Friday, February 8, 2019

Yield Rates

The cost of an electronic device is often related to yield rates.  What are yield rates and how do they relate to your personal life?

(Note:  If you thought this was going to be about bond yield rates, sorry!)

Mark is back in the studio, making chickens.   His tenure with ceramic birdhouses may be about up - everyone else is making those, now, and the market is saturated.   Of course, this might not have happened if he hadn't put on workshops on how to make ceramic birdhouses - talk about giving away your trade secrets!   But he's that kind of guy - giving.  And besides, you can't just make the same thing forever.  So now it's chickens - they've come home to roost.

Making things from clay is a multi-step process.  And at each step along the way, there is a chance the entire thing could go bust.   You sculpt or form or throw something from clay.   You may decide that what you made is ugly, and you smash it and throw it in a bucket, to be later recycled using a pugging machine, into fresh clay for something else.  Odds are, you will crush at least one in ten of your creations as not being up to snuff - particularly as you get better at it and have higher standards for your work.  The work at this point may be called leatherware.

But assuming you like what you created, you have to let it dry out before the initial firing.  And the drying process is complex.  If it dries too fast, it may crack - a problem in particular for larger pieces.  And once it cracks, you have to toss it - it won't heal itself in the firing process.   So at the drying stage you may lose one in ten of the objects d'art that you created.  The dried pieces may be called greenware.

Next, to the kiln, for a lower-temperature bisque firing.  Granted, in some cases, you might go directly to glaze firing, but that's a whole different story.  Bisquing the clay causes it to harden - fusing the silica in the clays to form a hardened terra-cotta like form known as bisqueware.   At this stage, a lot can happen.  If there are any air bubbles in the clay, the item can blow up in the kiln, destroying itself, as well as adjacent pieces (which will make you lots of friends in the pottery guild!).   Or it just may crack, and cracked pots are not deemed to be worthwhile to show or sell - many a potter has a yard full of such rejects.   Again, you might lost one in ten at this stage as well.

Now to glazing.   Glazing can get tricky - and it can be heartbreaking.   You make this great thing, you fire, it, and it looks great.  Then you glaze it, put it in the kiln, and high-fire it, and it comes out cracked, broken, melted, or just looking like baby-shit brown.   High temperature (Cone 6 and higher) glazes get tricky - they often are designed to "break" and form multiple layers of colors as the glaze runs down the side of the piece.  Put the glaze on too thick, and it runs down onto the kiln furniture (which pisses off the kiln master).  Again, such pieces are generally not deemed worthy of show or sale, if they have drips of glaze coming off them.   Put the glaze on too thin, and the bisque shows through.   It takes a lot of trial-and-error to get it just right.

And then the color can come out bad.  It is supposed to be green, but comes out a burnt brown.  The kiln-master used his own crazy ideas about cool-down times and caused the glaze to burn.  Or maybe a student dunked a pot in the glaze bucket and contaminated it with another color glaze.  Or maybe the glaze was not stirred up sufficiently.   There are so many ways to screw this up, many of which are not your fault.  You put something in a kiln and cross your fingers.   A lot of amazing pieces come out more to luck than to skill.

But if you find a glaze that works well for you - and you get used to using it - you tend to stick with it.  Using different glazes requires you re-learn the process (with the corresponding high scrap rate) - which is why when you visit a potter's studio, you tend to notice they use one or two colors of glaze almost exclusively.   You go with what works and what you know.

So at the glazing step you might lose another 10-20% of your work, although once in a great while, you might be able to re-glaze and re-fire a piece and salvage something.

If you add up all the losses at each step of the game, you can see that as much as 50% of what you make ends up as scrap.  And that can be very discouraging, particularly for beginners (where the scrap rate is much higher).  Mark thinks he manages to yield about 70% of what he starts with, which is a pretty high number, all things considered.  There is a pile of broken and cracked and rejected pieces out back of his studio, though.  Most studios have them.

In the semiconductor business, we call this yield rate.   When you make a mico-chip or a flat-screen television with thousands if not millions of transistors on it, there is a finite probability that one or more of these millions of microstructures may have flaws in them.  As a result, particularly early on in the business, you end up with phenomenal scrap rates - as high as 90% - with a corresponding yield rate of 10% or so.  Sometimes it is even worse.

When flat-screen televisions first came out, the yield rates were pathetic.  For every 10 television panels they would make, maybe only one would be worthy of sale.  The rest were scrapped, which is why flat-screen televisions cost so much back then - in the thousands of dollars.    Prices came down once production techniques improved and the yield rate increased.  But even today, in the semiconductor business, a lot of chips get culled from production if only a few of those millions of transistors on the chip are defective.  Yield rates are never 100% or even close, in most cases.

For microprocessors, the problem is even more complex.   Did you ever wonder why they sell the same microprocessor with different clock speeds?   It is not that they designed the processor to run at different speeds, but rather they build them all the same, and after they are done, they test them to see how fast they can run without undue errors.  Each chip is then tested to see what speed it can reliably run at and is sold as being rated for that speed.   Computer buffs and gamers know about this, and often try to save money by purchasing the slower processor and "overclocking it" to make it run faster.  The trade-off, of course, is that it may crash more often.

In other chips, if parts are defective, they simply change the designation of the chip.   I used to work with Intel's 8032 processor, which came in a number of flavors and was used for industrial controls.   In one version, it had 8K of onboard memory.  Another version had no onboard memory.  The difference was only that the non-memory version had a flaw in the memory from production, so they just sold it as a non-memory version, and you provided your own, off-chip memory.  So even if something fails, it can succeed.  You just say, "I meant to do that!" and no one is the wiser.

Yield rate is an interesting phenomenon.  Can you imagine if other products were made in the same genre?   Suppose GM had to scrap every 10th car that came off the line because there was one tiny defect in it?   It would certainly drive up the cost of cars!   Or imagine that every car was tested and rated for how fast it would go - and each car would achieve a different speed?  They used to do that, with Duesenbergs  - not one of them leaving the factory unless it could do at least 100 mph.

With regard to the latter, this does - or used to - occur to some extent with car engines.   If you got a car with a well-balanced engine (the pistons all weighed the same, the crank was balanced, etc.) and the valves were well-seated and the valve-timing spot on, it might generate more horsepower than it was rated for, and run smoothly.  On the other hand, if you got a Friday engine - built by someone in a hurry who cut corners - it might run rough and never achieve its rated horsepower.   Today, standards are a little higher, and variations in engine output are far less.   But there still is a bell-curve distribution of performance, nevertheless.

In your personal life, yield rates are also important.  As I have noted before, my estimation as to the "efficiency" of the human being is on the order of a few percentage points, if that.   Think about it, we waste hours every day eating, sleeping, defecating, and watching television (some people do little more than that!).  The few hours we actually "work" are mostly wasted with social grooming, taking coffee breaks, and doing online shopping (or reading this blog - shame on you!  Get back to work!).

Not only are we horribly inefficient, but most of what we do ends up on the scrap heap - or at least a substantial portion of it.   Failure is the norm, but that doesn't mean your efforts are fruitless.   In the early days of flat-screen televisions, nearly every one made was a failure.   But the few that did work would command hefty prices in the marketplace.   You could look at the pile of junked televisions on the scrap dock and be depressed, or look at the shiny products that did make it to the showroom and be elated.   The choice is yours.

Baseball is often seen as a good metaphor for life.  Unlike football, where players are exhorted to "give 110%" (which is physically impossible) or told that "losing is not an option" (when in fact it is a predictable outcome), in baseball, failure is part and parcel of the game.   Even the best of hitters would struggle to bat .500 - hitting 50% of the balls thrown to them.   And teams can lose game after game - against the same team - and still make it to the playoffs and even win the world series.  Whoever invented baseball knew more about real life than whoever invented football.   Baseball is all about coming back after defeat.   In football, it seems, there are no second chapters, for the most part.

In your life, you will make a colossal number of errors.  You can dwell on them and be depressed, or learn from them and move on.  And if you realize that the one or two things you did do right have paid off, well, your attitude toward life may improve considerably.  Even if you fail 90% of the time, that 10% success rate can blow those failures out of the water.

Not every investment you make will produce results.  In fact, many will fail utterly.  You will buy stock in a company that goes bankrupt - I have, more than once.   But other times, you will invest in something and it will do well.   It might just be a mutual fund that doubles your money after a decade or so.   And while that is not rags-to-riches success, it is something pretty damn good.

Predicating success on being successful all the time is one sure way to fail, however.  When I started investing early on, I had naive ideas that everything I would invest in would always make money and moreover, have fantastic rates of return.   I was going to be the smart guy who latched onto "the next big thing!" and made millions - or so I thought.   I mean, after all, such people do exist, don't they?    Why can't I be one of them?

Then I realized that the folks who make the mega-bucks are either really, really lucky, or have mega-bucks to begin with, and can manipulate markets - often by selling crappy investments to wanna-bes like me.  The harder I tried to "win every time" the more I ended up losing, ironically - or was it predictably?

Your yield rate in life is never going to be 100% or even very close to it.   We are all human and make mistakes and do dumb things.  You can obsess about this or just learn to live with it.  Put away thinking about what could have been or what went wrong, and look to the successes you have had.

Because odds are, they are pretty damn good successes!