Plans2Reality

Efficiency Doesn't Matter

Not for mass-market solar panels, anyway.

The goal of mass-market solar, otherwise known as grid-tied PV, is to generate electricity for the grid, so we don't have to burn as much coal, oil, etc. It's a fine idea.

But anyone in the fild will tell you that PV electricity is still a lot more expensive than coal-fired electricity, even when the sun is shining - never mind darkness. (And even if you factor in a pretty high carbon tax. Coal is cheap.)

Therefore, the goal is to get the cost of PV electricity down as far as possible - ideally as cheap as coal, even cheaper if possible.

Now, electricty is sold in units of kilowatt-hours (kWh). So we need to get the cost of PV, in terms of dollars per kWh, as low as possible. It doesn't matter how EFFICIENT a PV panel is, only how much it costs. Sunlight (unlike coal) is free, so it really doesn't matter if a panel is 10%, 15%, or 20% efficient - not if such efficiency comes at a cost.

There's nothing _wrong_ with efficiency, per se. And I am sure the boffins will find ways to improve it. But anyone who asks what a panel's efficiency is doens't understand the problem. The only meaningful question is, what's the cost per kilowatt-hour?

Fix your Mac at iResQ (iPhones and Pods, too)

I had a sick Mac Mini - it would not boot. I sent it to iResQ and they fixed it and had it back, all in less than a week, and less than Apple was going to charge.

iResQ sends you a shipping container. You put your computer in and send it back. They examine, call you, and you approve the repair. It's dead-bang simple, really.

So, five stars, two thumbs up, etc to iResQ!

They also fix iPhones and iPods, though I have not used that service.

Solar Panels: Why Quote the Wrong Spec?

Here at Plans2Reality World Headquarters, we have been looking at solar panels for electricity. This has provided an interesting and instructive lesson in marketing along the way. I shall christen it:

Bragging About the Wrong Thing

Solar panel makers love to brag about two things: their watt-peak (Wp) rating and their efficiency. Oddly enough, neither one matters. Nope, not a bit. Not one d*mn bit at all. 

(Let me interject here that _all_ panel makers are equally involved. In private conversation, many have told me they agree with the analysis I am about to present.)

First, Wp. The peak-watt rating is a measurement of instantaneous power output from the panel, taken under an incident light intensity of 1000 watts per square meter, and at a cell temperature of 25C.

So what's wrong with this? The panel will never actually see these conditions in the real world, and there is poor correlation between these "STC" ratings and real-world performance.

1000W/m2 does approximate the noon-day sun, but most of the time it isn't noon. Furthermore, most solar panels are NOT installed in sun-tracking mounting systems, so at best the panel "sees" full sun only at noon. The oblique angle the rest of the day reduces incident light. (Actually, most smaller arrays are installed so the panels aren't at the optimum angle even at noon - residential rooftops are usually not at the correct angle, but it's easier, and nicer-looking, to mount them flat against the roof.)

So your panel will never generate its Wp rating, due to sun angle. But wait, there's more!

PV panels are quite temperature-sensitive - some more than others, but all are to a fair degree. And when you place an object in the sun, it gets warm - warmer than 25C most of the time.

How much warmer? Well, That Depends.

The exact amount of heat gain depends on the specific PV material. Some panels will get warmer than others. In addition, the degree to which the panel cools off, that is, sheds heat via conduction and convection to the environment, varies substantially among panels. There have been attempts to quanity this (NOCT, PTC) but these are at best compromises.

Net-net of all this: the Wp rating of the panel represents a power output you will never see in real life. Couldn't be much more meaningless than that, huh?

So what specification SHOULD we be talking about? 

We should be talking about kilowatt-hours per dollar. Why? The answer is quite simple: Solar panels are installed to generate electrical energy. Electrical energy is measured - and sold - in units of kilowatt-hours. So why not measure our PV installation the same way the electric utility measures the electricity it sells us?

It's just that simple. Really.  I am putting up PV panels so I don't have to buy kilowatt-hours from the utility. Great. So tell me how much the kilowatt-hours from the PV panels will cost.

Nothing else matters, folks. Nothing.

Next installment: Efficiency doesn't matter either.  ;-)

The Wonders of Time Machine

I've been using Apple's Time Machne for many months now. It is indeed invisible in normal operation. I finally got a chance to test in. I had a corrupted file, so, off to Time Machine!

Couldn't be easier. Flip back thru time 'til you get to a date & time you want, then grab the files.

The more I know, the more I love my Mac.

Moore's Law: Still Not Dead

This time around, it's iSuppli presiding over the funeral. (via the Register: LINK) As is often the case with negative predictions of many types, it will be proven both wrong and right. Wrong, because Moore's Law is not dead yet, but right, because they have identified a real problem.

Heck, there's _alway_ a real problem. Usually several. When Moore coined his law in 1965, you could just about measure IC features sizes with a ruler and your eye. Masks were hand-cut in rubylith. The industry has had to solve many serious and significant problems - some directly relating to making wafers; others related to design. How the heck to you design even a one-million transistor circuit? Let alone test it....

Design tools and mask charges are huge - and iSuppli is correct that they cannot get much "huger". Great; so somebody will invent something. Design tool companies will not make any money selling tools if no one can afford to buy them. Maybe they'll rent. Maybe it's Software-as-a-Service. Heck, if I knew the answer I wouldn't be blogging, I'd be doing.

But somebody out there does know the answer. I wonder who?

A Decent DNS Tool

Mac OS X 10.4 and newer versions have a GUI for configuring the DNS. It appears not to suck. I was actually quite impressed with its ease. To be fair, it is a bit strait-jacket-y (is that a word) in that it has some very definite ideas on how config files should be named, and it definitely seems to force you to having exactly one name associated via an A record with an IP; all other names must be CNAME. (I think; I am still learning.)

But it works, and there is much to be said for that.

At Long Last, a Mail Upgrade

Over the weekend, I upgraded my mail and web server from an ancient (by computer standards) G3 tower running OS X 10.2 Server to an Intel Mac Mini running 10.5 (for mail) and a G4 Mini running 10.4 Server (for DNS, Web, and File Sharing).

My, the new CommuniGate mail system is nice. There are a few installation tricks, but I managed to get it all sorted out and running in a few hours. I do wish there was a proper, searchable (or indexed) reference manual. CommuniGate offers on-line (web) help screens, but there are dozens, and it's tricky to be able to look at the help screen and the UI at the same time. But overall it is a powerful and stable package. My last version ran trouble free for seven-plus years.

The new one intergrates VoIP and other unified messaging. Can't wait to try that out. It's fully compatible with all that Microsoft Exchange stuff, too.

Spring

The season and the mechanical device. I can't quite beleive I haven't posted since February. After a slow winter, business activity has increased. It seems as though there was a sort of pause last fall, waiting for the other shoe (or bank) to drop. Now, while we are by no means out of the woods, business activity is increasing again.

So, here at Plans2Reality we are springing forward as well. We've expanded our offices and our service offerings. There's increased emphasis on what I call 'Driving the Design Win Process' as we help our clients close the technology design wins they need to drive sales.

Got Design Wins?

Loving my iPhone more and more....

Don't have one? Get one.

If nothing else, the ability to control your music server (i.e. iTunes) from anywhere in the house is wonderful. We have speakers in all the rooms, and can have whatever sort of music wherever we want it. It's wonderful!

So, buy an iPhone. You can make calls with it, too.

Moore's Law Conquers Television

This is not really news - it happened a few years ago, although it is hard to pick a specific date.

TVs have gone fully digital, and I don't mean the broadcast format. Even today's TVs deal with analog broadcast by digitizing it and using DSPs and dedicated digital hardware.

Furthermore, with the takeover of LCD and plasma displays, the display itself is now subject to Moore's Law. (Gordon never said he was referring only to silicon, after all.)

What does this mean for the industry?

First of all, because TV is a high-volume market, it means that the larger chip companies will dominate it. Cost of silicon will be the key driver, and the bigs guys with big fabs always win that war.

Second, it means that TV will become a client-server architecture. One could say that TV has always been client-server, and that broadcast radio and TV invented the concept, so to speak, but client-server in the computer context has a different meaning. TV will consist of different types of servers offering content to a wider range of clients. Over-the-air broadcast won't go away, but numerous types of Internet delivery will become more common. Hulu, Apple's iTunes, and NetFlix on-demand are just part of it.

TV isn't going away. But it's going to look a lot more different in five or ten years than it does today. Moore's Law is at work.