On Fri, 03 Dec 2004 02:04:04 -0500, "J. Clarke"
wrote:

wrote:
much general snippage

Uh, do you understand what is meant by 'proof of principle'? Hint: It
is not a prototype.

If the principle is that the device can be used to replace a table saw then
the "proof of principle" is a device that replaces a table saw.

No that's a prototype.

I don't care about "proof of principle" that some kind of device
can be made

That's become painfully obvious. In fact it leads me to wonder why
you're so intent in participating in this discussion at all.

Your gut obviously tells you that the tool I am describing will never
exist. And that's fine. Your gut may even be right. However the logic
and facts you are attempting to use to support your gut feeling are
anywhere from fatuous to flat wrong.

What's more, your arguments are rapidly degenerating into a series of
flat statements with no support whatsoever. Which is increasingly less
convincing.

--you
claim that that device can do something, but you don't have any backup for
that claim at all, just brainless advocacy.

Wrong on both counts. I am claiming that _in another few decades_ the
woodworking tool I am describing could easily exist. Clearly a device
that I think will exist in a number of years can't be said to do
anything at all today.

As for backup for that claim I have cited a number of examples
demonstrating that the technology is coming into existence. By
contrast your 'evidence' so far has consisted of a single citation of
a paper from which you drew a correct, but utterly irrelevant
conclusion. (Of course microturbines get their advantages from being
small. That's the whole point.) It also appears you didn't bother to
read the entire paper -- or at least you missed a couple of tables and
discussion that answered one of your other questions.

I suspect that if the
engineers and scientists who are working on this are reading this thread
they are cringing and what you are claiming because they know that they
can't deliver it and it won't be remembered that it was _you_ making the
claims and not _them_ later.


If any of those scientists and engineers are participating I'd be very
interesting in hearing their opinions.

From my discussions with scientists and engineers involved in MEMS,
active structures and such I doubt seriously any of them are cringing

(And on the the side issue of flyng cars
Actually, the two big obstacles have always been cost and runways.

I am not aware of any case where runways or lack thereof had a
detrimental effect on flying cars. Are you? On the face of it, it's
difficult to see how they could. The essence of flying cars is that
the vehicle is both an airplane and a car. It was not, as generally
conceived, a personal helicopter. In other words it flew as close as
it could reasonably get to its destination and drove the rest of the
way.

Cost is a more difficult issue simply because it is more speculative.
However examination of the structures and components of various flying
cars shows that a lot of them could have been produced at prices which
would have given them a significant market. (Not as big as
automobiles, obviously.)

I'll disagree on both counts. The cost of some of the designs in
volume production would have been less than an luxury automobile.

You can buy new airplanes now for less than the price of some luxury
automobiles. Most people can't afford to drive a Ferrari though.

And many people can't afford to fly a private plane. However thousands
of people can afford it and do fly them. It's a non-argument unless
you're trying to claim that the flying car would have to replace the
automobile to be a success. That's a fairly nonsensical standard.

And
the runway issue was addressed by a variety of the designs in
different ways.

Addressed by what designs other than helicopters that actually flew well
enough for anybody but an experienced test pilot to survive the experience?

Again, the essence of a flying car is that it acts as both an airplane
and an automobile. It's hard to see how the runway issue would have
been significant. Especially given both conditions and attitudes in
the heyday of the flying car craze in the decade after World War II.
Towns and cities everywhere were building airports. So were private
individuals.

Helos
address runways but they still need a good deal of space and make a huge
amount of noise. While in principle I could keep a helo in my back yard,
in practice the neighbors would lynch me in a week. The new designs use
ducted fans for vertical takeoff but they don't promise to be any quieter

Well no. A major component of the noise from a helicopter is the
interference in the air flow between the main and tail rotors. If
you've ever heard a NOTAR chopper you'll see they are significantly
quieter.

Have you ever had one crank up in your back yard at 2 AM? "significantly
quieter" and "quiet" are not the same.

For a discussion of noise levels and reduction in
helicopters, see:

http://www.aviationtoday.com/cgi/rw/...=08rwcover.htm

Ducted fans and similar designs are even queter and can be made
quieter yet with active noise reduction technology.

Yeah, yeah, rah rah rah. Now, have you ever stood next to anything with a
high powered ducted fan as it spun up to full power?

Have you? As it happens I have.
But let's quantify this discussion. Give me an acceptable noise figure
(and profile) in EdB -- as well as a source for it -- and we'll have
something to discuss.

Try it sometime and then tell me how quiet it is.

I used to work across the street from the plant where Boeing (ex
MacDac ex Hughes) builds NOTAR helicopters, as well as Apaches.
There's also a helicopter flying school there. So I've been exposed to
a lot of helicopter noise. Even the difference between a conventional
helicopter and a NOTAR is considerable.



The ones that are furthest along promise both reasonable fuel
efficiency and a cost less than a high-end sports car. And this is
only the first generation.

Uh huh. If you've been around aviation long enough you'll have seen all
kinds of "promises" that were never delivered. And nothing that uses lift
fans is ever going to match the fuel economy of a Honda Civic.

No but it can be quite thrifty on a gallons per mile basis.

(Back to the main argument)

Now, what device that has been made or even designed has these capabilities
that you claim will be made available by this technology?

That would be a good point -- if I was claiming this woodworking tool
exists. I do not and in fact I don't expect such a thing to exist for
several decades. I don't know why you have so much difficulty grasping
this, or why it makes you so angry. But you obviously do and it
obviously does.

Now if you want to know exactly how these tools will be designed,
you'll have to find someone with a clearer crystal ball than mine.

In other words you don't have a clue whether your precious little MEMS can
actually do what you're claiming or how they might be used to do it if they
can. All you have is bad science fiction.

Wrong again. See previous discussion and citations. What I am saying
is that a lot of the design will depend on how the field develops. If
you think you can predict the exact shape of cutting edge devices even
five years out -- well, you're going to be seriously wrong more often
than not.

Given what I have seen already, and the way the industry works,

What "industry"?

Semiconductors.

The MEMs industry hasn't been around long enough for you
say anything about how it works.

MEMS has been around as an industry for more than a decade. That's
long enough to see the patterns developing and to compare them to
other high technology industries.

If you mean the electronics industry,
don't assume that MEMS is like electronics.

In what ways is MEMS different from electronics? Don't just wave your
hands, give specifics. Justify your answer with appropriate citations.

I can
tell you that something with those capabilities and using these kinds
of principles could be available in a few decades.

Or not, as the case may be.

There we agree.

Personally I'd say that "not" is the way to
bet. At least not based on the technology you are hyping. Some other
technology might come along that allows it of course.

Personally I'd say that it will happen, but that's what makes horse
races.

Trying to predict
exactly what it will look like or how the details of how it will work
will lead to something like that 'RAND corp. design of a personal
computer' that's making the rounds of the web. We just don't know
enough yet.

Was that "RAND corp" which is think tank or was that Remington-Rand the
computer manufacturer?

RAND stands for "Research ANd Development". It is a
government-sponsored think tank which concentrates on high technology.
It was established after WWII and AFIK has no connection with
Remington-Rand. As for the 'personal computer' . . . well, do a little
research and find out. No reason to spoil the joke for you.

In any case, at least they knew how a computer
worked. You don't have a clue how the devices you are hyping would
actually work.

I not only have 'a clue', I've seen the principles I'm talking about
demonstrated in the lab, in production or in other contexts. You could
get an excellent basic education in them if you were willing to read
the research papers, company literature on existing projects and other
reports.

Modern computers are small and inexpensive because the components from which
they are made are very small and there are only a few of them.

Even done a parts count on a modern PC? Even with the current level of
integration, there are still a lot of parts.

Computers are small because it is to our advantage to make them small.
If we had reason to make them large we could make them large -- and
still inexpensive. Do you seriously believe this thing is going to be
size of a modern laptop?

Now how are
you going to cut wood with that few pieces that small?

The sensors and actuators are going to be small. Where do you get the
weird notion that this tool is going to be made entirely of silicon?

Hmm? Or are you claiming that all of a sudden massive lumps of semiconductor-grade silicon
are going to become dirt cheap because they're being used to make MEMs
instead of microprocessors?

Silicon is going to get a lot cheaper but what makes you think the
active elements are going to be composed of semiconductor grade
silicon?

For the record: Some of them may well be -- if we're still using
silicon. But a lot of MEMS technology can be easily built with much
cheaper grades of silicon since the electronic charcteristics don't
matter.

Can you quote a single researcher who has actually developed such a device
who is making such claims?

Again the confusion over the existence of the tool. I'm talking about
several years out.


You've done a lot of "rah-rah" stuff but you haven't
demonstrated how something that is only cheap if it is made small is going
rip a piece of 2" lapacho in less than a month.

You're confusing the sensors and actuators (which are small) with the
complete tool (which isn't) and the cutting element -- which will be
sized appropriately for the tool.

I'm not confusing anything.

Incorrect.

You're claiming that this technology is going to be cheap

True

and it's going to be made entirely out of MEMs.

Wrong. I'm claiming it's going to incorporate MEMs elements as key
components. It is no more going to be 'made entirely out of MEMS' than
a modern desktop computer is made entirely out of silicon.

If that's the case

It is not.

I don't know what you're reacting to in all this, but it clearly is
not what I am actually saying.


then the active components have to be very small or it's not going to
be cheap.

The active components, in the sense of things like actuators and
sensors, will be small. They will also be cheap, but not just because
they are small.

In MEMS, as in electronics, economies of scale are a major
consideration. The cost to produce something in quantity, no matter
what the size, falls very rapidly.

Now, how much power can a MEMs actuator that can be made with less than, say
$200 worth of silicon produce?

Wrong question. The right question is 'how much power can a bunch of
dirt cheap MEMs actuators control?' The answer is 'more than enough'.

And there is no indication that the cost of silicon per se is going to go
down.

Untrue, as it happens. The price of silicon is on a long-term downward
trend. In 1959 metallic silicon cost a little over $1 per pound. By
1998 or so it was down to around 60 cents a pound and headed lower.

I see.

I hope so.

So it's come down 40 percent in 40 years.

Which directly contradicts your claim. You're apparently making this
stuff up as you go along and that is not a good strategy.


(The highly refined silicon used in making semiconductors is currently
running about $30 a pound. However that's pretty much irrelevant to
this discussion because of device differences and what drives prices
in that market. A couple years ago that same silicon was selling for
about $30 a pound.)

http://www.usatoday.com/tech/news/20...ar-cells_x.htm

So it's $30 a pound and it used to be $30 a pound and you just shot down
your own argument.

Oops. My error. A couple of years ago that highly refined silicon was
selling for *$3* a pound, not $30.

The first reason the cost of semiconductor silicon today is irrelevant
is that what drives prices in the semiconductor silicon market is
refinery capacity versus worldwide demand. The 2000 recession
disrupted that market and the recovery disrupted it in the other
direction.

The second reason it's irrelevant is that you don't have to use
semiconductor silicon for most of these devices. The reason we do so
today is that the methods of processing semiconductor silicon are well
understood. It's more convenient for researchers and it's cheaper for
relatively small production runs. However both researchers and
manufacturers are rapidly developing competency with other matetrials,
incuding less pure grades of silicon.



http://www.digitimes.com/NewsShow/Ar...ages=A5&seq=19

This reference takes you to a paid subscription site. Did you actually
look at it?


With processed wafers the actual computations are quite complex
because there are an enormous number of factors, both positive and
negative, in play. However if you hold the size (area) of each device
constant and the feature size constant (which almost never happens)
the devices end up being a lot cheaper as the wafers get bigger.

Define "a lot".

For starters you get about a 2.25 increase in device count, plus other
economies of scale -- principally in processing consistency. To
balance that you have the somewhat higher cost of the handling and
processing equipment.

And tell us how that translates to something large enough
to cut wood being cheap.

You're still hung up on this thing being built entirely out of
silicon. Again, that's like assuming that an entire desktop computer
is built out of nothing but silicon.

We're talking about components like actuators and sensors here, not
complete tools. And of course you're going to fit a lot of them onto a
wafer. But like current MEMS devices they will be diced and packaged
before use. You don't have to put the whole tool on a single wafer.

So what good are little bitty things going to do in cutting wood?

These 'little bitty things' are the control system. They replace the
expensive, heavy, high-precision components that we use today by
substituting active control for the passive systems based on weight of
material and mechanical precision.

Let's take a kindergarten example: An actively controlled fence. The
fence itself will consist of a strip of thin aluminium backed by an
array of actuators and the whole assembly is mounted to the saw guides
by not-very-accurate mounts. The actuators deform the aluminium in
response to signals from the sensors, mediated by the processors.

The fence actuators can be a strip array, like the array of LEDs in my
$100 Brother printer. They won't be much more complicated and in all
probability they'll be a lot cheaper. In addition there will be
another network of sensors to check the the distance of the fence from
the cutting element and their parallelism.

Mechanically, the 'fence' will be a cheap, low-tolerance, device, more
cheaply constructed than any Harbor Freight special. It will be sturdy
enough to stand up to shop use, but not much more. The mechanical
parts will cost only a few dollars.

The magic is in the active elements. The sensor array will constantly
track the movement of the wood, the cutting line and various other
factors such as temperature at the cutting interface and the cutting
speed and well as distance, parallelism, etc. And of course the
fence's processor(s)

Let's say you want to rip a 6" board. You crank your 'fence' over to
6" indicated. The tolerances will be loose as a goose, but you don't
care. The device will tell you when you're close enough, parallel
enough, etc.

Now, turn on the saw and start pushing the wood through. As the
sensors detect the cutting position, the actuators in the fence will
deform the aluminium strip to steer the wood exactly where it needs to
go. It won't need to move it very far because the fence helped you
line things up with sufficent precision before you started. The cutter
will contact the wood at precisely the right point on the right angle
to produce the cut you need. Accuracy is likely to be measured in
hundredths of an inch because that's sufficent for woodworking.

Now please note this is NOT a description of the kind of tools I have
been talking about. It's another one of those proof of principle
devices you seem to have so much trouble grasping -- albeit a more
advanced one. It is simply an example to demonstrate how these
technologies could be applied.


Given the way semiconductor fabrication works -- and given the
differences between MEMS devices and things like microprocessors or
DRAMs -- the prices of these devices will be extremely low in volume
production. And of course it's unlikely that most of the sensors and
actuators will be designed specifically for woodworking tools. They'll
be adapted from devices used in higher production devices.

Not devices big enough to do what you are claiming.

Wrong again. You're hung up on the idea that the whole thing will be
active.

I also don't think you grasp what I mean by 'cheap'. The active
elements in these devices are going to cost on the order of what a
transistor costs in a modern microprocessor -- for exactly the same
reasons. Each tool will contain a lot of them, but the the resulting
cost will still be very low.

Huh. So how will having 40 million tiny machines on a lump of silicon a
half inch square do anything useful in the way of cutting wood?

They're not going to be limited to a 1/2" square bit of silicon. Take
those 40 million devices, spread them out over several square feet
supported by an appropriately design mechanism you get something very
useful for cutting wood.

I know "sensors and actuators". And we're back to "what are you going to
actuate with the minuscule amount of force that such a small device can
produce that is going to be useful in woodworking?

The essence of a modern control system of nearly any sort is using a
combination of intelligence, sensors and relatively low powered
actuators to control larger forces. We do it every day, although
generally on a larger scale today.

Really, before you make these wild claims you should try to at least _think_
about how what you claim will be accomplished will actually be
accomplished.

Someone is definitely not thinking there. You've made that painfully
obvious in this message.

--RC


You can tell a really good idea by the enemies it makes

You can tell a really good idea by the enemies it makes

You can tell a really good idea by the enemies it makes