Silicon Valley Redneck

Another useful chip

From Micrel: MIC2876, a synchronous boost regulator with bidirectional load disconnect.

Just built up a test board (1/2" by 5/8", dominated by the input and output headers; actual circuit about 3/8" square) and verified what it does:

• When it's operating, it steps 1-cell Li-Ion voltage up to 5V (or other combinations; this pair is what I want it for), with minimal fuss and reasonable efficiency.
• When the ENABLE pin isn't being driven high, it disconnects the load from the supply.
• It operates at a decently high frequency, so the components can be small.

So, when you'd like to have a boost circuit, but you're worried about that pesky business of the input having a straight path through the inductor and rectifier to the output... this is a handy solution.  Boost regulator and power switch, all in one!

Mind you, there are also applications where you'd actually want a boost regulator whose "off" output was at supply voltage; I had one of those a few years ago.  The requirement at hand isn't one of those.

Update: With a 500mA load, and the particular 1.5µH, 1008 inductor I picked, efficiency is about 75%, and the little board gets rather alarmingly toasty.  I've got a 1µH part of the same footprint and lower resistance that I can try, later.  If I end up using this thing in the actual gadget, I may need to resort to a bulkier inductor (for which there'll be some room, on the back of the board, if I push some other stuff around).  I still need to get a handle on current demand (peak, average, and how long the peaks last) for the application.

So, I spent all day Thursday...

...at a Freescale event in Santa Clara, picking up information on, mainly, the rapidly-evolving Kinetis* MCU family.

This inspires me to do a bit more digging, and fiddling, and....

Well, the MCUs themselves are quite impressive, and, as I've noted, some of the low-end ones are seriously competing on price with low-end 8-bit MCUs.

The software support, though, is lagging a bit.  They're in process of moving from CodeWarrior (I've encountered other people using that, many years ago, on well-funded projects) to free Eclipse-based Kinetis Design Studio, and, well, things are in transition.  Real Soon Now, there's supposed to be KDS for Mac (it's currently on Windows and Linux, but one of my potential products is likely to have firmware developed for it by the sort of people who use iThings), and also key code ought to be migrated to the KDS/GNU environment, but that's Not There Yet, and importing example projects doesn't seem to work the way one would expect.

There was a hands-on KDS session at the Thursday thing, but, having been rather late in deciding I could attend at all, I was too late to get a seat in that session, and instead sat in on the sensor fusion session in the next room (which was interesting indeed, and I've downloaded those libraries, and sometime this spring I'll be trying to figure out how to use them).

Ah, well.  Another interesting development: some of the new chips (e.g., KL27Z series) come with a multi-interface bootloader in ROM, to allow installing firmware over USB, UART, I2C, or whatever, without at any time requiring use of a special-purpose debug/programming connector**.  Could be handy, and maybe for the project that was tentatively planned with a KL26Z, I should use a KL27Z instead (same size package, and, I believe, slightly more expensive but with more memory, and still price-competitive with an 8-bit solution when all's said and done).

 

Update: Now my FRDM-KL25Z board seems to have gotten frotzed.  I don't know if this happened as a result of my attempting to speak OpenSDA to it using KDS, or if it happened earlier, but now if I try to load an Mbed program I get the dreaded SWD ERROR, and, while there are various suggestions out there in Netland as to the cure for this, none of them actually appear to work.

The FRDM-K64F board does still work as an Mbed, and I'm not sure I want to sic KDS on it until I have a better idea what's going on.

As is typical In The Industry, the third-party tools (debugging pods and the like) are wildly expensive, at least the versions not subject to severe licensing restrictions (generally forbidding any manner of commercial use), and it looks like the firmware needed to make the vendor's own development tools work is also subject to awkward (and, presumably, widely ignored) restrictions.  So, it's not altogether clear whether one could, legally, sell knockoffs of the FRDM OpenSDA subsystem in pod form, at least in any usable state.  Which is really a shame, because (assuming it actually does work with KDS, if your target board hasn't gotten frotzed) it would make a nifty cheap programming pod, only a little pricier than one of the little USB AVR programmers and having actual debugging capability to boot.  (Though extending the pod design and its protocol to include control of target power would be a really nice thing for those of us who like to automate processes, hint hint.)

Update 2: Based on the debugging output when I run openocd manually, it looks like it's the target processor on the FRDM board that's gone bad... or maybe I messed up the SWD interface to it somehow, perhaps when I installed a set of headers on the board.

Anyway, for all I know, I've got a perfectly usable programming/debugging pod sitting there.  But if not... it looks like the "Multilink Universal" from P&E Micro might be about right, and the $200 price tag isn't all that painful as capital equipment goes.

 

* I still get this mixed up with Kinect and/or K'NEX.

** NXP has had this feature for years, but only supporting a UART interface, and with a somewhat poorly-considered convention for the use of modem-control signals.

Low leakage

That tiny gadget I've been mentioning has a pair of 100µF ceramic chip capacitors (1206 footprint, X6T dielectric, rated 6.3V).

These are across the output power bus.  In operation, they get charged up to 4.2V from the battery bus when the load switch is turned on.

I've just noticed something.  If I do some fiddling with the load disconnected, put the gadget into power-down mode (which turns off the load switch, natch), disconnect the battery and the external power source, set the board down all by its little lonesome, go fix dinner, eat dinner, watch a couple of vids, then pick up the board after it's been sitting for an hour or so, and plug in the load...

...There's still enough charge in those caps to get quite an impressive blink out of the LEDs on the load.

I'd say technology has improved dramatically since I was a kid, but then I remember that, back in those days, the phrase "100µF ceramic capacitor" would have been a ticket to the looney bin.  There were 2.2µF ceramic capacitors.  They were quite bulky, and lossy.  (Audibly lossy.)  And now we have these wonderful things that cost (checks last DigiKey invoice) 67.4¢ each in quantity ten.  And a 41.2¢ (qty. 10) load switch: rated 3.5A maximum, 1.1V to 5V, logic-level control input, 42mΩ ON resistance, and OFF resistance high enough not to be worth trying to measure, all in a 1.2x1.6mm package.

Yow.

Cool thermal imaging goodie!

Now from Spark Fun!

Well... not now, exactly.  A couple of hours ago, yes.  Sometime in the future, presumably.  Those sure sold out quickly!

Thing is, I've been aware of this sensor (or something remarkably like it) since Fall 2011; going into early 2012, JCM Consulting Technologies was negotiating, via a third or possibly fourth party, for the contract to come up with a manufacturing test and calibration solution for a FLIR module that I think would have looked a lot like this dingus (sans eval board; just the little square-thing-with-lens).  Specs weren't firm yet, timeframe looked a lot like "deliver it last week and we'll tell you what we need in a couple of months", and we didn't land the deal.

If memory serves, projected volume was in stocking-stuffer territory, with a product (whatever is was to have been) incorporating the module retailing for an enhanced-stocking-stuffer price, maybe around $100.

Well, I never did see that product, but now the imager module is (or was, and may be again) available on a little eval board for $350, which is quite a bit more than I'd hoped for, but just having it out there is kind of neat.

No, I didn't snap one up this morning; it wasn't in my budget (though a thermal imager, even a low-res one, is firmly on my tools-for-real-work wish list).  Maybe next year.  And I did have some ideas for products to design around it, so maybe I need one for work and one for tinkering with.

Update: I find someone selling the modules at $300 a pop, and someone else organizing a group buy at a bit over $200.  Also, FLIR now has a complete handheld camera built around this puppy, for $500 from various retailers.

For serious work use, if the need comes up again, I think I'd spring for a somewhat pricier model with better resolution.  But, for funsies, the little low-res module would be kind of neat (and isn't in this year's funsies budget).

Some amazing tech dribbles down

Been looking at low-end chips the past few days.

I find that Freescale is having a go at poaching customers from the AVR/PIC worlds, with the Kinetis E series: runs on 2.7V to 5.5V, is decently appointed with memory and peripherals for competing with the ATmega line... and starts at around $1 in onesies.

Then, today, I go looking at FPGAs for a rather specialized application (for which I can't easily find an MCU that'll do the job, within the given constraints).  The Lattice ICE40LP series, which I'd been considering for some bigger applications... starts under $2 in onesies, for a device that won't handle the task at hand but appears useful for other purposes.  And the slightly more expensive model that probably would do the trick, clocked at the necessary frequency for my purposes, looks like it'll run on under 5mW.

These things need feeding, of course.  So I look up dinky, low-power buck regulators, for getting down from lithium-cell voltage to MCU-and-FPGA voltage (assuming a 3.3V-or-lower MCU).  Well!  TI has a handy chip that comes in a SOT-23-5 package, runs at 1 MHz, and has a low-power pulse-skipping mode and 15µA quiescent current.  And, under a buck in onesies.

Might end up designing the whole mess around lower-voltage logic; maybe 2.5V rather than 3.3.  But there's also an analog section, so that means shopping for low-voltage CMOS op amps.

I may just be able to put most of the analog signal chain inside the FPGA, which would be mighty handy.  That would leave just some preliminary signal conditioning to be done with an op amp or two.

An unexpected parcel

I returned from afternoon errands to find a smallish brown box awaiting me.

From Mouser.

I haven't ordered anything from Mouser lately, have I?  Is this some long-forgotten back order?  Or did I win a drawing?

I won a drawing.  At ARM TechCon, I'd put my name in a hat (or something to that effect) for a chance to win one of several nifty eval boards, and apparently I won one of them.

So now I have a Motorola Mozzarella Fresca Freescale FRDM-K64F.  Got the chip (a right powerful little MCU), MBed debug interface, Ethernet PHY and connector, and application USB connector, plus headers that I guess are supposed to be double-row Aard-Wino style, and some other on-board goodies.  Oh, and even a MicroSD socket.

The form factor does appear almost conducive to using it as back-panel module for a programmable instrument, though the card socket is at the wrong end.

So, another thing to play with... In My Copious Free Time.

I'd picked up a lesser edition of the board, for a lesser processor, at the MBed booth at the show.  Lessee, now: yeah, a KL25Z.  So, maybe I should get up to speed on the Freescale Cortex M family, as opportunity allows.

Squee!

Aren't they the cuuuutest little things?  Actual size is 0.35" by 0.625", and in real life the top and bottom soldermask are the same shade of OSH-Park Purple; for some reason the camera didn't capture that.

Ordered the boards, let's see, Sunday, September 28^th.  There was unused space on an expedited panel, and my tiny boards fit, so I got a free upgrade, and the purple envelope turned up in my mailbox today, Wednesday, October 8^th.  All for $3.15, delivered, for nine pieces.  Digi-Key box with parts also arrived today, and I took a few minutes to assemble four of the li'l goodies.

Yeah, it's a stupid simple circuit, and, yeah, the same functionality is available in a single tiny chip... which needs to be mounted to a board, maybe with a capacitor, in order to be useful.  Ever try to solder one of those 4-ball micro packages?  It ain't easy!  This has all practical-sized parts, so I can build it by hand.

And something from Monday afternoon

One of the long-running projects involves a sensor which includes a piezo disc: effectively, a specialized microphone.  In fact, the output gets recorded as digitized audio, using a computer's or iThing's microphone port.

But, it needs amplifying.  The traditional solution is this.  Pretty flexible, but on the large side, and it needs its own battery.

Ergo, this:

Basically the same concept, except that I dusted off some good design practices vaguely remembered from high school (there's a reason for that bypassed resistor in the emitter circuit!), looked up a transistor characterized for low-voltage, low-current operation with moderately high gain (and factory-screened for a reasonably narrow range of gain), did some actual math, spent a few minutes throwing the design together in EAGLE, and sent the board file and three bucks for six copies to OSH Park.  Hey, it's 0.35" by 0.875"; not a lot of square-inchage to it.

And it works, first time!  As designed, it takes input from a piezo mic, and amplifies it nicely, with the iThing's mic port providing the power and the load resistor.

Might become a product, even.  It'll probably lose R1, R2, and C1, though: they're meant for biasing an electret mic, and, as powered by an iThing, there's just not enough power for both this and a typical electret module.  The preliminary respin is already down to 0.75" long; if those three components go away on account of pointlessness, it'll shrink even more, maybe to 0.35" by 0.625" (main problem then being that there's not much room on the back for the product name and my name).

Updates: (1) respin, now 0.625" long, is out to fab, and apparently got a spontaneous expedite on account of there was space on an expedited panel; (2) preliminary testing with the client shows that this works, in his application, at least as well as the Spark Fun product, with the possible exception of very low frequencies.

First impressions: Epson Workforce WF-7610

Really first impression: that's one darn big box it comes in!  And heavy, too.

Unboxing is much like any other inkjet printer (+scanner/copier/FAX), only with even more bits of blue tape.

The beast having been unpacked and de-taped, I wrestled it into place on my side table, having to remove even one more pile of paper to make space.

And I need to keep a good-sized area clear, 'cause changing the ink cartridges requires scooting it out from under the overhanging bookshelf.

Paper tray handles various sizes neatly; for the big sizes, like 11x17 (which is what I mostly need it for), the front of the tray extends.

Paper goes in the front and comes out the front, just like a typical office machine, but unlike inkjet printers I've had in the past.

I had to plug the USB cable into the workstation (Debian Wheezy) to get a version of CUPS that speaks WF-7xxx; the server (Squeeze) knows nothing of such devices.

I told CUPS it was a WF-7510, and it seems to work OK.  Test page came out zippy-fast, and nice and clear.

Next up: print a schematic from EAGLE.  Oops!  For some reason, EAGLE thinks that printer only does A3, A4, and Letter.  Hmmm?

Ah, but e-mail brings me an 11x17 schematic that I'll be needing on the morrow, ain PDF form.  Acrobat knows it's an 11x17 printer.  Schematic comes out fast, and decently usable... but I think I may want to fiddle with the driver settings, or choose a better default print quality, or something; it's a tad light.

So: gotta fiddle with the default settings, and try to figure out what's up with EAGLE.

Haven't tried the scanner yet.  Opening the cover is problematic in its current home.  But, hey!  It has a sheet feeder!  Might be usable as scanner or copier yet, as long as I'm dealing with loose sheets.

Nifties!

Well, Intel's been making some interesting goodies for the embedded-systems world lately, and the latest one... well... Spark Fun has the gadget, and an initial crop of add-ons.

The Intel Edison is a really teeny computer module, with x86 (Atom) processor, enough RAM and flash-disk to run Linux and get some work done, WiFi and Bluetooth, and a little I/O header.  For $50, in onesies.

There's no video output... which isn't exactly a bad thing.  Most of the applications I have for little embeddable Linux machines don't call for video, and the lack of video on this module implies that the Linux distro for it will be intended to run headless.

It's not the zippiest CPU out there, but you probably won't be running KDE and a database server on it.  From my perspective, it's competing with maybe a 200 MHz ARM9 (being something available in a non-BGA package and capable of running real Linux), so a 500 MHz dual-core Atom ought to be plenty.

Lack of Ethernet knocks it off my list for some potential projects, but by no means all.

Overall: Cool!  I won't be playing with one right away - got enough on my plate already - but it sure looks like it has potential.  Also, there's one pending project for which it might be just the ticket, assuming the project actually happens.

Disclaimer: If you think my scribbling here will have any meaningful effect on the price of my Intel shares, your notion of how the stock market works is at least two decades out of date.

Not having the most productive weekend

Well... I came to the end of Friday exhausted, and with a long list of Work to do, on three different fronts.  And those are just the urgent ones.

What with household stuff, family stuff, and being rather under the weather the past few days, I just haven't accomplished much over the weekend.

Kinda hard to focus, and over-use of caffeine to try to keep going doesn't help any.

I've made some progress on one of the work projects.  Possibly even the most important one at the moment.

I did find a distraction, as if I needed another one: addressing the issue of poorly-formed bottom layers on the FlashForge 3D printer.

Turns out, though FlashForge doesn't officially have any firmware updates, installing the latest Sailfish firmware (and using the corresponding version of ReplicatorG) does... well, it doesn't improve the quality that much, but the printer finishes jobs a bunch faster, even at the same nominal speed settings.  Then some fiddling with the Skeinforge settings, and... it's not exactly there yet, but I'm seeing some definite improvements in quality, while still having improved build times.

Also, I got a roll of ABS filament from a high-rated vendor, though it's looking like the roll of ABS that came with the machine is probably fine, and the firmware and settings will do the trick.  Hey, having another roll of filament around in a second color won't hurt.

And I built the current version of Slic3r, which seems to work (unlike, apparently, the version bundled with the released ReplicatorG 0.40), as long as everything's set correctly... like the offsets being 0, 0 instead of half-full-scale.

Update: And, the next step in the project whereon I have made some progress?  Involves enabling forwarding of some HTTP requests from the live(-ish) server to the test server, before I go replacing the entire content of the live server with untested material.  It's simple, and the code worked fine when that-which-is-now live was inhabiting my home server.  Just involves uncommenting the call to http_post_data.  Simple, right?  Er.  Gotta install pecl_http, which requires installing a couple other things first.

But, everything having been installed, and Apache gracefully restarted?  I still get Call to undefined function http_post_data().  And, if I check for that function being defined, and then call phpinfo()?  Why, I find that it's not defined, and yet:

---

http

HTTP Support	enabled
Extension Version	2.0.6

Used Library	Compiled	Linked
libz	1.2.7	1.2.7
libcurl	7.26.0	7.26.0
libevent	2.0.19-stable	2.0.19-stable

Directive	Local Value	Master Value
http.etag.mode	crc32b	crc32b

---

So it's installed, and enabled, and yet the function I want to use is not there.

The extension is version 2.0.4 on my server, where it worked before.  Surely the handy functions haven't been eliminated in a minor revision!

I think I should have spent the weekend sleeping, and not promised to have anything done by Monday.

Wunnerful: I look at the documentation for pecl_http, and there's nothing at all about the handy utility routines, which however are still described in the PHP manual as coming from that package.  I'd seen something about an API change, so I backed off to version 2.0.4 (no visible change) and then to 1.7.6 (the format of the phpinfo section for http changed, but it's still showing as enabled, and the function I want still isn't there.  The documentation does say there's a new namespace in version 2, but 1.7.6, as noted, doesn't work either, and anyway the documentation is entirely object-oriented, and (from the package side) has nothing whatsoever about convenience functions of any sort.

In the time I've wasted trying to get a stupid function call to work, I could have rewritten my script using some other (but less convenient, assuming both are working) library.  Or a raw TCP connection, maybe.  Perhaps I should rewrite that one function, using cURL or something, just to have it handy and avoid subsequent mucking about with cURL.

Friday the 13th come on a Friday this month; maybe the bad luck is hanging around all weekend.  Now I'm trying to remember whether Fred (the mostly-black cat) crossed my path on Friday morning.  (I'm pretty sure he didn't; I seem to recall encountering only the Dread Guardian of the Forest.)

Linode.com: first impressions

tl;dr: Yow!

Slightly longer: Well, a couple of months back I asked a friend about virtual-private-server hosting services, and he recommended Linode.  I'd kinda looked at their website.  Now, what with me wanting to move gumbyware and orgzone off my DSL line, and a client needing a Linux VPS, 'twas time to act.

I signed up for the entry-level plan: $20/month for 2 cores, 2GB of RAM, and 48 GB of disk space.  Looks, from what I've seen, like disk space scales with CPU cores and RAM, so when I start needing More Storage I might have to spring for more crunchpower along with it, but we'll burn that bridge when we come to it.  Should be a while before that comes up, tough; while I have a lot more than 48 GB on the current server, most of that is internal stuff, with no need to be internet-accessible.

Anyway: following the "Getting Started" instructions was easy, and creating a node with Debian 7.5 installed took next to no time.  I booted it, logged in via ssh, installed synaptic the command-line way, logged out and back in so's X11 port forwarding would start working, and commenced installing packages via synaptic.

Gotta carry over a simplified version of my firewall rules (the ones for home are rather arcane, and deal with three network ports, each with its own special cases)... probably before I spend too much time setting stuff up, being as how screwing up the rules on a VPS could lock me out entirely, leaving me to start over from the beginning.

Seems zippy-fast so far.

Update: with Apache, PHP, and PostgreSQL installed, plus piles of options and supporting stuff, here's the status:

root@liabc-xyz:~# df -h
Filesystem      Size  Used Avail Use% Mounted on
rootfs           47G  1.2G   46G   3% /
/dev/root        47G  1.2G   46G   3% /
devtmpfs        996M     0  996M   0% /dev
tmpfs           200M  188K  200M   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs           450M     0  450M   0% /run/shm

So, a couple of CDs worth of storage used so far, installing the software.  Haven't installed Subversion yet, nor various other stuff, so I'll probably be at 5% usage before I start adding actual data.  (And moving my Subversion server over will get Fun, because I didn't set up a cname for it, so everyone with a working copy of anything checked out is pointing to the true name of my current server, on the DSL line.  Lesson: when you set up a service, always create a cname, and use and distribute that instead of the true name.)

Typepad WTF: that update up there came out full of <span> tags that I didn't ask for and that have no apparent use.  It's so full of <span>, you'd think it was the Green Midget Café.  Wait: now the whole post is full of them.  Bloody Vikings.

Update 2: just set up a server for a client.  Took hardly any time at all, but then he has less baggage than I.  Setting up the reverse DNS was painless, and included a validity check.  About 2 hours from start, he now has the system and forward and reverse DNS all properly configured (apart from some configuration tailoring, which will require some actual thought).  Doesn't have the application goodies installed yet, but that's mostly a matter of deciding where to put them and then SFTPing them over from my old server.

New machinery!

So I mentioned G codes and/or ReplicatorG a couple of times lately, which might have constituted a Hint.

I'd been pondering 3D printers, of the FDM persuasion, for at least two years now.  Early this year, I'd come to the conclusion that there was a legitimate business case to be made for having one, and the now-discontinued Makerbot Replicator, dual extruder variant, was probably about right, and its $1200 Chinese knockoff, the FlashForge Creator, was probably the thing to get, especially given the reviews it was getting on Amazon.

Come Monday, finding the business case still there and my finances not totally depleted by taxes and car repairs, I decided it was time to buy... though, in the meantime, FlashForge had come out with a new model, the Creator X, which is basically the same thing in a metal-frame enclosure with plastic* panels, instead of plywood.  The price on the original Creator had dropped a little, while the Creator X was at $1300... only, on Monday, it was $1156.  Okaaaay, then!

Just after the shipping notice arrived, there came a message from FlashForge tech support, with a PDF of the instruction booklet and a request that I RTFM before unpacking the product.  And so I skimmed through it, and noted some important little details, and printed a copy to have handy for when it arrived.

The box showed up on my doorstep just after lunch today (being Wednesday, as promised).  Covered with Chinese shipping labels, and showing signs of rough handling, but with nothing rattling.  On opening, I found no signs of damage beyond the outer box.

At this point, I noted a deficiency in the unpacking instructions: they failed to address the issue of cat removal.  Huckleberry thought this looked like Fun.

Anyway...

The printer having been extracted from its somewhat-damaged outer box by its convenient handles, I proceeded to remove the assorted packing stuff, extricate the spools of filament (white Antilock Braking System, blue People's Liberation Army), install the extruders, install the spool holders, install the filament guide tubes, and that stuff.

Unpacking and setup took an hour or so; woulda been quicker if I had a bigger & less cluttered work area and no cats.  There were a couple of bobbles in the instructions, but nothing I couldn't figure out quickly.  Would have been fun to make an unboxing video, given a larger work area but including the cats.

A test: the "20mm calibration cube" (20x20x10mm).  Came out 20.06 x 20.07 x 10.5mm, but I think the bottom 0.5mm is part of the raft and meant to be discarded.  Finish is not too shabby; it's got some texture to it, and was obviously built up out of filament, but it's good enough for functional purposes, and I presume a bit of sandpaper and maybe a coat of paint would make it dog-&-pony-show-ready.  The machine says it took 20 minutes; I don't know if that includes warmup time.  Had I some black filament handy, I might do a 1x4x9 test monolith, Just Because.

Only technical issue encountered so far: the first couple of times it looked for the home position, the right-hand limit switch didn't seem to do its thing, resulting in many seconds of alarming noise and vibration.  I jiggled the connector, which didn't feel loose, and the next home operation went OK.  I think I need to inspect the switch/board/cable.

Oh, and a limitation: the build volume isn't big enough for Huckleberry.  A more standard-sized cat might fit.  I think I may want to add a front door to this thing, if only to protect it, and the noses of inquisitive felines, from each other.  Also, it'll need a dust cover.

Now to generate the build file for a somewhat larger and more complex object, for an actual revenue project.  It's something I would ordinarily machine out of randomly-acquired plastic stock, but if it'll 3D print, that'll save me an hour or so of machine-shop time per, and the client... um... the cost of my time.  Did I think there was a good business case here?  Well, I can use that time to do other billable work, while the machine whirs away all by itself.

Update:

The rather complex shape for Real Business took 38 minutes of unattended buzz-click-whir.  Finish on sides is quite passable, even on the sloped bits.  Finish on top (which has some odd curvature to it) definitely shows the layering, but is usable.  Finish on the bottom is a bit more problematic; I'm not getting a clean transition from raft/support to actual item.  I'm assuming this is a function of the way ReplicatorG structured it, and not of the hardware.  Also, the item doesn't quite fit the other piece, but that's because I left an important cutout out of the design.  Have to add that, and reprint, perhaps with revised settings.

I have a feeling I may end up tinkering a bit with the ReplicatorG source code.  Really, I shouldn't have to go into the G-code and edit the platform temperature.

Also, I gotta figure out how to put multiple parts in one job, given that most of the things I'll be making take up only a tiny fraction of the platform, and putting a bunch of them in one job to be run unattended seems a better use of my time than running one after another.

It's probably time to put a computer out in the garage, with WiFi, instead of running back & forth with an SD card.  A RasPi or BeagleBone ought to do the trick.  Or an old laptop with Ubuntu installed, for that matter.  No UI hardware needed; just ssh in from my workstation and run ReplicatorG from there.  (Though: I can see local control being Very Useful in some situations, and the front-panel controls might not quite be enough.)

D'OH:

Nothing wrong with the right-hand limit switch.  Way I assembled the thing, the wiring harness for the extruders was flopping to the right... and getting pinched between the carriage and something on the right side, so the carriage really was never reaching home.

More update:

Yes, of course it can be used to print cat toys!  The part I'm fiddling with needs to be built on a raft, and I keep finding my discarded rafts scattered around the house in little pieces.

* Correction: on closer examination, the panels appear to be some sort of processed wood product, with smooth faces, painted black.  Whatever.  The overall impression is much more serious than the wood-finish plywood model.

I probably want the second edition of this

Cool!  Via El Reg: a buncha Aussies is working on a conductive-ink PCB printer, the EX¹.

One of their maybe-when-we-get-ambitious goals is double-sided printing, with a camera for alignment... but...

A less-ambitious goal is a non-conductive ink cartridge for silkscreen.  And: the FAQ says they're working on multilayer.

Why double-sided, anyway? With additive manufacturing, why not include a (high-capacity) insulating ink cartridge for soldermask and insulating layers (maybe in addition to the different-colored one for silkscreen), and do multiwire?

Multiwire wouldn't be good for RF, absent some way of controlling impedances (which could, I suppose, be arranged, even to the extent of printing coaxial cables), nor would it match the characteristics of normal production boards, but 'twould allow quick few-offs of insanely complex designs.

Will life imitate my deranged imagination?

A couple of years back, I had an Idea - I don't recall whether I shared here, but since there's no prospect of my pursuing it, I'll go ahead and describe it.

The idea, see, is that you'd engineer strains of bacteria that (a) could live just beneath the surface of human skin, and (b) were bioluminescent, in a variety of colors.

Now put little colonies of them in itty-bitty cages so they'll stay where they're put.

Glow-in-the-dark tattoo ink!

And now... someone's gone and put bacteria in cages.

Though I'd assumed the cages would look more like buckyballs and less like serving dishes for monkey brains.

Cute li'l physical prototypes

Swung by the office today, and picked up a package from Purkle Porkypine* which I'd inexplicably had delivered there, containing some intricate little plastic parts with fiddly tolerances.

They seem to have come out OK.  I don't currently have all the other bits that need to fit together with the new bit, but the ones I have fit just so, and they certainly look like what I designed.

With a lick of paint, they'd even pass for production parts at a moderate distance.  (With a magnifying glass, no.  The surface texture is distinctly 3-D printed, not injection molded.)

Delivery was decently quick, and the price was reasonable for a quick-turn, high-detail job.  (Much as I'd like to have a Makerbot of my very own, it wouldn't have helped on this project; this called for a professional-grade printer.)

Assuming testing goes well this coming week, I'll be meeting with the injection-molding guru soon to discuss volume, and the design tweaks needed for good manufacturability.  (Actually, I figure to hand him the CAD file, along with some direction as to which dimensions are critical and which surfaces need a matte finish, and let him do the tweaking.)

Sorry, no photos, because Confidential.

 

* Oh, and I suppose you can pronounce "purple porcupine"?  Let's hear you say it ten times fast.

In which I rave about Solder

I had a few boards to rework today, for a client.  Pretty simple; a few small SMT parts to pull & replace.

Except: it was a lead-free board, so melting the existing solder was a Royal Pain.  My usual approach in such situations (drenching the board in flux and applying hot air until it chars not generally being acceptable) is to apply some tin-lead solder to the parts in question, and allow it to form a lower-melting alloy than the tin-silver-copper stuff that's already there.

Well, just this morning the postman brought me a syringe of tin-bismuth low-temperature lead-free solder paste.  I'd ordered it for doing the flip sides of my little boards, and in particular the Geiger-counter boards, which have proved highly problematic when the solder on the previously-installed components now on the underside of the board melts at the same temperature as what's under the to-be-soldered parts on the top.  (In fact, I ended up soldering the flipside components on the last one with an iron.)

Inspiration strikes!  Goober a little tin-bismuth solder around the old parts, apply just a little hot air... and it melts, gets right in there, and alloys into the old solder, quickly resulting in a removable chip!

The pads having been cleaned up, the tin-bismuth stuff proves right nice for installing the new parts.  Quick melting, excellent wetting, all the stuff you don't expect from lead-free solder paste... nor even from the leaded stuff.

Highly recommended.

Update: I wonder if anyone at the solder companies has considered fiddling the standard tin-silver-copper lead-free alloy with a few percent bismuth... or perhaps indium or gallium... in pursuit of an alloy with a melting point closer to the old tin-lead eutectic and maybe better flowing & wetting properties.  Surely they have R&D metallurgists on staff... or don't they?  The industry has seen, what, three new alloys in the last five decades?  And modern accounting treats R&D as pure expense.  Maybe some mad metallurgist working in his garage needs to build the better alloy.

Got another purple envelope today

Containing three of my latest purple board - yes, the one that goes with the mysterious East European item - and a purple sticker.

Update: assembled the right (i.e., Eastern-facing) end of the board this afternoon, and spent an hour or so playing mad scientist with this, various bench instruments (power supply, pulse generator, scope, voltmeter), and the Mystery Item.  It works!  Next up: populate the MCU and a few other parts, and throw together the basic operating firmware.  Then, the left end (USB interface), and some rinky-dink Ruby program for demo purposes.

I expect to be turning this into a Genuine Product (actually, 2 or 3 closely-related products) sort of June-July-ish, assuming an ongoing supply of Mystery Items can be negotiated.  Watch for an announcement.

Update, Sunday afternoon: MCU is now controlling the interface section (in a rudimentary way, as I'm missing a small but important component at the moment), and taking measurements.  Another few hours should have the dingus fully operational.  When those few hours will be... well, I've got a busy week coming up.

Update, Thursday lunchtime: hardware is now fully armed and operational.  I'm sending measurements back to the host PC via the FT201X.  Just need to implement the rest of the firmware features, and it'll look remarkably like a Product (modulo packaging and suchlike details).  This afternoon, though, there's Real Work to be done.

New toy

My Beaglebone Black just arrived!

Cute li'l rascal.  And, though it sports a micro-SD socket, the OS lives in 2GB of on-board flash.  So, the add-on memory card is optional.  Looks like a good configuration for designing into custom programmable instruments, with the power, Ethernet, and USB connectors exposed to the world.

Haven't powered it up yet.  Got work to do, and here it is the end of the month, Walpurgisnacht, and all that.

But!  I now have the mechanical sample that allows me to make sense of the cape template I found in someone-or-other's library (prolly Lady Ada's).  Looks like I pretty much have to keep my tall components - in particular, things that go at the ends, such as connectors - topside, unlike the RasPi crust.

For now, back to work.  Assuming, of course, that I can get the darn furry Huckle-beastie to go take a nap someplace that isn't on top of me.

Last of the summer gadget boards

That board I ordered from OSH Park just over two weeks ago?

Here we go. Nice and shiny, if incompletely depanelized (a few seconds with pliers took care of the rails).

Continuity check shows the inner layers connecting the things they're suposed to connect.

Finish, like Kryten's final words, is ENIG.

I was going to spend the afternoon stuffing first the power section and then the processor, but a busness lead just came over the transom, so I may be chasing near-term revenue instead.