Electromigration in interconnects, wire

[efzauner]

Just finished a 3 day course in Microwave IC design. One of the topics was electromigration on the gold interconnects on the IC. It is a problem for ICs. The suggested rule of thumb is a maximum current of about 10mA for a gold interconnect of 1micron wide by 1 micron thick. Much more than that and the gold atoms will be pushed around by the electrons and can cause opens. So with all this snake oil abound, I am sure that some cable companies make use of this fact to sell their super fancy expensive crap. A quick calculation reveals that the current would be 10000 amps in a 18 gauge wire!

So any snake oil suggesting that electromigration is a problem in interconnects is a load of BS.

[Guest " "]

does this vary by frequency...is the 10ma at DC

[Cal Blacksmith]

Apples and oranges. Don't try to apply MICROWAVE theory to audio;
there is enough confusion already we don't need to add to the problem.

Wire is
one of the largest snake oil scams in audio. Just use wire of the right size
(12 to 10 gauge) and use good quality terminators for speakers and buy good quality
interconnects (rat shack gold line or blue jeans cable) and don't give it
another thought. Many will disagree but the placebo effect is strong and if you
BELEAVE you will hear a difference, you probably will though in scientific
tests, you won't find any audible differences.

EDIT: OBTW, Nickel connectors are just fine, there is no difference sound wise with them over gold. Gold will cut down corrosion over time if the cables are in an extreme
environment such as by an ocean with the salt atmosphere but other than that,
they only look prettier. Even then, if you just give your connections a twist every 6 months or so, you won't have a problem then either.

[efzauner]

does this vary by frequency...is the 10ma at DC

The concern was with bias supply conductors to the transistors.

[efzauner]

Electromagnetic theory applies at ALL frequencies. It is just that at audio frequencies most of the effects are so small to be insignificant. The problems is that the snake oil people find some neet effect at RF/Uw and then tell everyone that it is important at Audio. Knowing EM theory lets us see thru the BS. The reality is that only someone self taught or with a college degree learns enought to understand. The remaining 95% of the population are the targets.

The Electomigration effect on ICs is actually a DC effect, but only a problem on very very small conductors as found on ICs. If we do some estimates, like I did, we can quickly see that it is caused by current densities a few orders of magnitude higher than found audio interconnects. That is how we take theory or observations, do the calculations and determine what we need to worry about at the power/frequencies we care about. An experienced designer knows what to worry about and does not make all the detailed calculations every time. The anoying thing is that the snake oil people make us do the same calculations over and over again to prove them wrong every time. And then they give us more bullshit so that they save face and can continue selling $10,000 speaker wires.

Another area applicable to both microwave and audio is wavelength and characteristic impedance. In Microwave or RF design we can regard a transmission line simply as a lumped LC if the length of the Tline is less than about 20% of the wavelength. This also applies to audio.

On an IC at 20GHz, a wavelenght is a few milimeters and we cannot always make the lumped element simplification.

At 30KHz, a wavelenght is about 6 miles, so we do not care about characteristic impedance of any normal length interconnect, just their lumped LC.

At video frequencies of 30MHz like those on HD component video, wavelength is about 10 meters, so we do care about 75 ohm characteristic impedance of video cables.

So you get an interview between a knowledgable audio guru and a snake oil company president that goes like this.

Audio Guru: Tell me how your speaker wires work:

Snake oil pres: We have patented blah blah based on Maxwells equations covering the blah blah effects of electromagnetics on blah blah.

Augio Guru: but at audio frequencies those effects are only .0000001%. Do you have any measurments or double blind test to back up your claims:

Snake Oil Pres: But Einstein proved that blah blah .00000001% is significant because of blah blah. We have 1000 happy customer that never returned the products during the 90 day money back return period. Plus Patents blah blah cover our design, and research paper blah blah by Sigmund Freud shows double blind tests to be unreliable.

Audio Guru: But that Einstein calculation has to do with light traveling to the next galaxy, not on audio signals traveling thru speaker wire. And that patent covers a method on how to manufacture better sounding piano wire by putting some silver in it, not how adding .0000001% silver improves the sound of a piano played thru copper wire. Furthermore, that Freud study showed that pubescent boys could not tell the difference between B and B flat when cheerleaders practiced outside the music class window and the double blinds where open. So again, please prove to me your statements.

Snake Oil Pres: You must be jealous that I am making millions of dollars selling this crap and you are obviously biased agains my company and if you publish any of this you will hear from my lawyer.

[SilverSport]

Cal...I read your post and am sure your recommendations were made to help but I wonder why you limit your speaker wire to 10 or 12 GA (that WAS speaker wire, yes??? [])...I have been happy so far with 14GA from Signet Wire (came with my kg2s from HiFi Hutch in "the day") as well as some nice Mon$ter wire given to me (VERY pliable clear jacket) to Carol brand "extention cord" and lamp wire...I have not really heard any difference between any of them...just their visual look and the pliability of their respective jackets (for some reason that makes me THINK quality...more snake oil???)...I tried 12GA (the aluminum and copper-silver colored and copper clear jacketed) unknown brand of speaker wire from Home Depot and didn't like it (but the twisted ends I use were difficult with my older Heresys and amps)...as for ICs...I agree with your choices...add RCA branded clear jacketed with a silver colored woven shielding from Home Depot or Fryes (I believe this to be identical to Radio Shack gold series...at least the connectors are identical AND they are cheaper than RatShack I believe)...I have tried other ICs that friend's or my brother has had...Nordost Black Knight (flat wire) had a bit nicer highs but lost a bit of bass...may be in my head...[])...currently trying Tributaries-Delta Audio...nice connectors (gold plated) very flexible blue jackets but...I don't really hear any sound advantage (or disadvantage for that matter)...over the RCA branded ones I have...I had tried these other ICs hoping NOT to hear a difference...I don't think I could bring myself to pay what these companies want for their product...seeing as I bought MANY RCA branded ICs for the cost of those TWO Nordost ICs...I am curious to hear if the 12 or 10 GA wire is any improvement (providing the proper termination and better wire than I tried)...Bill

[Cal Blacksmith]

Bill;

Yes I was referring to speaker wire, though I might like to

see what 10-gauge coax looks like! 14-gauge wire is fine for 99% of the users

out there. I use 10 gauge stranded O2 free speaker cable that I bought at my

local electronics retail/wholesale store for $34 / 50 spool and could not tell

any difference between it and the 14 gauge wire I had run before, except that it was softer and more flexable and looked cooler. Is O2 free wire needed? I dont think so,

the main reason for that alloy of copper is so that it draws better than

standard copper and thus is easier to form for things like pots and pans etc.

it imparts no sonic qualities that I know of.

For long runs I would go 10-gauge, for normal runs 12-gauge

just because the price difference between 14-gauge and 12-gauge speaker wire

(not snake oil wire!) is only a few pennies per foot and why not go with a

little larger wire? I come from an electrical (not electronic) background and

run 12-gauge wire in walls just for the extra comfort it gives me knowing I

have overbuilt the circuit.

I have some interconnects that cost about $100/3 that run

between my pre-amp and amp but they were given to me by the seller of the amp.

I dont think there is much of a chance I would have bought them on my own! For

most of my connections I use either Rat shack gold line at about $15/3 or some

China made A/V cables I picked up at the same electronics store as above, for

about the same price as the gold line cables, heck they both probably came out

of the same factory!

I guess the advice I take to heart is that if I am happy

with what I have, I quit looking! If you are happy with your speaker wire, it

is fine! I really doubt you will hear any difference by going to a larger gauge

wire, I know I didnt and it was a much larger jump than what you are thinking

about!

[Guest " "]

I really enjoy these discussions about speaker wire, line conditioners, RFI/RMI mats, cone brass spikes for speakers, etc.

One thing I do notice, is that the playing feild is not level....everyone does not have the same equipment, source material, power grade, etc.

Everyone does not have the same line voltage and same grade of house wiring. Single house, one apartment amoung 100 in a complex.

There are so many technoligies (transformer coupled, capacitor coupled, direct coupled, push-pull, single ended, OTL, and all the classes of operation A, AB, B, D, etc) .

Some can tell the difference between the platforms others can not.

Some can tell the difference between wires, others can not.

Some can notice a change if you replace the cord going to the amp from a 4 dollar cable to a 99 dollar monster powerline 6 ft electrical cable that has 2 layers of sheilding, highly thick and pure copper strands and a speacial plastic center and coating.

I have often wondered if the folks who are on one extreme have simular platforms, environments, listening preferneces one way or another.

Persoanlly I think there is a point were the cable, power grade, topology can be incrmentaly improved with in a cost range....but a good example of where a line can clearly be drawn is this buisness about premium and economy grade fiber optic inter connect cables ...like the type used from cd, video, satellite, etc. The advantage of fiber optics is that it is ditigal 1's and 0's...it is either on or off....a ditigal signal, in my mind, is uneffected by the median used to transport it.

[efzauner]

regarding digital signals, they are just as much affected by the media aka wire or fiber that transports it. Yes, they may be just ones and zeros, but add noise, loss, and other more complex distortions and one and zeros get mixed up and cause errors.

For the fiber optic TOSLINK, there is the electro to optical conversion and back again at the other end. The fiber itself, the light having to go thru the air interface of the connector, connector reliability, dirt on the connector etc. It is not as black and white as that. The same digital signal on an cable with RCA connectors can be just as reliable if not more. We have been transporting electrical signals for a hundred years or so and we understand the issues very well.

As for the difference between the $4 power cord and the $99 dollar one, I have yet to see any reliable double blind study that shows that anyone can tell the difference.

[SilverSport]

Cal, thanks for taking the time...cool thread...and no one blew their top...[]...yet!...Bill

[Guest " "]

I will remain open minded about the fiber optics. We have about 10,000 miles of it running through out our enterprise network that spans 15 locations. We send audio, video, data, telephone, security systems through it. Seems on the surface...the cable is less of any issue than the converters on both ends. I thought...maybe I am wrong...that these transmissions in optical wire used CRC error correction and if there were bits of imperfection the bit would get retransmitted. I am not an engineer....well I'm not certified as one anyway...I seen CRC error status frequently and retransmission rate data...and for the most part it is a very small and barely measurable part of the overall transmission.

So we are back to the topology in which the optical cable may be a tweaking point if the audio devices on both ends are so good that the optical cable is the weakest link. Has the technology reached that point already...I don't know...I have owned and still own models of high end equipment that were state of the art...but later...another model came out with better something or another.

On my 4000.00 reciever and my 1200.00 CD player I can not tell a difference with differenct grades of optical cable...as much as I would like to say a 79.00 optical cable sounds better than an 8.00 dollar optical cable.

Maybe comparing a 8 dollar optical cable to a 79 dollar one is not a good enough spread, and thats why I can't tell a difference.

[Jay481985]

regarding digital signals, they are just as much affected by the media aka wire or fiber that transports it. Yes, they may be just ones and zeros, but add noise, loss, and other more complex distortions and one and zeros get mixed up and cause errors.

Just from what I heard that digital has a contial loop of the information a few times over to fill up any gaps and loss and noises that digital tends to lose. Nasa perfected the technique for sending and relaying informations to satellites and those mars probing devices. By sending the signals back and forth to mars I believe it takes like 1 minute for the signal to reach plus having cosmic radiation effecting the transmission and images is more akin to having bare copper wire in a bath of salt water while sending electricity through it. But we still have control of those rovers and the images are still clear. This is do to redundancy of the digital system. By sending the information say a million times over your bound to complete the losses made from noise and errors. Though I do admit I have no knowledge how well the redundancy on say a digital connector is on a reciever.

[efzauner]

I have nothing against digital, and I did not say that TOSLINK is no good. I just wanted to say that digital and optical links also have transmission issues. Those 10Gb fiber optic cables need amplifiers and signal reconstruction at regular intervals. To compare a Toslink to a 10 GB single mode fiber link to sending signals to Mars is apples and oranges. That is why I said "other kinds of distortion". An in depth technical discussion of the issue is beyond the scope of this forum and is better suited for www.ieee.com. But if you say that digital is OK because it works for your thousands of miles of fiber, the we should also say that analog is better because we have millions of miles of telephone cable. It all depends on the execution and the tradeoffs made to achieve the overall goals. Both systems will work very well if done properly. Both can result in bad music if done poorly.

The problem with the snake oil people is this: If you have not done grade one math then you are not capable of refuting their statement that their $1000 cable makes 1+1=3. So you believe them and spend the money. The same is for EM theory: Only in advanced undergraduate university levels to you begin to learn enough to be able to see thru the bullshit that is spewed from their mouths. The layman is doomed.

[Jay481985]

oh no.... the thing is I believe snakeoil fools the misguided too. People believe that engineers or anyone in the know uses the best quality products. When I worked at a lab we basically winged it and tried to recycle and reuse products to fit our needs instead of buying the best. Its funny how overinflated checkbooks become overinflated ego on "the best"

Also the digital debate, I was just trying to say that digital has "redundacy" (sp I'm tired) to fill the gaps and signal losses which occur just as much as analog, just that analog has no redundancy issue so I was trying to apples to apples the fact that the digital signal from here to mars is the same digital used to send through a toslink, though different algorithms always apply. But the basic concept still there.

[efzauner]

OK on the spelling, hope you give me some slack too! Family was over today, had some BD cake and coffee too late. Cannot sleep now.

I am not sure why Toshiba developed it since the digitized dvd soundtrack is not that high data rate and is easy to send over coax/RCA. Look at 100MBit lan cables over twisted pair! They are now working on 10Gbit over very much the same cable! Perhaps it was a marketing thing to differentiate themselves. To be honest, I do not know if the TOSLINK has error correction, not all digital links do, and there are many different forms of digital communication, many ways that the analog signals can be encoded into digital. Like I said, Digital and Analog can both work when done correctly. Compare CDs to SACD. Both are digital, both use error correction, but they use different sampling and encoding techniques and sound different.

As for engineers using the best quality etc., yes we do, we use the best quality that will make the product meet its technical, reliability and cost requirements. Agree on overinflated checkbooks=overinflated ego, but that is normal that you will want to feel good after you spend the money. I have said in other posts, I don't so much worry about the millionare spending $1000 on a power cord, but more the average Joe spending $1000 on a HT system and $500 on fancy speaker wire and interconnects. The $500 would have been better spent on better equipment, and some simple good quality but not overpriced cabling. I do turn green and my clothes start ripping off my back when I hear stuff like putting speaker cable on top of glass spacers to keep them off the floor; just like glass insulators are used in high voltage power lines. Or a lot of the other BS going on out there.

[Jay481985]

OK on the spelling, hope you give me some slack too! Family was over today, had some BD cake and coffee too late. Cannot sleep now. I hear you, I have super insomnia with my best was 84 hours!

I am not sure why Toshiba developed it since the digitized dvd soundtrack is not that high data rate and is easy to send over coax/RCA. maybe they were looking toward the super future? Look at 100MBit lan cables over twisted pair! btw that twisted pair, if you ever open one up is well shielded and the wires are probably 32 gauge if that, I know that the bell wire at radioshack for science fairs are thicker by a large margin. and cat5e or cat 6 is good to 200 feet lengths and handle much more data /information than any audio cable will transfer. They are now working on 10Gbit over very much the same cable! Perhaps it was a marketing thing to differentiate themselves. To be honest, I do not know if the TOSLINK has error correction (me either), not all digital links do, and there are many different forms of digital communication, many ways that the analog signals can be encoded into digital. Like I said, Digital and Analog can both work when done correctly. Compare CDs to SACD. Both are digital, both use error correction, but they use different sampling and encoding techniques and sound different. true but in theoretical realm there should be no difference but there is......

As for engineers using the best quality etc., yes we do, we use the best quality that will make the product meet its technical, reliability and cost requirements. well then there is always the we cannot afford it, try with the materials at hand approach that schools other than MIT, Cal Tech, etc etc do. Agree on overinflated checkbooks=overinflated ego, but that is normal that you will want to feel good after you spend the money. I have said in other posts, I don't so much worry about the millionare spending $1000 on a power cord, but more the average Joe spending $1000 on a HT system and $500 on fancy speaker wire and interconnects. The $500 would have been better spent on better equipment, and some simple good quality but not overpriced cabling. well said I do turn green and my clothes start ripping off my back when I hear stuff like putting speaker cable on top of glass spacers to keep them off the floor; just like glass insulators bahahahhahhaaahahhah they use glass insulation fibers not glass.... I think the material is a different silicate too..... are used in high voltage power lines. Or a lot of the other BS going on out there.

[Guest " "]

" putting speaker cable on top of glass spacers to keep them off the floor; just like glass insulators are used in high voltage power lines"

good one...thanks for sharing.