"Holocaust survivors said Monday they are through trying to negotiate with the Mormon church over posthumous baptisms of Jews killed in Nazi concentration camps, saying the church has repeatedly violated a 13-year-old agreement barring the practice."

I think that the Mormons' practice of baptizing Holocaust victims is really the new definition of chutzpah. Inmitigated gall, indeed. More ...

Leaders of The Church of Jesus Christ of Latter-day Saints say they are making changes to their massive genealogical database that will make it more difficult for names of Holocaust victims to be entered for posthumous baptism by proxy, a rite that has been a common Mormon practice for more than a century.

In May, the Vatican ordered Catholic dioceses worldwide to withhold member registries from Mormons so that Catholics could not be baptized.

Now, I don't particularly care if the Mormons want to have ten wives and marry their sisters and cousins and what-not (though it boggles the mind that they support California's Prop 8 and its definition of marriage as being between "one man and one woman, but hey, hypocrisy is everywhere).

But seriously: Fuck the Mormons.

That's the only explanation that makes any sense for why there's such the outrage over the Rev. Jeremiah Wright's statements.

As Jon Stewart rightly points out, to no less a Republican douchebag than Newt Gingrich (who personifies the "IOIYAR" double standard), why do weapons-grade assholes like Pat Robertson and the wormfood Jerry Falwell get free passes, and in fact are actually PRAISED for their absurd comments?

Hagee should be excoriated, not sought out for their support and endorsement.

I really don't think it's because Wright is black. It's because he's a Democrat, pure and simple.

One day the fuckin' chickens will come home to roost on the Republicans, and that will be a glorious day indeedy.

But the problem is simple: Every time I hear someone say, "I'm going to give you some straight talk," I know that I'm in for a mound of bullshit.

And in McCain's case, it's total weapons-grade bullshit.

I'm sure that I am not the only person who thinks this way.

Sure, he's a war hero, blah blah blah. That didn't stop him from ignoring the Swiftboating of decorated war hero John Kerry, or the disgusting insinuations by the wingnuts that THREE limbs Max Cleland left in Viet Nam were somehow self-inflicted.

And, sure, he spent five years being tortured in the Hanoi Hilton. But that didn't stop him from standing by and not actually doing anything about George Bush's policies of torturing "detainees" in Abu Gharib and Gitmo and other, more secret, detention centers.

Typically, McCain talks the talk but never -- NEVER -- walks the walk.

So our new mantra should be:

"John McCain is the kindest, bravest, warmest, most wonderful human being I've ever known in my life."

Kinda just rolls off the tongue, eh?

Now, who's dealing a deck full of Queens of Diamonds?

An easy demonstration design is to use the display as a text terminal. I could have put a micro of some sort into the FPGA but the '340 is more convenient, and I figured I'd use the micro's USB interface as the data source. So anyways, I partitioned the design and put some of the smarts into the micro and some into the FPGA. The basic breakdown is as follows:

FPGA: The GLCD interface is in the FPGA. Here, we generate all of the lclocks and strobes and data lines needed by the GLCD. The FPGA also contains the dual-port display buffer, implemented using Xilinx block RAMs. The top-level module is glcdtest.vhdl.

The FPGA gets its 48 MHz global clock from the micro's SYSCLK output. The FPGA interfaces one side of the display buffer block RAMs to the micro's EMIF. Note that the micro side of the display buffer is read/write, so if the micro wishes to read a location (useful for blinking characters at the cursor location), it can do so. This is handled by the dispbufctrl.vhdl module.

The 48 MHz clock is also fed into a DCM where it is divided down to a 4.8 MHz clock (which is four times the 1.2 MHz LCD clock). This "4x pixel clock" runs the GLCD timing and buffer memory access. The DCM is instantiated in gdcm.vhdl.

glcd.vhdl contains the logic that interfaces to the GLCD. It has counters that keep track of the current display row and column as well as generating the display buffer address for the current pixels. All clocks and strobes for the GLCD are generated here.

ps2kbdif.vhdl provides the logic that reads keypresses from a standard PS/2 keyboard. (The Digilent board has a PS/2 port.) The raw keycode bytes are then sent to the micro over SPI using the spimaster.vhdl module. The SPI clock is sourced by the FPGA and is the same as the 4x pixel clock. This means the SPI port is clocked at 4.8 MHz.

Micro: The C8051F340 micro gives us an RS-232 interface (see uart.c) to a host computer (USB at some point). It implements the character generator ROM (see charmap.h), and deals with writing to and reading from the display buffer (display.c). The main program loop (in main.c) keeps track of the current cursor location, which is given in terms of character size (8 x 8 pixels). This means that when the cursor is at location (4, 2), the upper-left corner of the cursor is at pixel location (32, 16). Note that (0, 0) is always the upper left corner.

It's interesting to note that the microcontroller sees the display buffer memory as 8 bits wide but the display controller sees it as 4 bits wide. This is further complicated by the fact that the display is divided into four quadrants, and the display interface's 4-bit data bus loads one pixel for each quadrant at a time. The micro sees the display buffer as 8 bits wide and loads 8 pixels in the same quadrant at a time. The comments in the sources should (hopefully) explain how each side of the dual-port buffer addresses things, as it's not exactly straight-forward.

Here are the FPGA sources, fitter project and test bench. The fitter project is compatible with Xilinx ISE 7.1i SP4 tools. The test bench has been tested using ModelSim XE III 6.0a.

Here are the microcontroller sources. The project file is for Keil uVision 2 and C51 version 7.04.

To Do:

Note that the micro doesn't do anything with the SPI input as of now (11 Jan 2007 with firmware rev r840). This means you can't connect a keyboard to the FPGA board's PS/2 port and have things you type show up on the display.

I have a LinkSys NSLU2 network-attached storage device running Debian Linux (instead of its native firmware) and it'd be nice to use this thing as a display terminal, so (for example) when booting I can see console messages.

It'd be great to make the display a generic graphic display, instead of a text terminal display. The FPGA logic should remain the same as it's just the display buffer and timing control. The question then becomes how should the micro handle writing to the display?

Obviously, communicating to the host over USB instead of RS-232 would be cool. One thought is to make the display compliant with the USB Video class, advertising itself as a simple monochrome graphics terminal. Then we'd have to handle how we draw things. This could be interesting.

My home stereo uses a pair of powered studio monitors, so I don't have an integrated amplifier/tuner. Instead, I have an old NAD Model 1000 stereo preamp, which is really just a fancy switchbox with a phono preamp (all discrete!), tone and volume controls, and a headphone jack.

One thing it doesn't have is a remote volume control. For a long time, this hadn't been a problem, since all I used the stereo for was listening to music. Now that we've got the TV, the Mac mini (used as the DVD player as well as a DVR with EyeTV) and the cable box hooked up to it, lack of remote volume and mute control has become quite annoying. (We HATE commercials.)

But I'm an engineer ... I can build my own remote volume control! And I did. Here's how.

a) Use the remote's Volume Up and Volume Down buttons to increase or decrease the volume. When the volume is changed, the Gain: display and bar-graph meter are updated appropriately.
b) Use the remote's Mute button to mute or unmute the audio. When muted, an (M) annunicator appears in the display.
c) Use the remote's numeric entry (0 through 9) and ENTER buttons to directly enter a volume level (in the range 0 to 255). Numbers greater than 255 are ignored.
d) Use the remote's POWER button to turn the LCD backlight on and off.

Digital control of analog audio volume control is easy if you use a TI PGA2310 stereo volume control chip. It uses a simple 16-bit SPI interface to load two eight-bit gain-set words, one for each channel. The gain ranges from +31.5dB to -95.5dB in 256 0.5dB steps. The chip uses ±15V rails for its audio amplifiers and +5V for the digital section. (The much-cheaper PGA2311 uses ±5V analog rails.) Basically, each channel consists of a stepped attenuator where laser-trimmed resistors set the attenuation, followed by a variable-gain buffer. The buffer is capable of driving 600Ω loads, so an external line driver is not necessary (for single-ended applications).

I chose to use the Philips RC5 IR remote protocol, which I won't go into here as there are a whole bunch of other web references for it. I imagine that most universal remotes support it, certainly if you configure the remote to talk to a Philips TV set. Rather than use a discrete phototransistor to receive the IR light bursts and then demodulate it in a microcontroller, I chose to use an IR receiver module. These are simple three-pin jobs, with power, ground and data-out pins. Vishay offers a full range of receivers in all sorts of packages. However, Radio Shack stores stock their part number 276-640, which works well enough.

Finally, I use a microcontroller to decode the receiver's data bits and turn the keypresses into gain set values that are sent to the PGA2310. Pretty much any microcontroller with an interrupt input, an SPI port (you could bit-bang, too) and a timer should work.

Here's the guts of the box. Yep, it's a hack. I had the chassis lying around and used a step drill and a nibbler to cut out the holes and such.

I'm partial to the Silicon Labs 8051 micros. There's a bunch from which to choose, with various analog and digital peripherals as well as RAM and flash. The best thing about these parts is they all have in-circuit debug/emulation, and US$35 gets you a USB JTAG dongle, which is like two orders of magnitude less expensive than the old Nohau 8051 emulators.

I had a C8051F226 development board floating around that had a standard 16x2 LCD attached to it, so I figured I'd use the LCD to show the gain and mute status. I've already written code to interface to the LCD. Plus, the LCD was useful in trying to figure out what codes the keypresses actually generated. This was much easier than using a 'scope and counting transistions on the serial data line.

Here's the micro board. The 10-pn connector on the bottom (next to the LEDs) is the JTAG programming/debug header. The LCD is on the right. The +3.3V regulator is on the lower left. The devel board's expansion header is at the back.

So, there it is: IR receiver module, micro, LCD, analog volume control chip. One more point: the analog input comes from the NAD preamp's Tape 2 out (because it doesn't follow the volume control like the Outputs 1 and 2), which is unbalanced on RCA jacks. My speakers have balanced inputs on XLR connectors, so I used a couple of TI DRV134 balanced drivers for the outputs.

Here's the connector panel. The two RCA jacks are the inputs and the two male XLR jacks are the outputs. The RCA jack channel is the same as the XLR below it. I didn't bother labeling them as "left" and "right" because I adjust left and right identically; there's no balance control (although one could be implemented easily in the firmware, I chose not to do so because there isn't an obvious balance control on the remote). The jack on the lower left is the input for the wall wart. (Sorry about the crappy mechanicals here. Maybe I'll get a drill press and/or some punches for Hanukkah?)

For power, I use a 16VAC wall wart. On a piece of perf board, I built up a simple ±15V supply using a couple of 1N4004 diodes, 470 µF caps, and LM317 and LM337 regulators. I then put the PGA2310 and the two DRV134s the perf board, and wired up the serial signals back to the micro board. I also connected the +15V wire to the micro board's +3.3V regulator as well as the 78L05 regulator I use for the LCD (which also goes back to the analog board as the PGA2310's +5V supply). So the sandwich is pretty ugly, but functional, and when it's in a box, you don't see the ugliness.

Here's a close-up of the analog/power-supply board. The PGA2310 is on the adapter board in the upper left. The two DRV134s are in the middle left. The ±15V power supply is on the right beneath the connectors. The ribbon cable wiring to the micro board is at the upper left.

Here's a close-up of the driver chips and the PGA2310. Note that the PGA2310's package is not the same as the adapter board so a little creative pin-bending was necessary. (Don't be afraid of SMT!)

You can find a schematic here. Note that this is based on what I built using the micro devel board, which explains why the +15V is the input to both the +5V and +3.3V regulators. If I were to do a custom board, the +15V supply would feed a beefier +5V regulator, and that +5V regulator would be the input to an LDO supplying the +3.3V.

A micro port pin and a 2N2222A transistor are used to control the LCD backlight. Since the '226 device doesn't have PWM hardware, and I didn't bother to do one in software, the backlight is either full on or full off. Note that the backlight LED gets its current from the +15V supply, because the 78L05 can't do enough current for the backlight and everything else. (Again, had I not started with something I'd already had, I would've used a heftier +5V supply.) The LCD backlight specs 90 mA, so that means a 1-W 120-ohm resistor. I didn't have any available, so I paralleled four 560-ohm 1/4-W resistors.

Here's a close-up of the display. The IR receiver is in the hole on the upper left.

The firmware is here. I used the Keil C51 compiler and µVision 2 environment. It's reasonably well commented and hopefully you'll be able to understand why I do what I do.

Things to do, if sufficiently motivated:
a) Vary the backlight intensity using PWM.
b) Update firmware so it powers up to the previous volume level, rather than at zero.
c) Update firmware so that holding down Volume Up and Down buttons for a "long" time should increment/decrement volume level quicker.
d) Make a PCB and a nicer box.
e) Maybe set up to use the Apple Front Row remote?
f) Add relays or FETs to prevent power-on/off glitches from sending potentially damaging "pops" to the speakers.

Summary: now we have a convenient volume control for the stereo and TV. It's got very low distortion and noise, and a nice display. Enjoy!

DISCLAIMER: please don't e-mail me with questions like, "How do I do this with a PIC (or other micro)?" since I haven't used any other micros in a long time. Please don't ask me for detailed help regarding how IR remotes or LCDs work; as I said, there are a lot of web resources.

WARRANTY: there is none. I don't guarantee this to work at all, or work well enough for your particular application. I believe the schematic to be correct, and the firmware compiles using the Keil 7.04 compiler and loads into the 'F226 device using the SiLabs IDE. If you blow up your stereo, I'm sorry but it's not my fault. If you incorporate it into a product which you then sell, and the product explodes in a customer's face, I am not responsible and will be held harmless from any and all legal liabilities.

LICENSE: I can't keep track of which open-source license is which, so here's the deal. This code is © 2006, Andy Peters, ASP Digital. Feel free to use it and/or modify/extend it in any way you wish, with the following requirement: if you re-publish it or any derivatives anywhere, my copyright notice must be retained. If you find it really useful, please consider picking something out of my Powell's Wish List (enter andy@latke.net in "Your friend's email address." If you find any bugs, please tell me.

Good night, and good luck.

New stuff has been recorded here in Tucson with Jim Waters and might be released before the guys turn thirty.

Go see 'em while they're still playing in front of fifteen people at the local shithole. You'll thank me later.

I can think of another country that has a national language, one that demands that all business be conducted in the national language.

That country, of course, is France.

Yes, the same country these nitwits go out of their way to denigrate ("Freedom fries," anyone?).

So, basically, the wingnuts are saying that America should be more like France.

Make up your fucking minds, douchebags.

Next, I was able to see the repo from my Mac Book Pro, which was still running Subversion 1.3.2. I copied the source tarball to the MBP, ran the configure script, built and installed the new svn. svn --version told me I was running 1.4.0. So a simple test: the

$ svn ls http://svn.mydomain.foo:8888/svn/

returned an odd error: "Unrecognized URL scheme." Weird. Some web searches indicated that something was up with ra_dav not being loaded, and doing a

$ svn --version

confirmed it: ra_local and ra_svn were available but not ra_dav. So I did the reconfigure and rebuild thing, and during configure, I noticed that there was a comment about "can't find neon." A quick web search for "neon subversion" lead to the neon homepage, so I downloaded the latest and greatest. Buried within the subversion configure script output was a note about "get neon-0.25.5.tar.gz, unzip/tar it into the subversion source directory, and rename it to neon." I did that, reconfigured, and then got a complaint: turns out that the latest/greatest neon is 0.26.1, and Subversion really wanted 0.25.5. I downloaded the correct version of neon, put it in the right place, and reconfigured. There was no complaint about neon, so I did a make and a sudo make install, and voila! I was able to access the repo using the http protocol. Yippeee!!!

Anyways, like everyone else, Chase has online access to accounts. One login gets me to a credit card account and the mortgage account (there used to be four Chase credit card accounts, the result of their buying spree; I closed all but the oldest account).

You can pay the credit cards online and the payment is credited the next day, which at this point, all of the credit-card issuers let you do.

You can also pay the mortgage online. However, instead of electronically debiting the checking account and crediting the mortgage account, you have to sign up with their bill-pay service. And get this: rather than doing what's done with the credit card accounts, the bill-pay service debits your checking account, writes a fucking paper check and mails it to Chase Home Loans, a process which can take up to six business days. Yep, that's right: they mail a check to themselves. All that spewing to government regulators about how the big banks needed to combine all of their operations for the consumer's benefit turns out to be complete HORSESHIT.

You know, if I wanted to mail a check, I'd just write out a check and drop it in the mailbox at the post office (which is what I do). And it's worth noting that Wachovia's online payment system works as you'd expect: you schedule the date on which the payment will be made, and it's done and you don't have to worry about it.

A second issue with Chase's online mortgage-payment system is that there's no way to indicate that you want to pay any extra towards your loan principal. In fact, the system is too stupid to know your regular payment amount! So, you just fill in whatever amount you want to pay and they send a check to themselves and if you include an extra amount, you have no idea whether it'll be applied towards the principal or to the escrow or to the interest.

Since Chase tends to fuck up simple things like "please apply the extra payment amount to the loan PRINCIPAL instead of the escrow," I've had to call them and deal with their customer service. And every time I do so, I say, "oh, by the way, when will it be possible to do instant online payments like you do with credit cards?" and the answer is always, "I'll forward your request ..."

It turns out that there's a Real Good (from Chase's point-of-view, anyways) reason why they don't handle online mortgage payments like they handle credit-card payments: they offer their "Fast Pay" process, where you have to call them on the phone and authorize an overnight payment, a "benefit" for which they charge $12.

One imagines that there's a lot of people who have to do an overnight payment, and if you were Chase, you probably wouldn't butcher a cash cow, either.

But as a(n unwilling) customer, fuck that.

And fuck that black hole known as the Chase Mortgage Research Department. If your issue is referred to them, you're hosed. They want YOUR money immediately, if not sooner, but G-d forbid you close the loan and need to get an escrow refund.

The Fall are touring with and sharing backline with a North Carolina four-piece called The Talk, who do pleasant if generic noisy indie rock, and they namecheck other NC indie bands (Superchunk, Archers of Loaf). Also on the tour is this silly vidiot who, between bands "mixes" video images and music and it's all quite annoying. It's the sort of music you'd play to drive people out of the club after the show ...

Their set started on time, although Sean had to play therapist and convince the band that they really should play. They play, and Smith ambles on stage, and proceeds to remove the rack-tom mic from the drum. He takes off his jacket and hangs it on the ride cymbal overhead mic. Guitarist Ben's vocal mic is immediately removed from its stand and put on the floor. Smith decides he doesn't like the next vocal mic and drops it on the floor. He wanders around the stage, putting his vocal mic in front of the bass amp, the keyboard amp, wherever. He picks up the mic he dropped and sings into both at the same time. He puts a vocal mic inside the kick drum. I guess he doesn't recognize my SM91 in the kick drum as a mic so we have a kick sound most of the night, until late in the set it gets flipped over. (So I unmuted the vocal mic in the kick drum and it sounded pretty cool.) I use all of the PA (big surprise, that) and the Crest 8001 on the mids occasionally thermals and cuts out for a couple of seconds.

They play a tight, angry set, giving the crowd (165 paid) their $20 worth.

They play an encore. The crowd is stunned.

After the encore, I start zeroing the mixer. They come back onstage and do a SECOND encore. The crowd is astonished.

Yeah, the best show of the year, so far. Suffice it to say that the rest of the tour won't go as well.

Oh, yeah, they're coming back to Solar Culture on June 15. You have no excuse for missing them.

It's a bi-annual exercise in pure exploitation, that's why. Bands get paid $100 each. (Why an out-of-town band would accept $100 is beyond me, unless they're getting paid more, in which case the whole thing REALLY stinks.)

PA companies basically lend their services and their gear to a for-profit enterprise. (Jeb seems to think that a band bringing in their practice-space PA for $100 is somehow acceptable. I say: fine. This is proof that nobody cares if the bands sound good, and it's all about the beer.) If the reason for doing this gig as charity is because "we'll get other gigs," then jesus-h-fuckin'-christ: WAKE UP. There ARE NO OTHER GIGS that the Crawl's promoters handle. What, you'll do the Crawl so you get the priviledge of doing Club Congress outdoor shows for next-to-nothing, too?

There aren't enough good bands in Tucson to warrant the number of stages and venues. And it doesn't take a rocket scientist (or a sound man) to realize that putting big PAs in close proximity to each other is a recipe for disaster. Yeah, a couple of years ago it was a lot of fucking fun trying to get the acoustic performers on the "Americana" stage (on 5th St between Hotel Congress and the MLK apartments) audible over the noise from the Congress parking-lot stage (punk bands) and the big stage on 5th between Congress and Broadway (more punk bands). I mean: why bother?

I have other complaints about how the Crawl runs day-of-show, but those would just make this post too long.

Here's some more proof that it's all about selling beer (and not even good beer--I won't drink Budweiser). I realized this a few years ago when I was working on the Main Stage (with NO support or secuity people at all) and I discovered the source of a bad buzz in the PA: they plugged the fucking beer truck into my PA distro! So, I unplugged it and the noise went away. Then I plugged the beer truck into an outlet in the Rialto. That's when Paul Bear jumped my shit and plugged the fucking beer truck back into the PA distro, ignoring my protests.

Shame on the Tucson Weekly for continuing to promote this exploitational clusterfuck. Shame on all of the bands who think that there's a benefit to playing the Club Crawl.

So I woke up on Sunday morning with my back out of whack, and went to the chiropractor on Monday to get it sorted out. And I was a lot more consistent bowling on Tuesday than before. I averaged about 163. I finished out the first game with four strikes in a row, then started the next game with a strike. Plus I was picking up a bunch of spares.

All I need now is a pair of my own shoes and I'm all set.