Silicon Valley Redneck

Ugh.

Just in time for the long weekend (not that weekends, per se, really figure into my schedule), I woke up with a case of the cruds.

Doesn't seem to be a case of swamp fever, transmitted by them pesky skeeterbugs of a few nights ago, nor yet plague from Wednesday morning's murine menace.

For one thing, I don't have a fever.  Actually, I have more of a negafever, with my body temperature remaining obstinately at its nighttime setting.  This typically happens if I'm awakened from deep sleep by some loud noise, such as an alarm clock or telephone, but every few years I'll have one of these episodes of my thermostat sticking and just not going into day mode.

And so I'm shambling around, zombie-like, with a mild case of hypothermia, resulting in dizziness if I move incautiously, plus an unhappy stomach and whatnot.  Oh, and seriously impaired judgment, though not impaired enough for me to try driving in this condition.

On top of which, my freezer is devoid of Bird's Eye Frozen Braaaaains.

The things you see when you haven't got a proper camera

Out walkies at Rancho San Antonio yesterday morning, I noticed a fuzzball ahead, alongside the road before the farm.  As I'd left my camera back in the car, all I got were some mobe snaps.

I'm not much of an authority on matters murine, but that right there is either a dwarf long-tailed chinchilla or a dusky-footed woodrat, no?

It was scrabbling in the dirt with its forepaws, and apparently finding tiny things to nibble on - seeds, perhaps, or the remnants of a trodden-on Frito.

Looky them whiskers!

At one point, the critter wandered over and sniffed at my toes, whereupon I cautioned it against running up my leg.  Wood rat, roof rat, or regular old plague rat, loitering by daylight in an area where people are traipsing through just ain't normal behavior for a rodent that isn't someone's pet, and I didn't feel like taking chances with whatever might be afflicting it.

Update: scale is not clear from the photos; it might be mistaken for some sort of blunt-nosed mouse.  It was definitely rat-sized, and a fair-sized rat at that, though it was a great exaggeration for a passing blonde to say it was as big as a dog.  (Her perceptions were perhaps affected by the lack of a table to stand on and scream.  Or, maybe, she has one of those teeny-tiny purse dogs for reference.)

Pattern recognition

Ordered a VariCAD upgrade several minutes ago, and the phone hasn't rung yet.

I guess BofA's fraud-prevention algorithms have realized that, yes, I do occasionally make payments to a Czech company via a Dutch payment-processing service.

(Those yearly upgrades have had about as much price creep as a local newspaper subscription over the past few years, i.e., a whole bunch.  Still, I continue to use the product for business purposes, and expect to make more use of it in the coming year, and the upgrades have in fact been worthwhile in the past... though I haven't gotten around to installing the latest, nor indeed anything in the past few months; perhaps I should do that.)

This really gets my feckles up

Headline at NY Sun, via Insty:

‘Second American Century’ Hoves Into View As Ann Romney Steals the Show at Tampa

Aaaargh!

"Hove" is the past tense of "heave."  Attempting to form the third-person singular present tense of a past-tense verb makes no sense.

(Hmph.  It appears that, Kryten's remarks notwithstanding, "feckles" has become a word on the Internet, though it still doesn't mean what the Cat thought it meant.)

Mosquitoes from Mordor!

Ugh.  Haven't been sleeping well lately, what with one thing and another.

Last night: anotherer thing.  Mosquitoes in the bedroom.  Uncommonly nasty ones.  Not much buzz, but... well, I ended up crawling out of bed around 0300 because I wasn't sleeping anyway; at that point, I had a blotchy itchy rash hither and yon.

A couple of hours later, the rash blotches have dwindled to standard-issue mosquito-bite bumps, though I don't normally have even this much reaction to mosquito bites, never mind what I had when I got up.

Guess I'll find some time this afternoon to spray the outside walls, fences, and shady spots with Mordorthion (well, a residual spray that claims to work on mosquitoes for 2 months), and the bedroom with some sort of indoor bug spray, probably the kitchen stuff.

How to tell metals apart

Have a few-off PCB build in the works, using Real Solder (with genuine lead).

Assembly house wants to know if the board finish is tin-lead or silver, presumably for the non-ROHS paperwork.  ("It might be contaminated by ROHS materials!")  The PCB fab house didn't provide that info in the paperwork for the order.

Surely there's some simple test.  Like, rub it with a poisonous mushroom and see if it changes color?

The real trickle-down economics

Spotted on Yahoo:

The Federal Reserve needs to do more to keep money flowing through the economy because the recovery hasn't been forceful enough, Chicago Fed President Charles Evans, a known backer of additional easing, indicated Monday.

Yeah.

Dude, the high-level capital markets are saturated.  Governments and large financial institutions are already borrowing at negative real interest rates.  Banks don't want more money - or at least they don't want it enough to pay more than a pathetic fraction of a percent interest on deposits.

In aeronautics, this is called a stall, and continuing to haul back on the stick is called pilot error.

In politics, allowing people to keep more of their earnings in the hope that they'll spend it on things they want is sneeringly called "trickle-down economics."

What then shall we call a policy of shovelling money into governments and banks, in the hope that it will somehow end up on Main Street?  And what shall we call it when its only results are to drive up our collective debt and drive down the value of our retirement savings, and its proponents insist that the only possible course of action is more of the same?

Home repairs

'Tis a weekend for fixing broken stuff.

Mostly accomplished: fixing the living-room computer, as drives the home-theater projector.  It's a 2005-vintage small-form-factor thing, with a 1.8 GHz Semprini Sempron, and nice features like many-channel audio and an optical audio output.

Back in '09, the hard drive started going flakey.  By that time, compatible drives (original SATA, <640 GB or some such) had largely disappeared, but the drive started working again while I was fiddling with it, and continued working until a few weeks ago, when it started giving great flurries of read errors.

Well, in the meantime I'd replaced my old server, freeing up a 160 GB classic-SATA drive.  This left only the small matter of a suitable Linux install.

The original, 2005 setup had a rather brittle configuration of Slackware, which was carefully nailed down with support for a TV tuner card and one of those ATI/X-10 RF remotes.  The brittleness had become a bit of an issue, what with lack of support for some of the newer codecs.  But!  That 2005-vintage analog TV tuner is obsolete anyway, so no need to support it.

Try installing generic Debian-stable (Squeeze).  It installs, but: KDE doesn't work (croaks partway through startup), and the sound system seems to be broken (as in, Xine and other players will sometimes manage to get sound to the speakers, but usually not, and it comes and goes during a playback session).  Presumably something to do with the specific hardware, given that Squeeze works fine (KDE and all, not sure about sound) on my lab machine and on a VM on my workstation.

Next up: XBMC/Ubuntu.  Installs fine.  Comes up with the big shiny UI... and brings the CPU to its knees, or maybe even its elbows.  Doesn't have enough cycles to handle mouse movement.  Totally unusable on this machine.

Okay, so: try the latest Kubuntu, which turns out to be 12.04.1.  Installs fine.  KDE runs fine.  Sound works, and it doesn't take me too long to find the KDE settings thingy for choosing a sound profile with digital output.  Xine works.  There's a script provided for installing, um, a certain illicit library that could be handy on a multimedia machine.  Also, there's an easy setup for lirc, which (after a couple of tries) gets it working with my ATI remote and doing reasonable things for Xine - haven't brought in my old irexec configuration yet.  (The pause button doesn't work, but that's because the physical button is apparently worn out.)

Button it all up and stuff it back in the stereo cabinet: there's one job done.

Next up, the laundry tub.  My second cheap plastic laundry tub having come down with a severe case of the crumblies, and now having cracks below the waterline even for careful use, I'd rounded up a sturdier-model replacement.

Get it put together, disconnect the old one, go to install the new one: oops.  Cut the tailpipe 3/4" short instead of 3/4" long.  Lacking a way to saw an inch and a half back onto it, back to the store for a suitable start-over tailpipe.

Oops II: the feed lines provided are for 1/4" NPT inlet, or something along those lines.  I have 1/2" NPT connections (on a rusty kludge installed by the last guy, and it's rusty on account of being iron pipes connected to a copper system, so I really ought to replace it, so...).  Back to the store in search of adaptation.  Don't find what I'd been visualizing, and brass fittings have gotten expensive, but: come home with a pair of 1/2" NPT feed lines.

Proceed to install and connect the sink.

Oops III: when I open the cold-water valve, the iron kludge springs a leak.  Upstream of the valve.  Grrr: at minimum, need to unscrew the kludge and put some fresh tape on it.

Oops IV: the shutoff valve outside my front door no longer shuts off the water to any meaningful extent, which makes DIY leak-fixing problematic.

So, I call the $$$unday plumber to come deal with the shutoff valve.  Apparently this is a busy Sunday, but somebody should be able to come out between 3 and 7.

Now awaiting a call back from the plumbing company.

Oh, it all makes work for the Working Man to do....

Update: the plumber came (5:30ish), he saw, he didn't have the parts to fix the valve; apparently the dispatcher hadn't believed me when I said the shutoff valve wasn't closing.  However: he managed to get the water turned off at the curb for a few minutes (I've never managed to turn one of those ferschlugginer valves, but keep thinking I should acquire or fabricate a suitable wrench) and we got a temporary fix in place for the lundry tub, which is now usable.  He's supposed to return tomorrow (giving me a half-hour warning on my cellphone, in case I'm at a client site) to deal with the shutoff valve... after which, I'll be able to address the garage plumbing situation properly at my leisure (though sometime fairly soon, 'cause some of that iron pipe is rusted nearly through, and maybe the leak was actually a hole in the nipple).  Oh, and the new faucet's aerator promptly clogged solid with rust and needed cleaning.

Update 2: plumber's colleague showed up Monday morning and replaced the old shutoff valve with a shiny new ball valve.  Maybe this coming weekend will be time to remove the last of the iron and put in nothing but proper brass fittings.  Mmmm... mayhap I should check real industrial suppliers for a better selection of fittings, too?

Reflected sound of underground spirits

The kid is off at some sort of Bavarian Illuminati Secret Training Camp for a couple weeks, in an undisclosed location in the Swiss Alps.

It occurs to me that she might encounter Gnomes there.  What with her not having picked a career path yet, might she decide to become a Gnome?

This leads to thoughts.

First of all, I'm pretty sure no university this side of the puddle offers a program in Gnomeology.  Economics, yes, but that's hardly the same thing - more like philosophy sprinkled with math to make it look scientific (and the math being heavily dependent on assumptions, which come from the philosophy or (after graduation) from wishful thinking on the part of Upper Management).

Then, I realize that none of the U.S. financial institutions employ Gnomes anyway.  Gnomes get in the way of the modern way of doing Big Finance, which involves a lot of getting in front of the trend and not much analysis of whether the trend makes any sense, nor how much it resembles, e.g., a trend from the 1630s that turned into a huge fiasco then.

So, probably not a career option.  Which is a shame, because we really need to put the Gnomes back in charge of the financial institutions.

Amazing how misdesigned things become standard

I mentioned as how the ARM SSP (synchronous serial port, for SPI and suchlike) lacks an "operation entirely complete" interrupt?

Well!

Turns out that, as an SPI slave (at least as observed in the NXP LPC13xx), it insists on having SSEL deasserted after each and every frame (byte).

This is severely annoying.

I'd observed long ago that SSEL in master mode was per frame, not per block, so the AGROS driver, by default, simply ignores the availability of SSEL and uses a GPIO pin as a per-block select, as required by most SPI peripherals, such as chained drivers, SPI memory devices, and so on.

The corresponding limitation in slave mode is seriously crippling, if one is trying to work with a master that uses SSEL to define a block transfer.  (The Aardvark does this.  I'm still waiting for the detailed specs on the gadget I need to work with.)

Oh, and the ARM implementation doesn't work in a daisy-chained setting either, so that's no justification.  Set the frame length to 8 bits, bring SSEL low, clock in 48 bits, and what comes out of the FIFO is the first 8 bits, not the last.  (Besides, were it daisy-chained, the next device in the chain would be getting the data loaded into the transmit FIFO, rather than the excess bits from the SPI frame.)

And yet, somehow, this cell has become The Standard for use in ARM-based MCUs.

Update:

• This malfeature is indeed baked into ARM PrimeCell PL022.
• Turns out ST doesn't use PL022; the STM32's SPI cell looks like it handles SSEL the way it's normally done in the real world.
• So, maybe I use an STM32 eval board for this gadget, instead of the LPC1xxx eval board.  (Or I could probably use an AVR, whose SPI looks to be simple and well-behaved.)
• But - and this is the fun part - I think the other side of the interface is an STM32, which may make life interesting, given that the side I'm trying to emulate is apparently built around an LPC, and their SPI implementations are incompatible.

Oh, well... the abovementioned fun part is Somebody Else's Fun, and a workaround is possible, albeit kludgey.

That could have been worded better.

From this article:

Some patients have died of overwhelming infections, including some Asians now living in the U.S.

Um.  For sufficiently undead values of "living"?  Is this another thing we gotta worry about?  Do I need to start carrying silver bullets, or broad-spectrum antibiotic?

Print me out, Scotty!

Seems there's a company working on a 3-D bio-printer.

If they can get it working, and someone comes up with a corresponding bio-scanner, we'll have teleportation!

(And, yeah, they're currently only talking about teleporting food.  As in the New Yorker cartoon from many years ago: "Fax Food," showing a fax machine printing out a 4-color pizza with extra pepperoni.)

Ah, more fun with the pod creatures

Starting on a little application that's to run on an LPC1343; specifically, on an LPCXpresso board, 'cause that's a cheap way of getting a chip of appropriate capabilities mounted on a board with I/O connections, just right off the shelf.

Anyway: seems appropriate to set this up on the workbench in the lab.  I'll be connecting the LPC board to a Total Phase Aardvark adapter for testing SPI operation, and that's already set up on the lab machine.

So, I pull in the udev rules for LPCXpressoes and such, and lo and behold the IDE actually does find the debug pod this time, and I can debug a small program.

So I decide it's time to activate the lab copy of the IDE.  Oops!  Activation works, but now it's giving me Java out-of-memory errors and heap problems, even after I fiddle with the .ini file as recommended.  Maybe it's something to do with only having 2GB of physical RAM in that box, though there's also a perfectly good swap partition that hasn't been touched.

So, back to the workstation, where LPCXpresso is known to work properly.  Install the Aardvark Control Center.  Oops!  No adapter found.  Unhelpful.  Maybe some 32-bit library issue on this 64-bit Linux system?

Virtualbox to the rescue!  Boot up that 32-bit Squeeze box from last week, set it up for Aardvarking, tell Virtualbox to give that machine the Aardvark, and hey presto!  So now desktop #7 has the 32-bit VM running the Aardvark software, and desktop #8 has the LPCXpresso IDE, and both pods are physically plugged into the same computer, and all's well.  (And JEdit is on desktop #12, with the demo app's source files open.  And people don't understand why I need 16 virtual desktops and regard Windows as horribly crippled for not supporting such things.  Well, among the other reasons for regarding Windows as horribly crippled.)

Summer gadgets: signs of life

Having to correct a few bonehead blunders along the way, but things seem to be more or less right so far.

L to R: Store-boughten STM32F4 Discovery board, being used as an SWD pod; my programming/debug/serial-port adapter; my handset board (whereof the exact nature shall, for the time being, remain shrouded in secrecy).

Few patch wires and kludges:

On the adapter board, I designed in a chip that I'd used before, and that had since gone obsolete in that footprint (SOT23-8).  Installing a TSSOP part in that footprint was tricky, but worked for the prototype.

Also on the adapter board, I connected an entirely wrong modem-control signal, which makes for a cut & patch operation.

Further, that half-pitch DIPswitch really needs to be normal-sized, maybe with protruding levers.

On the handset board, I carefully checked the pinout before using an STM32F1-series library part for the STM32F4xx that's in the actual design... but I didn't check that all the ground pins were still ground.  Oops!  Two of them are actually outputs from the core voltage regulator, and are supposed to be bypassed to ground, not shorted to it.  Fortunately, those pins are adjacent to corners, so lifting them post-installation was merely a royal pain.

Then, when I finally hooked up the debug cable: no talky.  Poke things with a scope probe.  Wiggles here, wiggles there... hey, when I probe SWCLK at the target end, the debugger signs on!  Ship it with the scope attached!  Or, perhaps, tack a 22pF, 0402 cap (what happened to be to hand) from SWCLK to GND on the target board, and make a note to terminate it properly in the future.

Anyway, I loaded the DAC_SignalsGeneration demo, and now I get waveforms on the DAC1 and DAC2 lines, so things are more or less working.

Couldn't find a handy demo program that talked to USART1 in an immediately identifiable way, so maybe I need to write one.

Then: get AGROS running on the Cortex-M family, write a driver for the STM32 UART, get it signing on, start writing drivers for other peripherals (SPI, I2C, ADCs, that sort of thing), and it'll be Application Development Time!

Summer gadgets, n-1

Actually, it now stands at 2(n-1), as the 2-layer design whereof I'd ordered two pieces arrived in today's mail, and there were four of it, in keeping with the pattern of BatchPCB delivering twice as many boards as I'd paid for.

So, that's the radio platform board here.  Yet to come, the Navi-Brain prototype.  Maybe that'll get here Monday.  (Maybe?  Hm, only two tracking numbers for the order, and I've received two envelopes.  Hope there wasn't some screw-up.  Ah: the envelope that arrived today bore the numerically-lesser tracking number, by several zillion.)

Meanwhile, I made a start on assembling the radio handset, and discovered that the ergonomics are, shall we say, suboptimal.  Well, that's why we build proof-of-concept gadgets before rushing into production, right?  It should come out usable, but I'll be omitting at least one pair of toggle switches, and thinking that the switches should have been across the bottom in place of the row of buttons.  Long's it works electrically, it'll be a good start on the production model.

Update: oopsie?  While emptying the lab trash, I found the envelope from the first board shipment (the 2-layer designs), and it bore the numerically greater of the two tracking numbers.  Maybe I need to rattle a cage via e-mail, come Monday, 'cause there was definitely no Navi-Brain board in either shipment, and I was kinda hoping to have that'n for show & tell in a month or so.

Update 2: seems I was misinterpreting the order-status page, and "shipped" doesn't necessarily mean that all 4-layer boards have been shipped; my Navi-Brain board is probably just later in the 4-layer queue, and should turn up eventually.

Another peripheral is short of interrupt options...!

I've been known to kvetch, from time to time, about the proliferation of 8250-descended UARTs, and how their lack of a "transmit complete" interrupt makes life difficult for those of us who sometimes have to deal with an old concept called "half duplex" (y'know, like some old-timey modems, or modern 2-wire RS-485).

Well, the ARM SSP (Synchronous Serial Port), as used by NXP, has a similar lack: while the status register has a "transmit FIFO empty" bit, it's not brought over to the interrupt status register... so there's no way to get an interrupt when the last of the data you'd loaded into the TX FIFO has been sent (or, at least, loaded into the shift register) and it's time to rescind the "data available" signal you're sending to the SPI master.

And, even when you are the master, it's highly convenient to have an interrupt for "the last byte you loaded into the TX FIFO has now been clocked out of the shift register, and the TX path is bare."  For, y'know, controlling SSEL outputs when you might have several or the built-in SSEL just doesn't do the right thing.  (In the AGROS driver, I ended up using the receive-timeout interrupt for end-of-transmission, though that's delayed by some unspecified "timeout period.")

So, another pollish driver.  Fortunately, the application at hand is a tiny standalone thing.

Wonder how that'll work out.

Heard on the radio: the Obama and Romney campaigns have agreed to a day off from negative advertising on September 11.

Meanwhile, a SuperPAC with no connection to any candidate is preparing an ad that shows two guys whose last names and party affiliation all begin with 'R' crashing airliners into Social Security and Medicare....

Meticulosity

Been going through the Subversion logs for the Advanced Tactical Toaster's firmware, identifying what changed at each commit, and documenting in one place everything that's changed since release 1.00, based on the diffs, which are much more detailed than the commit messages (especially since the commits are often for things in some other part of the project tree, relating to entirely different projects).

Lotsa detail.

And I really should (since no one else seems to be doing so) review the official Firmware Control Procedure, which is derived from the procedure I outlined, lo these many months ago.  Might just be worth making sure that 1.03 (coming Real Soon Now, pending some executive decisions on feature changes) is done exactly according to the Book.

Another brain malfunction

Just got an e-mail announcement that yet another science-fiction author has died.  (Harry Harrison this time.)

I keep seeing such announcements.  Before long, we'll run out entirely, right?

I never see an announcement when a famous author is born.

Brian L. Marzelvane (2012-2107), Hugo- and Rory-winning author best known for his Ammulite trilogy, was born last night at his parents' home in Riley, Oregon.  His first published work will be "The Secret of Bamberga", the second-place entry in the 2021 Skyway Soap Short Fiction Competition.  His wife, Annette, will be born next year, and they will marry in 2036.

Guess I'd better wake up properly now, and get some real work done.

Did I mentioned sotware rot?

Going on two year back, I got the LPCXpresso software to work under Debian-stable (Lenny, at the time).

More recently, I'd installed the latest version (4.2.3) on the new workstation, which is running -testing (Wheezy, which, as I've noted before, tends alarmingly to the -unstable, but which does play nice with my Sandy Bridge hardware).

Well!  I've got a project coming up in the next week or so which actually calls for using an LPC1343, probably in the form of an LPCXpresso board - to simulate a peripheral, in a manufacturing test setup.

So, time to fire up that IDE and rummage through the example code.

IDE launches, and reports being fully licensed and good for 128K code.

Debugger does.  Not.  Work.

Grrrr....?

So, I try the funny incantations of dfu-util and crt_emu_lpc11_13_nxp, and I get pretty much the same messages as all those other poor schlubs out there, most of whom eventually solved their problems in one way or another, but none of the published solutions appear to work for me.

Figuring it's a problem with running on 64-bit Linux, I try a couple of different 32-bit installs on virtual machines, as well as the -stable (Squeeze) installation on the 32-bit mini-box in the lab.  Same results.  Likewise if I try the manual invocations of dfu-util and crt_emu_lpc11_13_nxp as root.

Hmmm.  Try running the IDE as root.  It woiks!  Goes through the "set up your workspace" bit, then I install the examples, and it finds and uses the LPC-Link.

Play with the udev permissions.  Still doesn't work for me as myself.

Hokay: maybe something's wonky with the configuration.  Run it again as myself, and switch to a fresh workspace.  Hooray!  It finds the board, and debugging works!

So, something stale in the old workspace.  Great.

(And I'm thinking that the VM and lab-machine failures, which showed exactly the same symptoms, actually were a permissions problem on  account of my not having brought over my udev script.)

But!  All's not yet well!  When I try a tiny example, it runs fine, but when I try to debug the alphabetically-first example, being adc, I get "License restriction: Flash image too large for this license".  Looks like it's a bit over 8K, which is the limit for the unregistered version.  So, WTF?

Apparently the system updates over the last few months - different kernel version, and ifconfig returning the Ethernet MAC address differently - changed my system's identity enough for the debugger to think it wasn't properly licensed, while the IDE happily reported being fully licensed.

Solution: generate a new serial number, and re-license.  Hey, it's free.