PDA

View Full Version : 13-Year-Old Boy Improves Solar Panel Efficiency 50%



Fxc2hh
08-26-2011, 12:27 PM
13-Year-Old Boy Improves Solar Panel Efficiency 50%

"The only problem is that most solar panel arrays aren't exactly the most efficient replacements for electricity due to their need for the perfect positioning and good weather."



He created a tree-like stand for solar panels and attached them in a Fibonacci-like manner and compared the results with a standard flat solar panel array. His simple creation yielded an incredible 50 percent jump in efficiency between the traditional method and the Fibonacci design.


http://www.tomsguide.com/us/Boy-Solar-Energy-Efficiency,news-12247.html

Now if only Parallax would design a tree stand for some solar panels.....

prof_braino
08-26-2011, 01:16 PM
Comments say he compared 10 static panels (on a flat mounting) and compared this to 20 panels arranged on a tree.
http://www.smartplanet.com/blog/thinking-tech/why-13-year-olds-solar-power-8216breakthrough-wont-work/8261

Pretty good try for a 13 years old. The first of many small stumbles before an astonishing success?

Duane Degn
08-26-2011, 02:03 PM
From:

http://www.eco-scams.com/archives/746



How did this confused science project became international news?


It turns out this is worse reporting than the "rocket powered" robot arm (http://forums.parallax.com/showthread.php?133732-robotic-arm-powered-by-rockets).

It still falls short of the "cold fusion" hype.

Duane

Edit: Here's another good link about bad science:

http://optimiskeptic.com/2011/08/21/this-is-where-bad-science-starts/

Clock Loop
08-26-2011, 03:09 PM
Damn, so this is what makes news?

No study on the mysterious processor frequency rate reduction for the past 10 years?

1980-8088 = 2mhz.
1990-pentium = 75mhz.
2000-pentium4 = 3ghz
2010-Core i7 = 3ghz ****STAGNANT PROCESSOR DEVELOPMENT SINCE 2000*****

Uhh, we did NOT hit mores law wall yet. nor will we.

This is all about CONTROL, and keeping us stupefied. (thus why the above article made the news)


You want solar panel innovation? Try looking into what frequency solar panels absorb. And I am sure the range is deplorable. Even plants only use red and blue frequencies, but they manage fine.

Our panels are so inefficient, because they are not tailored to the frequency ranges our sun and atmosphere create. Even this is already known, but still nothing gets developed in this direction.
I am really getting tired of the globalist that has billions of dollars but can't make a simple good contribution in the invention area, but yet poor people innovate all the time with little or no money.
Einstein, tesla, just to name a few.

But we still allow these massive companies and people to dominate so badly that they oppress technology in fear of their own demise. Im sick of it.

Dave Hein
08-26-2011, 03:21 PM
Why 13-year-old’s solar power ‘breakthrough’ won’t work
(http://www.smartplanet.com/blog/thinking-tech/why-13-year-olds-solar-power-8216breakthrough-wont-work/8261)
Too bad the internet wasn't around when I invented a perpetual motion machine at age 13 :)

Martin_H
08-26-2011, 03:44 PM
Dave, I beg to differ.

I invented perpetual motion in 1973! I took my Erector set motor and coupled its output shaft to the input shaft of the generator. I wired the generator to the motor and gave the shaft a spin, expecting it to turn forever. Now granted there were bugs to work out as it was harder to turn than before they were electrically coupled. But I was sure the eight year old me would get it sorted out eventually.

Dave Hein
08-26-2011, 03:56 PM
Mine pre-dates your by 8 years. It used two magnets mounted at either end of a plastic tube. The tube rotated around a nail passing through a slotted area at the midpoint of the tube. A magnet was mounted below the tube, which would repel the one mounted at the end of the tube. This would push the tube up slightly, which would cause the heavier end at the top to come down. When it got to the bottom the magnet would push the tube up again, and so on. It should have run forever, but my construction skills were too limited at that time. That seems to be a common problem with most perpetual motion inventors.

Phil Pilgrim (PhiPi)
08-26-2011, 04:06 PM
Even though the kid's results have not stood up, his basic idea was truly inspired.

I hope he doesn't become discouraged from a career in science because of all this. There should be someone close to him to explain that fame is, by nature, fickle, that not all experiments stand the test of time, and that it's okay if you stub your toe once in awhile. You just have to keep trying.

-Phil

Martin_H
08-26-2011, 04:22 PM
OK Dave you convinced me that you have prior art. I agree to not contest your perpetual motion machine patent when it is awarded by the US patent office.

Phil, I agree. I hope someone explains to him that there's plenty of failure in the careers of most success people. My recent favorite example is Andre Geim won the Nobel Prize in physics for the discovery of graphene. But ten years prior he was a bit of a laughing stock when he won an IgNobel for using magnets to levitate a frog.

Gadgetman
08-26-2011, 04:24 PM
Our panels are so inefficient, because they are not tailored to the frequency ranges our sun and atmosphere create. Even this is already known, but still nothing gets developed in this direction.


The panels used on the ISS and on sattellites all use cells that absorb at least 3 different frequency ranges.
The problem is that for the moment they have to make 3 different cells and sandwich them together to do this.
Result: 3x efficient cells costs more than 3 times the price of a single-efficiency cell.
The same with most that uses more than one colour frequency; colour lasers, colour photography(analog), offset printing...

Alternative solar-power systems(focusing mirrors to heat salts or other 'high capacity' media, mostly) are gaining ground, but as they require a lot of 'off panel' energy extraction systems, they're not exactly suitable for small-scale(home) usage.
And unless there's a very large market for the product, they won't be produced.
And we NEED a 'home usage' sized system to efficiently drive down cost.

BTW: The easiest way to increase the efficiency of a solar panel today is to use a sun-following mount.
Strange that the kid never noticed sunflowers...

The second easiest is to use mirrors(must be cheaper than the cells) to focus additional sunlight onto the cells.

erco
08-26-2011, 06:34 PM
Solar cells are also more efficient at lower temps. They get very hot (duh) out in the sunlight. Seems like combining a solar water heater/photovoltaic array/heat exchanger would be a win/win all around.

Big money. Forget I mentioned it. Where's that delete key?

GordonMcComb
08-26-2011, 07:33 PM
2010-Core i7 = 3ghz ****STAGNANT PROCESSOR DEVELOPMENT SINCE 2000*****


Well, there have been numerous articles and papers on why the CPU fabbers have concentrated on multiple cores rather than raw speed increases. Multiple core processing is more power efficient for one thing (and better for battery-operated devices) -- you can turn a core off if it's not needed, yet the other cores can continue their threads at rated speed.

The i7 has six cores, and adding more cores is now the natural progression. It's no longer pipeline speed. I don't think stagnant is a fair word to use.

Now to this poor kid, and the usual show of pettiness you see when discussing an engineering and scientific discovery (or lack thereof): give the little guy a break. So he got excited that his research somehow centered on the coolness of Fibonacci sequences. He's not the first to think there's something more in that progression than meets the eye.

The quote from a responder on the energy blog, and reprinted in the article cited above, demonstrates why we have so few kids willing to enter science. Loaded works like "nonsense" and "horribly" are typical of engineer speak. I get so sick of that crap as they attempt to prove their intellectual or educational superiority (and certainly proving their lack of social grace). They're so used to these kinds of negativistic "cut them down to size" phrases -- they employ them daily among their peers and management -- they don't even know they're using them.

Just say the findings are unmerited, and the concept needs further research, and then explain in simple terms why.

-- Gordon

treborz17
08-26-2011, 08:10 PM
Actually, solar power credit should go to the pioneer who discovered Yuma Arizona, my place of residence, where tourist are given free hotel rooms, and in some cases free meals any day the sun doesn't shine. Of course we have a few inconveniences such as mail is still provided by Pony Express, and a shortage of gun racks inside of grocery stores, restaurants and churches.

Hich up and visit a spell,

Robert

erco
08-26-2011, 08:20 PM
Well-said, Gordon & PhiPi. It's easy for armchair experts to overreact and dogpile onto the "flame de jour". Any naysayer can tear down. It takes discipline and restraint to apply gentle correction. Blame the media for the hype, not the 13-year old boy.

I'm usually a bit skeptical when I see claims of fantastic breakthroughs. It wasn't the kid's specific claims that caught my attention. I was more impressed by the professonalism of the his presentation. It's a lucid and thoughtfully presented writeup that is noteworthy. So definitely, let's laud his observations and nurture his scientific apirations. Otherwise, he'll turn out to be just another bitter old skeptic like me. :)

bsnut
08-27-2011, 04:43 AM
I agree with erco post 100%. The 13-Year-Old did a good job in how he presented and did his project.

He also found different way to catch more sunlight on the panels. When you get more sunlight on the panels, the voltage will stay constant for as is there is sunlight.

It's just good has the panels laying flat or facing south. Hack, I'm not getting this kind of result on the ship that I am restoring and my panels are laying flat. But, I don't have much choice, because laying the panels is the best way for me.

Clock Loop
08-27-2011, 05:19 AM
Well, there have been numerous articles and papers on why the CPU fabbers have concentrated on multiple cores rather than raw speed increases.

Numerous articles and papers? Duh, they need to explain the lack of progress (at least in the civilian sector).
Thats similar to the bank robber who is actually the security guard for the same bank. You expect him to tell on himself? No he is going to tell everyone that he saw the bank robber get away.

We will find out in the future what DARPA is doing right now with clock speeds behind the closed doors, and you will see that I am right.



I don't want to take the thunder of this theads topic, but im sorry, copying a core and increasing the wafer size to accomodate is not innovation, its keeping the dumbed down public somewhat happy that they think they are seeing innovation. Dual cpus are not innovative, and also, they could have developed multicore stuff along side the frequency. If parallax can make a 8 core mcu, intel should be making a 100core cpu, not 4 or 8. That kind of innovation was SUPPOSED to be a mid point in the year 2003.

Back to the solar panel, the kids idea was mostly a failure due to improper scientific testing standards. Hopefully he will learn that, and hopefully he will also learn that most inventors make 100 designs that don't work(failures) and only strike it on the 101st invention.

If that kid gets discouraged by all this talk, then he doesn't have what it takes to be an innovator, thats ok, many don't.

rod1963
08-27-2011, 06:47 AM
The kid deserves a A for effort and should be encouraged. Not raked over the coals by a bunch of egomaniac armchair engineers and scientists to make themselves feel good.

No wonder kids don't have much interest in science and engineering. Can't blame them when these technical people act like creeps towards them.

Heater.
08-27-2011, 10:14 AM
100 cores dipping in and out of common shared RAM, as we get now for 2 or 4 cores, would be kind of dumb. The RAM access would be a bottleneck that limits performance. As I often say, an infinite number of cores sharing the same RAM would have zero performance.

Looks to me like there is a sweet spot around 8 cores where core processing and RAM accessing reach a balance and performance is maximized. As in the Prop. For example I feel that 16 cores for a Prop 2 would have been a mistake.

prof_braino
08-27-2011, 04:44 PM
Back to the solar panel, the kids idea was mostly a failure due to improper scientific testing standards.

Sorry Clock Loop, the kid's idea was NOT a failure, it was an investigation. Any investigation that returns a result that can be examined is a success. It is a success because we analyze the snot out of it and determine if found he something interesting, and discover where he made mistakes. Kid still deserves an "A" for the project.

"Improper scientific testing standards" - remember, this is a kid, his job is to play and learn. Playing and making mistakes is how we learn. The same applies to adults, but adults try not to admit it.

The "failure" is the news reporting that reported his misinterpretation of his results as a breakthrough without checking.

Number of processor cores and processor speed is a separate topic, why don't you start a separate thread? That would be a good discussion, I thought there were a whole bunch of factors at play.

Clock Loop
08-27-2011, 05:22 PM
"Improper scientific testing standards" - remember, this is a kid, his job is to play and learn.


Teach proper testing methods, for instance why was more/less solar panels installed in the Fibonacci vs flat configuration? Proper scientific method suggests using equal panels in either situation, and compare results on graph, simple and proper.

prof_braino
08-27-2011, 05:43 PM
Teach proper testing methods, for instance why was more/less solar panels installed in the Fibonacci vs flat configuration? Proper scientific method suggests using equal panels in either situation, and compare results on graph, simple and proper.

Very good. You demonstrate that you have the same understanding of proper methods as the rest of us. Now are you going to go to the kid and teach him about it? That is where the information is needed. The OTHER part of proper method is getting the information from the party that has it to the party that needs it when it will do the most good. Telling the kid BEFORE he finished his report would have done the most good, tell him now will still do some good, but telling us who already know and agree will not have much benefit.

Clive Wakeham
08-28-2011, 12:26 PM
Well, there have been numerous articles and papers on why the CPU fabbers have concentrated on multiple cores rather than raw speed increases. Multiple core processing is more power efficient for one thing (and better for battery-operated devices) -- you can turn a core off if it's not needed, yet the other cores can continue their threads at rated speed.

The i7 has six cores, and adding more cores is now the natural progression. It's no longer pipeline speed. I don't think stagnant is a fair word to use.


-- Gordon

The i7 has eight cores, at least my does.

ratronic
08-28-2011, 03:44 PM
@Clive Wakeham and GordanMcComb - which core i7 models do you have? I have a 2600k and it only has four cores but eight threads.

GordonMcComb
08-28-2011, 04:36 PM
AFAIK, the i7 only (currently) has a max of six cores. These are the Gulftown chips, part of the Sandy Bridge chipset, and you may remember Intel had to recall those earlier this year due to an issue where a portion of the chip could degrade over time. (I actually bought one of these SB machines, from Costco, but they had me immediately return it.)

A four core i7 has eight threads. These are often indicated as having eight cores, which is true to a point, but they're virtual cores, not actual physical cores.

-- Gordon

ratronic
08-28-2011, 08:42 PM
Thanks for that Gordon. I dealt with the recall issue myself with an Asus motherboard, the recall came less than a week after I purchased it.

Clock Loop
08-29-2011, 03:20 AM
A four core i7 has eight threads. These are often indicated as having eight cores, which is true to a point, but they're virtual cores, not actual physical cores.

-- Gordon


Most also don't understand that if you enable this technology, known as Hyperthreading, you cut the maximum speed any single program can run at by 1/2 your max processor speed. (at least when the tech first came out) but even so, today, if your core that is split, gets 2 programs using 100% then each program really only processes at 1/2 the speed of the full core.

So either way, hyperthreading sucks, is a joke, and was just an example of lame technology replacing the development of higher clock speeds.

prof_braino, actually many users here are young and the same as that kid, who have now learned that using a good scientific method is good advice, for any adventures into the physical manipulation and measurement of our world.

Phil Pilgrim (PhiPi)
08-29-2011, 07:23 AM
Any time my shop PC's BIOS memory gets reset, due to battery replacement -- or, who knows? something going bump in the night? -- I wonder why its performance sucks so bad. Ohhh, right! I've got to go back in and disable hyperthreading again. Otherwise, while one thread is running my program, the other will be lollygagging in System Idle Mode.

-Phil

erco
08-29-2011, 04:28 PM
Any time my shop PC's BIOS memory gets reset, due to battery replacement -- or, who knows?
-Phil

Who keeps a computer long enough for that button cell to go bad? :)

Phil Pilgrim (PhiPi)
08-29-2011, 05:16 PM
I've still got an ancient Gateway Pentium II PC running Win98 that's on all the time. It's tethered to the main WinXP machine via a KVM switch and knows tricks that XP and later versions could never master (like how to read a parallel-port security dongle). I suppose, once the magic smoke has completed its slow leak, I'll have to upgrade the software to the newest version, which doesn't require the dongle.

-Phil

jbarry1506
09-04-2011, 02:55 PM
I have trouble believing this kid is being beat up like this for attempting something that most 13 year olds couldn't even dream of doing on an educational forum! This young gentleman is a success and will most likely grow into an accomplished and seasoned engineer. Remember Thomas Edison and this important lesson:

"After we had conducted thousands of experiments on a certain project
without solving the problem, one of my associates, after we had
conducted the crowning experiment and it had proved a failure,
expressed discouragement and disgust over our having failed to find
out anything. I cheerily assured him that we had learned something.
For we had learned for a certainty that the thing couldn't be done
that way, and that we would have to try some other way."

The quote above comes from an interview with Edison that was published
in the January 1921 issue of American Magazine.

kwinn
09-04-2011, 05:21 PM
@jbarry1506, while I agree that he should be commended for attempting something that most 13 year olds couldn't even dream of doing I have to point out that most of the respondents on this thread were supportive of his efforts. The vast majority of the criticism was aimed at those who picked up the story and disseminated it without checking the facts.

I expect that people who report on science and technology would have enough of an understanding of the field they are working in to be able to separate the possible from wishful thinking, bad science, and outright fraud. Sadly, I have had to lower my expectations considerably. That so many members of this forum questioned the results certainly separates them from the babbling bumblers that did report it without questioning it's validity.

Gadgetman
09-04-2011, 05:55 PM
My oldest 'in use' system is an old P133MHz/96MB RAM system that happily runs eCS(repackaging of OS/2 v4.0), but I also have a 386/25MHz I'm prepping for use.
(It has 2 x high-speed UARTs, so I want to use it for serial logging, and I also happen to own this card: http://www.jschoenfeld.com/products/catweasel_e.htm which really needs a good home...)

Being a science reporter today seems to mean 'was able to switch on the GPS and use it to travel to the tech show'... And if it's a 'fuzzy, huggy' show(green tech, school fairs, maker faires and so on), it seems that sending out an intern is 'good enough' for most... You know, they have to save the 'real' journalists for the important news, such as who is sleeping with who, got divorced, is using drugs, or the fact that yet another Homeland Security droid screwed up at an airport and nicked an iPod or groped a woman.
Journalists or wannabees that cover these things never have the articles checked for correct science, only spelling, grammar and readability.
(It's a test to see if they can be trusted with more serious cases like firemen rescuing cats out of trees and such.)

Heater.
09-04-2011, 06:10 PM
I never understood why firemen don't get cats out of trees with a quick blast of a fire hose.

kwinn
09-04-2011, 06:11 PM
@Gadgetman, I see that we share the same regard and level of respect for the vast majority of the news reporters.

kwinn
09-04-2011, 06:17 PM
I never understood why firemen don't get cats out of trees with a quick blast of a fire hose.

That would probably discourage cats from doing it again but OMG the uproar from cat lovers would give the nitwit news brigades several field days worth of reporting.

prof_braino
09-04-2011, 09:13 PM
I never understood why firemen don't get cats out of trees with a quick blast of a fire hose.

I suspect it has to do with the time and effort to put the hose away. I did a few quick experiments and fought almost any hose that will reach will move just about any cat out of just about any location, and quickly.

In the course of which I tested the myth about cats "always landing on their feet"; with the proper english or sidespin they can be made to land in just about any orientation.

Also below a minimum height, cats CANNOT land on their feet, and get very annoyed and run away after only several dozen repetitions.

Unfortunately I was unable to test the British technique for measuring rooms by swinging cats, as in "big enough to swing a cat in", as they had all run away by that time.

Gadgetman
09-04-2011, 10:01 PM
Also below a minimum height, cats CANNOT land on their feet, and get very annoyed and run away after only several dozen repetitions.


I believe that is slightly under 3' in height?
(Yes, someone actually researched it, and even used high-speed cameras to film what the cat did... Science.. don't you just love it... ;-)

And the show Brainica that ran on Discovery Channel once tested the 'room enough to swing a cat', but the pansies used a plushy instead of a real cat, so their results can't be trusted.
(The weight was wrong, which affects how you pivot your body during the swing. Possibly, the cat is able to bend its body, too, if it's swung at a low speed?)

erco
09-05-2011, 03:04 AM
My favorite unwritten children's book: "Some Kittens Can Fly".

Browser
09-05-2011, 03:16 AM
I never understood why firemen don't get cats out of trees with a quick blast of a fire hose.
we are not amuzed

-browz

ElectricAye
09-05-2011, 04:47 AM
I never understood why firemen don't get cats out of trees with a quick blast of a fire hose.

Don't make the climb if you can't do the time.

http://i51.tinypic.com/2ltivdd.jpg

Mark_T
09-05-2011, 09:50 AM
No study on the mysterious processor frequency rate reduction for the past 10 years?

1980-8088 = 2mhz.
1990-pentium = 75mhz.
2000-pentium4 = 3ghz
2010-Core i7 = 3ghz ****STAGNANT PROCESSOR DEVELOPMENT SINCE 2000*****

Uhh, we did NOT hit mores law wall yet. nor will we.

In case anyone's wondering the internal speeds _are_ increasing, its just that a chip/system wide clock speed is limited by the speed of light (half a clock cycle at 3GHz on a PCB is about an inch). Modern processors use transistors with switching times of a few ps IIRC - a local clock of 100GHz would be possible but could only remain coherent over a fraction of a square mm. Clock-skew across the chip is a major issue in VLSI design now.


Anyway back to the subject: its an indication of the poor state of science journalism - and in particular the lack of any fact-checking by news agencies who propagate 'interesting sounding' stories world-wide in hours. According to a recent article on the BBC news website on rare earths they claimed that a major use of the element lanthanum is hybrid car engines. In fact its the major consistuent of NiMH batteries (its the commonest choice for the metal M)... Even 1 minute check on wikipedia would have put them straight but 'news consumers' aren't apparently worth that much effort...

I sympathise for the poor guy at the centre of this story because he's been let down by a succession of people who could and should have spotted the error and nipped this media frenzy in the bud.

Heater.
09-05-2011, 12:40 PM
Prof_Braino,



"...tested the myth about cats "always landing on their feet"


As a young lad I had to do those experiments in "feline dynamics" as well.
It just bugged me that supposedly "for every action there is an equal and
opposite reaction". So how could they do it? What were they pushing against?.
Experiment made it all clear and I also found they need a minimum height drop in
order to pull it off.

Hmm...wonder if we could build a catbot that can do that...

"...measuring rooms by swinging cats, as in "big enough to swing a cat in""

and Erco:

"My favorite unwritten children's book: "Some Kittens Can Fly".

Cat based measuring systems and cat ballistics were strictly disallowed after my
parents witnessed the first experiments.

P.S. I actually like cats. What makes my stomach churn is all those millions of
supposedly cute and humorous cat photos on the net with dumb captions added.
It's disrespectful of cats. Why do we insist on belittling these superior
beings?

prof_braino
09-05-2011, 02:09 PM
My favorite unwritten children's book: "Some Kittens Can Fly".


"feline dynamics" ... a catbot that can do that...
Cat based measuring systems and cat ballistics
... cute and humorous cat photos on the net with dumb captions added.

NOW yer talkin'!

Certain research, using a red laser pointer, indicates that cats can accelerate to two to three meters per second on a level surface, but cannot make 90 degree turn in either direction left or right, and end up skidding on the smooth tile surface and smashing into the wall. Repeatedly.
However, the SAME set of cats were capable making a 90 degree turn STRAIGHT UP at the same wall. Sometimes they could two, maybe three meters. Unfortunately the batteries ran out before accurate measurements could be recorded.

A green laser pointer did not yield the same results. I wonder if the green laser pointer is not as visible to the cats' black and white vision?

What parts of the cat need to be simulated in a catbot? I think the "meow" is necessary, as the Doppler shift is useful in determining velocity, distance, and rate of rotation.

Tor
09-05-2011, 02:40 PM
NOW yer talkin'!

Certain research, using a red laser pointer, indicates that cats can accelerate to two to three meters per second on a level surface, [..]
oops.. now you did it.. using a laser pointer to play with a cat is patented (http://www.freepatentsonline.com/5443036.html), you know?

-Tor

Bean
09-05-2011, 04:04 PM
oops.. now you did it.. using a laser pointer to play with a cat is patented (http://www.freepatentsonline.com/5443036.html), you know?

-Tor

Yet another example of the utter stupidity of the people running the U.S. Patent system. But that is another discussion for another day...

Bean

GordonMcComb
09-05-2011, 04:39 PM
Every couple of months I get a call from some patent (busting) attorney looking to find prior art for a patent they're defending against. Old books are a gold mine for them, because books have a verifiable availability date, and they fill the PO's requirement for a widely published resource. If the "invention" is described in enough detail in the book, and the dates of publication and patent filing fall into place, the idea becomes prior art. The patent is effectively invalidated.

In a 1988 book I briefly mention using small lasers to -- among other things -- entertain and exercise animals (it works with most dogs, too). I was hardly the first to think of the idea. If this "method of exercising a cat" patent had ever gone to litigation it might have been disallowed, based on previously published disclosure. They did not reference my book in their patent, which had been out in wide circulation for at least five years before their initial filing.

It's not clear that the patent was filed as a serious contender. Sometimes these types of patents are filed to fill out a portfolio, or simply to prevent someone else from doing it, and disrupting their business (was this filer an importer of penlight lasers, for example?). In some cases the filer has no intention of enforcing the patent.

As a point of interest, this particular (and now well known) patent expired four years ago. Apparently its filers lost interest and didn't pay the maintenance fee.

-- Gordon

Phil Pilgrim (PhiPi)
09-06-2011, 02:14 AM
I think some patents are filed just to demonstrate how out-of-control the patent process is. Here's a perfect example:


http://www.google.com/patents?id=T2QKAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false

I assume the patent trolls will now be monitoring playgrounds around the country for infringing behavior.

-Phil

Tor
09-06-2011, 09:57 AM
I think some patents are filed just to demonstrate how out-of-control the patent process is.[..] Yep. In 2001 an Australian lawyer patented the wheel just to show how broken the system is (easily found by A Famous Web Search Engine). Afterwards (after the publicity) the patent office complained about the lawyer not having been truthful about the originality of the invention.

-Tor

Clock Loop
09-06-2011, 06:31 PM
In case anyone's wondering the internal speeds _are_ increasing, its just that a chip/system wide clock speed is limited by the speed of light (half a clock cycle at 3GHz on a PCB is about an inch).

Oh no, don't even start this up. For the sake of letting this thread die, im just going to say that most signals don't travel over an inch. Do you have ANY idea the REAL size of a i7 DIE? You will find all the moron speak you want, because its out there in abundance, afterall they must explain their lack of progress with the billions of dollars with which they research.
Moron speak, which will reinforce the excuse for a lack of progress, add that to the fact that intel is continually increasing DIE size to compensate for a lack of following mores law... and you have most people happily satisfied that all is ok.

You CAN make it look like you are following mores law, by increasing the DIE size.
Most electrons that propagate through a chip, rarely go long distances before they reach the slowdown in the semiconductor(even this could be made into a speed UP), let alone any odd long distance traces, even the memory controller is being cpu integrated.

And you are seriously accepting this info as excuse? What about laser interconnects, (which intel has had that tech for decades) inter cpu radio transmission, etc. We have many avenues of eliminating the many barriers that our BILLION DOLLAR COMPANIES LIKE INTEL have no problem with researching AT ALL.

PERIOD.

But morons will repeat morons into eternity, to make themselves feel better about the lack of progress that is going on, all around them. So they can be happy about the 3ghz cpu they just overpaid for.

Roy Eltham
09-06-2011, 06:47 PM
Clock Loop,
You should look up what Moore's Law actual is, because it really has nothing to do with performance or clock rates. It's all about transistor density. Also, the Quad Core i7 die is about 263mm on a side (so just over an inch by just over an inch).
Also, I hope you don't think I am a moron, because I don't agree with you at all on this topic.

Duane Degn
09-06-2011, 08:16 PM
I didn't realize Moore's was a mandate of what semiconductor businesses had to do. I thought they were in it to make money.

In my opinion using terms like "moron speak" weakens your argument (as does most name calling in a debate).

Do they need an excuse not to follow Moore's law? I don't understand why they have to explain their lack of progress. (They probably have to explain lots of things to shareholders but I image shareholders care more about how much profit they make than how fast the processors are.)

I'm also disappointed processors haven't been getting faster. I personally don't know how to solve the problem. I don't know the reasons semiconductor businesses haven't been able to make faster processors but I bet their decisions are financially motivated. I doubt they worry much about Moore's law.

"But morons will repeat morons into eternity"

If one of us quotes a semiconductor business, does that make us a moron?

Duane

Roy Eltham
09-06-2011, 08:30 PM
Duane,
Processors have been getting faster. Just not higher in clock rates. Today's Core i7 CPUs run circles around old school Northwood/Prescott Pentium 4's even though their clock rates are lower by default.

Heater.
09-07-2011, 07:04 AM
The speed of light is an important issue today. In 1ns light only manages to crawl about 1ft. That is going to limit your computer to 6Ghz if your memory is one inch away from your CPU. In most computers the memory is further away than that.
Having optical or whatever connection does not help with that fundamental limit.
"Moores Law" is no such thing. It was just an observation at the time.

Gadgetman
09-07-2011, 10:19 AM
Clock Loop,
You should look up what Moore's Law actual is, because it really has nothing to do with performance or clock rates. It's all about transistor density. Also, the Quad Core i7 die is about 263mm on a side (so just over an inch by just over an inch).


I assume you meant 26.3mm each side? as 263mm on each side would be larger than some motherboards?
If so, then 26.3mm is EXACTLY one inch... One Norwegian inch, but still an inch.
The Swedish 'Working inch' was 24.74mm(small hands on the Swedes, it seems) until they went for a 'metric inch' of 29.69mm...

I remember seeing a picture of a DIGITAL Alpha chip once, and the clock alone took up around a quarter of the die. That's what was needed to 'push' the signal around the die and to all parts without distortion.

Tor
09-07-2011, 11:04 AM
I think the old Norwegian inch is obsoleted by now.. :-)
The official SI inch is 25.4 mm

-Tor

Gadgetman
09-07-2011, 11:26 AM
It was still in the textbooks when I learnt carpentry back in 84... Gosh, maybe I'm obsolete?

Besides, SI,don't they define a KiloByte as '1000', too?
http://xkcd.com/394/

I think some marketing drones once noticed that their competitors all sold computers with 64KB RAM/ROM, and looked at the specifications of their own product and noticed 'Hey, we've got a whopping 65 Thousand of them litle thingies and the others only have 64, let's use that!'
(This was back in the days of the 8bit)
I think the entire home computing press (and thousands upon thousands of proto-geeks, too... ) laughed themselves silly over it. And that's probably who and why there so much push to bring 'metric' standards into the computing industry...

Tor
09-07-2011, 12:42 PM
It was still in the textbooks when I learnt carpentry back in 84... Gosh, maybe I'm obsolete? Not you, I think, but maybe your textbooks.. the old Norwegian inch was 26.14mm or 26.15mm (before or after ~1870), 26.3 was much older. As far as I know the official Norwegian inch changed to the SI inch when it was introduced, somewhere around 1959. All my old carpenter rulers ("tomstokk") that I checked used 25.4mm, but a colleague owns one from the seventies which uses 26.15mm. It could be one specially made for boat builders - if you want to build an old wooden boat you'll want one with 26.15mm inches. He told me though that in construction (building) it wasn't unusual to use the 26.15mm definition way into the 1990's.

I for one wish most of the old non-metric measurements good riddance, we don't need another Mars lander incident. I can even stomach the new SI definition of MB, I can manage to write MiB when I want to use the value we computer folks need. :-)

-Tor

Gadgetman
09-07-2011, 03:46 PM
It wouldn't surprise me if the boatbuilders used the old Norwegian inches a while after the change, so yeah...

I really should know about it as my father was a boatbuilder, and his father after him.
(My father learned the craft at Skorgenes, later 'Tresfjord Boats' back before the advent of plastic. He even exported a fair few but had to stop when fibreglass boats started flooding the market. Of course, most of his boats still floats, but the fibreglass boats from that time have mostly rotted away. My grandfather took over after that as he 'needed something to do' and his eyesight wasn't good enough for him to continue as a tailor. I doubt he ever used a decent ruler, though... )
But unless you're using some very special formers(templates made of wood placed along the keel to help the boat hold its shape, when building a wooden hull, it probably doesn't matter all that much which inch you use as long as it's consistent.

MiB? That acronym always makes me think of a movie...

Roy Eltham
09-07-2011, 04:16 PM
My apologies, the Quad Core i7 die is really closer to one inch by half inch in size (not one inch by one inch. Of course, it varies in size based on the amount of cache and that doesn't count the variant that has the GPU built in (those are larger).
The size of one core on that die is in about 1/4 inch by 1/4 inch, but they do all have interconnects that run the length of the chip.

Beau Schwabe (Parallax)
09-07-2011, 05:12 PM
This thread is amusing to me.

There is a finite speed depending on the design process that determines the ultimate clock speed. BUT that's only for 1 clock stage. What I mean is... for example:

A 180nm process might have a maximum clock speed of around 300MHz ... this is due more to parasitic capacitance in the layout itself. So how do you get clocks above that? It's possible, and it's done all the time.

The trick is to setup what's called a ring oscillator with various 'taps' ... each tap represents a slightly different amount of Phase delay from the 300MHz clock. Each tap is buffered and OR'd into a single output or 'master clock'... for example a 300MHz clock with three taps placed so that the propagation delay between them was about 1.1ns would effectively yield a 900MHz clock ,,, 4 taps placed 833ps would yield 1.2GHz

So now that you have a clock from essentially a singular source, what you have to consider now is the 'time of flight' it takes that clock signal to propagate from point A to point B.

Think about fiber optics.... with todays speeds you can quite literally have the 'optic cable' act as a memory device. What I mean is that when you send data over a fiber optic line. As soon as it leaves your computer, the data is in transit propagating through the line before it ever reaches it's destination.

In a way the clock signal or the data becomes a FIFO (First In First Out) the 'speed' at the beginning is the same 'speed' as it is at the end, however there is an initial propagation delay... the same sort of thing happens over a fiber optic line and even within the IC from one side of the chip to the other usually just causing an error in Phase delay. ... but it's the propagation delay due to parasitic capacitance and IR drop (<--current drop over distance due to resistance). Dealing with the speed of light at the moment is the least of our concerns.