Silicon Valley Redneck

Losing sight of the numbering scheme

One of the nice features of the 74*00 line of logic is that, for the most part, the function is consistent across technologies.

But...

I find myself in need of some open-collector drivers to fit in a small space.  The outputs may be subject to significant voltages in the OFF state.

A 7407 fills the bill nicely.  Well, apart from the "small space" aspect.  And, er, the 0 to 70°C rating.

The signature feature of the 7407 is the high-voltage outputs, rated for 30V.  If you only need 15V outputs, there's the 7417.  I don't recall offhand whether there was a 5V-output counterpart, and don't feel like digging out the ancient TI databook just now.

So: what's a modern version of a 7407 (hey, even a 7417 would do), with at least an I-temp rating, and available in a TSSOP?

Well... I find things like 74LV07, and 74AUC07.

Which, well... when I look at the absolute maximum ratings, I see an output voltage rating that's the same as the supply voltage rating.

Not very '07 at all, really.

A bit of further digging reveals that the 74LS07 is still around, has 30V outputs, and is available in a TSSOP... but the 54LS07, which would have met the temperature requirement, is obsolete.

Grrr.

Ah, well.  More digging to do.  If all else fails, use discrete FETs in tiny packages....

Another point of worry

Fabius Maximus notes, as others have, the shift in education statistics, with women now greatly outnumbering men in college.

I noticed something a couple of months ago that might be related to this trend.

I looked around the "young adults" section of a large chain bookstore.

About half the books were the latest vampire-and-werewolf romances from Crepuscular Cloneworks.

Overall, at least 90% of the books were obviously aimed at girls.

There were a few token boy books, and a few more that, at first glance, might have been gender-neutral.  But the rest?  The covers just screamed out "girl stuff."

So, is no one publishing books for boys because boys don't read, or are boys not reading because there aren't many books written for them?  Chicken, or egg?

Gosh, a real miracle drug!

From this gee-whiz article on statins:

The cholesterol-busting drugs also reduce the chances of death from all causes by 12 per cent.

All causes?  Wow!

And yet, I see two ways of parsing that.  Does it mean that, thanks to those daily pills, I'm 12% less likely to die in a car crash, 12% less likely to be killed by a meteorite, and so forth?  Or does it mean that there's a 12% chance that I'm immortal (so long as I keep taking them)?

Where did that come from?

When I went for my morning walkies at Rancho San Antonio on Monday, something was different.  There... in the big meadow in the county park... by the picnic area...

It wasn't there on Friday.  So what happened?  Did somebody plant a monolith seed on the solstice, and it took a week to sprout?  How big is it gonna get?

I have an urge to dance around it, waving a bone.  Or, perhaps, to invite a bunch of people to show up in monkey suits....

There's no inscription.  Yet.  Will one appear tomorrow?  Is this destined to be California's headstone, when the absolute drop-dead must-have-a-budget date can't be put off any longer?

First chipmunks, now unexpected standing stones.  What next, crop circles in the oak trees?

Just what the world needs!

A blog dedicated to the greengrocer's apo'strophe.  (Via Insty.)

RAM. We gots RAM.

Box from Newegg arrived this morning.  Oh, goody!  Unpack 2x 2GByte Corsair RAM modules.  Plug into workstation, on top of the 2 x 1GByte modules that were already in there.  Reboot.  BIOS sees 6G.  Linux sees 2.7G.  Er...

OK, rebuild the kernel with the "more than 4G of physical memory" option set and reboot again.  Re-install the Nvidia driver to go with it.  And...

Yep.  Six flippin' gigabytes of RAM.

Well, I tend to leave the system running for months at a time, with many programs spread across 14 virtual desktops, and some of those (Firefox, Acrobat, and such) are cumulative memory hogs.  And here I am getting ready to route some fine-pitch, multi-layer boards with EAGLE, which gets memory-hungry in a hurry when you start using a 0.001" grid for the autorouter.

And, after all, the extra 4 gig cost me less than I paid for my first kilobyte of RAM, back in the day.  (Eight 2102s, at many dollars each - I had to mow a lot of lawns to pay for that.)

Now to see how long it takes for this system to go into paging again....

Update: well, after a busy afternoon, memory usage stands at 2.2G, so it would have been paging by now without the upgrade.  (And the screen-blanker is the first thing to get paged out, and it doesn't get paged back in, so once paging starts, my monitor stays lit up unless I turn it off manually.)

What a difference 21,290 miles make

About $1300, in this particular instance.

A couple of times in the past, the Prius had exhibited a Known Issue: the engine would start up a bit too slowly on a cold morning, and the computer would declare a timeout and light up all the panic alerts.

On those occasions, the code somehow always cleared before I got it to the dealership.

Well, this warm morning, with the engine warm from earlier driving, it happened again, and it didn't clear.  So, time to get it fixed....

Now, all it really needs is a parameter tweaked, to accommodate the slower-than-expected startup of a cold or aging engine.  Alas, the ECU in the NHW11 doesn't accommodate field firmware upgrades, so the solution to error code P0300 ("random multiple misfire") is to replace the entire freakin' engine ECU with one that has the new version of the firmware.

And it couldn't, naturally, have done this repeatably while it was under warranty.  Oh, no.  It had to wait until the warranty had unmistakably expired before it would do that trick for the service department.

Oh, well.  It's back to normal operation now, and I expect to keep it for another few years.  I don't think there are any further Known Issues with the NHW11, aside from early failure of the auxiliary battery... which hasn't happened to me yet, in the 7 years I've had the car.

More parts bite the dust

I just came up with a totally cute use for an LM3900.

Alas, the 3900, along with its I-temp sibling the 2900, was discontinued a decade ago.  Besides which, what I'd really want would be an LM1900, which, as far as I can tell, never existed in the first place.

Well, maybe an LM159 would do the trick.  Except, er, only the LM359 is still around, and C-temp parts just don't cut it.  And it's not like the quantity is large enough to get National to make (or qualify, or whatever) a run of 159s.  And I don't want to design in a part whose future availability is suspect, anyway.

Oh, well... I wanted it for, basically, a cuteness feature (with some practical application in troubleshooting).  A little more looking around for suitable parts may yet be in order, but it's not a disaster if I have to drop the feature.

Well, isn't that special.

The One is trying to take unofficial credit for inspiring the protests in Iran?

And now, of course, the protesters, having supposedly been motivated by his noble rhetoric, are on their own.

Bay of Pigs?  '91 Shia uprising?

Projecting weakness, anyone?

Yes, by all means, rise up against your rulers who are our enemies.  Just don't expect us to stand by you in any meaningful way when you do it.

Or, this time, even in symbolic ways.  Cultivating good relations with the tyrants is so much more important, you understand.  They're people who matter, and who serve caviar.

Verrrry interesting

It seems that a pair of Columbia U. students have analyzed the Iranian election results and found evidence of human-generated random numbers.

A study at MIT back in the 70s (OK, so it was a few students chatting in a hallway) determined that 17 is the most random number, 137 is also a highly random number, and that, in general, numbers consisting of odd digits and ending in 7 are random.

And what did this new analysis find?  Yes, a disproportionate percentage of the vote counts ended in 7.  In fact, 17% of the results ended in 7... which suggests meta-randomness.  Further, only 4% of the counts ended in 5; it is well known that numbers ending in 5 are not random.

So, too many 7s and not enough 5s could be an indication that the numbers were chosen to look random rather than made up... as opposed to actually being random down at the last digit.