so that the data transmission can be run at a much higher speed than that which the hardware would support without errors. Correctable errors are expected to occur. You give up some of your bandwidth to redundancy of the coding, but you more than make it up in transmission speed.
If error correction and speedy digital data processing are good things, why is it that CD "burns" made at high speed typically sound "inferior" than a burn made at 1X or 2X ? At the same time, why do burned CD's have increased readability from player to player when dubbed at lower speeds than those at higher speeds ? As far as i know, digital is NOT like analogue where spreading the signal over a wider / longer surface area increases dynamic range. I'm not trying to be a smart-ass here, just trying to better understand what's taking place and interject "real world" situations into this theoretical debate. Sean
>
sean...When someone asks a question that we can't answer off the top of our head we generally say "send us your data". So: what evidence exists (other than anicdotal) that CD copied at low speed "sound better" and play more reliably?

In this case I will make an exception and take a stab at it. The Reed-Soloman coding used for CDs (and for any other application) is configured to deal with a certain bit error rate. In addition to the R-S error correction process I believe that CD players implement a data interpolation process to minimize the impact of errors that are not fully correctable, perhaps because of physical damage to the disc. Discs copied at high speed may have bit error rate in excess of the design value even when there are no scatches, interpolation may be invoked, and you can hear it. Warning...this is just my educated guess.

I will not attempt to explain how error correction codes work because I would not do it very well. I know there is an excellent explanation on the web (and a lot of confusing ones too) and I will try to find it again and point you to it.
Thank you El. I'm not saying that you are wrong, but that my experiences and those of many others seem to contradict some of what you are saying. I also know that theory is called "theory" because it is not always "reality". Granted, "theory" is reality in many cases, but the manufacturers "short cut" approach won't let the design perform as it should. Sean
>
This is beginning to sound like the old "bits is bits" discussion. To the naive engineer (such as myself), the digitally encoded music signals would seem to avoid all the mechanical vibration issues which dominate analog turntable playback. Yet, there is a shocking difference in the sound of various CD players - error correction or not. The degree of vibration isolation in CD transports and players is a distinguishing product feature. In my experience, the sound of a CD player can be affected by vibration coupling or isolation - this has been reported by many people in these forums.

From my perspective, the effects of vibration on an audio system are undeniable and can be profound. What I would like is a "model" which correctly describes this. (Perhaps that was the question that started this post). Vibration can affect the analog circuits in a CD player, but how do they affect the digital stream or the DA conversion? Perhaps there is a mechanism for modulating the timing of the converted signal? Or are there also errors involved?
Guys, extraction/write speed doesn't directly relate to a sound quality, it depends on HOW do you extract the audio data to HDD, and what CD-R drive you use to write it back.

Getting back to the subject :)...
Sean, you play a lot with isolation. You haven't find any satisfying explanation of the phenomenon (impact of vibration to a CD player other then just data read error which may not be a problem if Eldartford is right) from engeneering perspective?
Sean...And I'm not saying that you are wrong either, (about sonic effects) but that the technical explanations which heve been put forward don't seem adequate. I will remain a skeptic until one of the following happens:
1. An objective test of some sort (I suggested one) shows that vibration affects the equipment (to an audible extent).
2. A terchnical explanation that I can't fault.

If I personally hear an effect, it will be hard for me to remain a skeptic, but I will try to do so because I know how easy it is to fool oneself.
So if you knew it were to vibrate and you knew of a way to direct the vibration away from its own self would you be inclined to dampen the initial vibrating device?...Tom
Vibration is bad, except in loudspeaker cones. Isolation and coupling are good. But at some point, the law of diminishing returns sets in. My equipment does not require the extreem measures which some other folk find necessary.
My 2 cents

1) I have observed that CD's copied at a higher speed are less reliable. I attribute this to the shorter amount of time the laser has to burn the pit into the substrate.

2) Much is made of the error correction affecting the sound in CD players. Unless your CDs are pretty beat up, I would guess the error correction is used far less than you think.

My first CD player was a McIntosh MCD7000 (mid 1980's). It had an indicator on the front that would light up EVERY TIME the player was interpolating data (correcting for an unreadable error). Even with that early transport (high speed transports for computers hadn't yet been invented) the light ALMOST NEVER came on. Maybe once every few CD's.

I attribute the differences in CD player sound to other issues, such as D/A's, analog output stages and jitter (another issue altogether - not related to error correction).
Sometime before I made quite a lot of searches in internet about DAE issues. My observations show that even the most modern CD players/drives have a lot to do with data reading errors, and it very varies from drive to drive. And that normally CD players (it's actually their firmware) don't check data integrity as good as CDROMs do, which, however, have got improved with the years. The reason of why that indication almost never came on might be a 'weak' criteria when interpolation is really required. While interpolation methods also got improved, on 16/44.1 they still may not be refined enough comparing to the 'original' (otherwise all transports would sound the same). You can't get the 'original' when playing CD in CDROM either. So, it seems that the only way to get the 'original' is proper DAE onto hard drive that doesn't have reading issues, and then playing it from there., which is another story with other issues.