Laen's blog

An itty bitty power provider

Laen — Mon, 12 Jul 2010 20:49:56 -0500

SiliconFarmer found this awesome little solar cell on Digikey that puts out 4V at 50 microamps. Not a lot of juice, but add a 0.33F supercapacitor, and leave it in the sun for about 7 hours, and you have enough energy to run an Atmega168V in active mode at 1MHz for about 22 minutes, or in sleep mode for 4 days (!).

Parts List
Solar Cell: CLA289-ND on digikey
Super cap: P11064-ND on digikey
Diode: MMBD914-FDICT-ND on digikey
PCB from the DorkbotPDX PCB Order

OPEN LAB! Sunday, June 27th at PNCA

Laen — Thu, 24 Jun 2010 16:09:52 -0500

Location: Pacific Northwest College of Art, NW 12th and Johnson, room 205 Date: June 27th Time: 1pm to 5pm

Hey everyone, it's that time again! Fight the weather and join us at our quarterly/semi-annually Open Lab!

This one's promising to be pretty active.

We'll have:

- Erik Walthinsen giving Xmega demonstrations (is that right, Erik?)

- Jim Larson, Scott Dixon, and I (Laen) will be teaching hot-plate surface mount soldering (if you want to give it a try yourself, bring something to solder, or we'll have parts you can buy and cook on the spot).

If you have something to teach, something you want to learn, or a project you'd like to work with others on, reply to this post and put it on the list! (Or just show up. It's just that kind of workshop.) The more people that come, the more value everyone gets out of it.

Thanks!

-Laen

DorbotPDX Open Lab Unworkshop - March 28th, 2010

Laen — Mon, 15 Mar 2010 15:04:29 -0500

Where: PNCA, NW 12th and Johnson
When: March 28th, 2010 1pm-5pm
What: OPEN LAB!  Bring your projects and work on them in a group setting. Bounce your ideas off other dorks. Void warranties.

We're treating this as kind of an un-workshop, which means its up to you to set the agenda. Reply to this post with your interests and try to get a group together to hack on them.

Some ideas:

Propeller hacking. I know from the PCB orders that we have a bunch of people geeking on the Propeller platform.
PureData play. We have tons of PureData folks in the group. This would be a good place to get together and work on patches.
Circuit board designing. The PCB order goes out the evening after this Lab, so if you want help with your design, this is a good place to get it.
Arduino/Dorkboard/Teensy hacking.
Soldering practice. I'll have a bunch of Drawdio kits available, if you want to practice your soldering skills. They have spots for both surface mount and through-hole components, so take your pick.
Making things move. Messing with motors, servos, soleniods, etc. BYOMSSE (Bring Your Own Motors, Servos, Solenoids, Etc)

PCB Order #4, The Quickening

Laen — Tue, 16 Feb 2010 11:00:00 -0600

Announcing the fourth PCB Order!

The next PCB order will occur on March 1st, so submit your designs by 11:59PM on February 28th.

Now's your opportunity to have your design fabricated by a professional board house in the USA!

New this order:

Silkscreen on both sides of the board is now official. The board house has been giving it to us anyway, but we were only quoted for single-side.
7mil minimum annular ring. This is the width of the copper pad around a drilled hole.
Faster turnaround. The boards should be mailed to you within a week of the order deadline.

The Eagle CAM file has been updated:

The "Milling" layer has been removed from the board outline, since that's not actually supported in our order.

The Eagle DRU file has been updated to include the new minimum annular ring size.

What you get

Three 2 layer boards (see design rules below)
Soldermask on both sides
Silkscreen on both sides.

The Costs

$5 per square inch for three copies of your board. So, a 2 square inch design would cost $10, and you'll get three copies of your board. Shipping by US Mail is included, and there are no setup fees.

How to submit your order

For now, just email your board designs to "pcb-order@denizen.org" as either an Eagle .brd or a set of Gerbers.

Include in the email:

The number of boards you want in sets of 3.
Your mailing address

If you're making Gerbers from Eagle, please use this CAM job -- this is a new CAM file for this order, and was updated on February 15th to include a remove the Milling layer from the board outline

http://content.laen.org/pcb/LaenPCBOrder.cam

Help!

If any of this is confusing, or you'd just like some help with your board, feel free to drop me an email.

Design Rules

(Eagle DRU File: http://content.laen.org/pcb/LaenPCBOrder.dru) -- This is a new DRU file for this order, and was updated on February 15th to include a narrower annular ring.

6 mil trace width
6 mil spacing
15 mil clearances from traces to the edge of the board
13 mil minimum drill size.
7 mil minimum annular ring

Enjoy!

-Laen

Circuit Board Order #3!

Laen — Fri, 15 Jan 2010 19:02:36 -0600

=== Announcing the third PCB Order! ===

Now's your chance to get that design you've been working on professionally fabricated in the USA.

This is open to DorkbotPDX dorks, Make:PDX makers, OpenTechSpace hackers, and anyone else that wants in on it. Orders must be received by February 2nd, and will be distributed by mail by February 10th or at the DorkbotPDX meeting on February 15th.

=== What you get ===
- Three 2 layer boards (see design rules below)
- Soldermask on both sides
- Silkscreen on the top side only.
- Optional Mylar solder paste stencil ($5 extra)

=== The Costs ===
$5 square inch for three copies of your board. So, a 2 square inch design would cost $10, and you'll get three copies of your board. Shipping by US Mail is included.

Scott Dixon has agreed to make Mylar solder paste stencils (for easy application of surface mount componants) for $5 a board stencil. If you'd like to take advantage of that, just include the "tcream" file the CAM file below produces, and tell me you want a stencil.

=== How to submit your order ===
For now, just email your board designs to "pcb-order@denizen.org" as either an Eagle .brd or a set of Gerbers.

Include the number of copies you want in multiples of 3.

Include a mailing address if you'd like your boards mailed to you, which is the fastest way to get your boards.

If you're making Gerbers from Eagle, please use this CAM job -- (this is a new CAM file for this order, and was update on January 25th to include a solder paste layer)
http://content.laen.org/pcb/LaenPCBOrder.cam

If you want a mylar solder paste stencil, say so in the email and include the tcream gerber file.

=== Help! ===
If any of this is confusing, and you'd like help, feel free to drop me an email.

=== Design Rules ===
(Eagle DRU File: http://content.laen.org/pcb/LaenPCBOrder.dru) -- (This is a new DRU file for this order)

6 mil trace width
6 mil spacing
15 mil clearances from traces to the edge of the board
13 mil minimum drill size.

Enjoy!
-Laen

An ISP/SPI shield for the Teensy

Laen — Wed, 13 Jan 2010 18:12:49 -0600

With the inclusion of a functional AVRISP MKII (and PDI!) programmer in LUFA, I thought it would be fun to make a programmer board to take advantage of it.

This is an programmer "shield" for the Teensy2. Solder on some headers, plug in your Teensy, flash it with the LUFA AVRISPMKII code, and you have a functional ISP programmer.

This also works as a Serial, SPI, and i2c breakout board, since those pins are brought out into convenient headers.

Put a jumper on the "PGM" pins, and you can use the ISP pins to flash program the teensy itself. Useful if you've lost your bootloader somehow.

The EAGLE board files are released under a Creative Commons Attribution-Share Alike 3.0 United States License.

Surface Mount Soldering Workshop - Saturday, December 19th at TechShop

Laen — Fri, 11 Dec 2009 14:58:25 -0600

So, you've designed a circuit board in Eagle, gotten it fabricated, and you just got it back. Now you see that the parts which looked so big in Eagle are actually REALLY DANG SMALL. What are you going to do? How are you supposed to solder that?!

Working with tiny Surface Mount parts is challenging but can be done in the comfort of your own home. Come to Dorkbot's Surface Mount Soldering Workshop, presented by Jim Larson, held at OpenTechSpace at TechShop, on Saturday, December 19th, 2009 from 1pm to 5pm.

When you're done, you'll have a step-up voltage module to take a AA battery up to 5V.

To sign up, email smd-workshop@denizen.org.

Cost will be $35, payable at the class.

Please bring:

A soldering iron
A pair of tweezers
A pair of wire cutters

Class will be held at:

OpenTechSpace at TechShop
10100 Southwest Allen Boulevard
Beaverton, OR 97005

CC Licensed Circuit Boards

Laen — Sun, 29 Nov 2009 11:47:37 -0600

I've set up a wiki page for publishing our circuit board designs and linking to others around the net.

I threw in a couple of my own, a few of Ladyada's great designs, and Leah Buechley's LilyPad. Help build this resource by adding your own!

http://dorkbotpdx.org/wiki/creative_commons_licensed_circuit_boards

PCB Group Order #2

Laen — Fri, 20 Nov 2009 19:04:36 -0600

=== Announcing the second PCB Group Order! ===

Now's your chance to get that design you've been working on professionally fabricated in the USA.

This is open to DorkbotPDX dorks, Make:PDX makers, OpenTechSpace hackers, and anyone else that wants in on it. Orders must be received by December 10th, and will be distributed by December 21st.

=== What you get ===
- Three 2 layer boards (see design rules below)
- Soldermask on both sides
- Silkscreen on the top side only.

=== The Costs ===
$3 square inch for three copies of your board. So, a 2 square inch design would cost $6, and you'll get three copies of your board. Shipping by US Mail will be $2 extra.

If you're part of the workshop (see below!), you'll receive your order at the December 21st DorkbotPDX general meeting.

=== How to submit your order ===
For now, just email your board designs to "pcb-order@denizen.org" as either an Eagle .brd or a set of Gerbers.

Include the number of copies you want in multiples of 3.

If you're making Gerbers from Eagle, please use this CAM job:
http://content.laen.org/pcb/PCBGroupOrder.cam

=== Help! ===
If any of this is confusing, and you'd like help, feel free to drop me an email. Also, be sure to check out the DorkbotPDX PCB Workshop.

=== PCB Workshop ===
Don't know how to make a PCB, but still want in? Come to the "Designing Circuit Boards for Fabrication" workshop presented by Don, Dorkbot, and Tempus Dictum Inc.

Details at: http://dorkbotpdx.org/workshop/circuit_design_for_fabrication

Going to the workshop gets you 2 square inches of design space in the board order.

=== Design Rules ===
(Eagle DRU File: http://content.laen.org/pcb/PCBGroupOrder.dru)

8 mil trace width
8 mil spacing
15 mil clearances from traces to the edge of the board
20 mil minimum drill size.

Enjoy!
-Laen

DorkbotPDX Open Lab, June 28th at PNCA

Laen — Sat, 27 Jun 2009 17:22:38 -0500

Come one! Come all! To the DorkbotPDX Open Lab, generously hosted by the Pacific Northwest College of Art.

As the name suggests, this is an open lab for working on whatever projects you want to work on. It's a great place to show off works-in-progress, get help with a troublesome project, or to lend your expertise to someone doing something cool. It is also an opportunity for people who have attended workshops to follow up with anything that needs more work.

Bring your art projects, that could use the spark of electricity! Bring your electronics projects, that could use a touch of art!

Bring your Arduinos! Your Dorkboards! Your PICs! Your ARMs!

Bring your soldering irons, your protoboard, your motors, your sensors! your passives, your actives, and your power supplies and power strips. Bring your imagination, and the stuff to make it happen.

Time: 1pm to 5pm, Sunday, June 28th Place:PNCA, Room 205

Adventures in Etching - The test pattern

Laen — Sat, 30 May 2009 11:56:18 -0500

I made a test pattern to determine what my etching capabilities actually are:

(PDF)

This one is a bit better than my last one. It's smaller (2x2 inches vs. 5x4 inches) so you don't waste as much board, has test points, and a bunch more common surface mount components.

(results inside)

My method was:
* Print to magazine paper.
* Transfer to copper using 12 trips through the laminator.
* Etch in a muriatic acid / hydrogen peroxide bath.
* Remove toner with acetone.

Here's the final result:
Board 1:

The 8 mil trace is good!

Board 2:

Only the 10 mil trace and higher are good.

Experimenting with Toner Transfer

Laen — Tue, 28 Apr 2009 01:58:56 -0500

I spent some time over the weekend trying to see what my board etching capabilities actually are. I use toner transfer, and I thought that kind of limited my options, as far as how fine of detail I could get. I made this test pattern with a variety of surface mount components and line sizes, then tried to transfer them to copper with a couple of different methods.

First, I used a clothes iron, several different temperature settings and applying differing amounts of pressure. This is the one that turned out the best:

(Full resolution)

Then, I switched to a laminator. After sending it through a dozen times (top and bottom), I got this:

(Full Resolution)
MUCH better. I can do .001" lines sometimes, and .005" lines consistently. I can _almost_ do TFQP-100, and I can certainly do any of the other TQFPs.

Atmel's ATXMEGA64A1

Laen — Fri, 17 Apr 2009 19:49:12 -0500

I just got these beauties from Mouser today:

This is one of the first of the ATXMega line of chips, and _boy_ what a beautiful chip:

* 8 USARTS.
* 4 i2c busses
* 4 SPI busses
* AES and DES Cryptoengines
* 2 eight-channel, 12-bit ADCs
* 2 two-channel, 12-bit DACs,
* a built in 32kHz factory calibrated oscillator for the Real-time Clock
* built in 2MHz and 32MHz factory calibrated oscillators
* A 31x PLL which can be used as the system clock (for a top chip speed of 200MHz).

All for $6.40 a chip, quantity 10.

Now I need to make a board for it.

I might try doing a seeedstudio run.

Success at Etching!

Laen — Mon, 24 Nov 2008 00:37:21 -0600

It tooks some tries, but I think I'm doing pretty well at etching now.

Step 1 - Design the board in Eagle.

This was probably the most time consuming part for me, since I spent a lot of time trying to "get it right". Trying to make a board single sided can be a lot of work. Making a ground plane (drawing a polygon and naming it "GND") helps. Making fat traces (24mils) is pretty much required for a beginner like me. Also, having really wide clearances (again, 24mils) helps, though it makes it even harder to make things single sided, since that means no wires going between IC pads.

Step 2 - Print it.

I had bought some expensive toner transfer paper, and had pretty good success with it, but this time I just used magazine paper. It's cheap paper, so it has really tiny fibers. That makes it easy to remove from the board. The toner transfers off of it _really_ well.

It has to be a laser printer, since toner is a plastic that can be transfered to copper and ink is just, well.. ink.

Step 3 - Prepare the copper clad

I have 6"x11" 1/2oz copper clad board. That's too big for my 6"x6" Tupperware etching tank, so I cut it down to size using "Aviation tin snips". They're an $18 pair of scissors for cutting metal. It's still hard to cut the board, but at least it's possible with this stuff. I was tempted at first to just use my Dremel to saw through it, but I've heard that fiberglass dust is carcinogenic if you breathe it, so I want to make as little of it as possible.

I made my boards large enough to fit several copies of my design on it, just so I'd have redundant boards in case some didn't make it through to the end of the process.

I cleaned the entire thing with soap and water, scrubbing it down _hard_ with scotchbright. This gets fingerprints off of it, and gives the toner something to sink into when you transfer it over.

Step 4 - Transfer the print onto the copper clad.

First I tried to use a laminator, but I wasn't getting very good transfers. Many of my traces would come off during etching. This time I just used a cheap-o clothes iron. I ironed the hell out of it for probably 5 minutes, until the toner was practically soaking into the paper. This was probably too much, but it worked.

Be sure that it _all_ gets ironed, even the corners. I lost two of the nine copies, and they were both corner pieces.

Then I dumped it into a bucket of water, gave it a few minutes, and pulled off as much of the paper as possible. The remainder I removed with a tooth brush. You can be somewhat rough here. If the toner is transfered, then it's stuck hard to the board.

Step 5 - Etch it!

I used a muriatic acid and hydrogen peroxide etch. It's good stuff, and nearly infinitely reusable (I hear). The fumes kind of suck, so I etched outside. My mix was 2 parts hydrogen peroxide + 1 part muriatic acid as this seems to be the consensus online. I hear it's important to always add the acid to the peroxide, and never the other way around. Oh, and I always wear rubber dishwashing gloves when the acid is out.

I made enough etchant to cover the board with about an inch and a half, then heated it up to near boiling. That's not strictly necessary; I've heard of people etching with this stuff at room temperature, but it really speeds things along when it's hot. The etchant is clear at first, but turns windex blue then green as it eats copper. I kept the board moving in the etchant (both side to side and up and down, so the bottom would get etched well too).

It seems like nothing is happening until the last couple of minutes, then it all happens at once.

When I couldn't see any more copper for it to etch, I pulled the board out of the mix and dumped it into a bucket of water I had ready.

Step 6 - Clean it, cut it and drill it.

At this point, I have a bunch of etched boards, still covered with protective toner. Acetone (nail polish remover) removes that. Standard practice seems to be to remove _all_ the toner (and that's certainly what I did this time), but I wonder if
one could just remove the toner around the solder points. The toner would then serve as a little bit of an insulator if you were putting the board in a metal enclosure.

With the aviation tin snips, I cut out my individual boards from the large etched board.

To drill my boards, I used a Dremel tool drill press stand and some micro carbide drill bits from Harbor Freight. A number 58 drill bit seems to be a pretty good size for making holes that component leads can easily fit through.

Once again, careful of the dust, as rumors on the Internets says it's carcinogenic. I just made it a habit to exhale/blow at the drill bit when drilling and inhale between drillings. This is really producing more large shavings than dust, but better safe than sorry.

Step 7 - Soldering it.

Since I gave myself such large clearances in Step 1, I didn't have to do anything special to solder the boards. I did have a bottle of water soluble flux handy, just in case I was having problems, but I didn't need them.

Single Sided PCBs

Laen — Wed, 12 Nov 2008 20:09:58 -0600

In preparation for making PCBs, I've created some little one-sided PCBs in Eagle.

They are:

* Moto - An L293D H-Bridge breakout board.

* Tiny45-USB-SMD - A tiny45 and the circuitry necessary to talk USB.

* MagicHat Controller - An Atmega168 based board with an i2c header for driving 12 vibe motors.

AVR Power Saving Modes

Laen — Thu, 30 Oct 2008 16:12:04 -0500

AVR chips have a couple of built-in ways to reduce power usage: "sleep" modes, and "Power Reduction Modes". Sleep modes stop the running program (which is the biggest power hog), waiting for an interrupt to wake it up, while Power Reduction Modes can be used any time (including while the processor is asleep).

Sleep Mode:
The sleep modes available to you depends on which type of chip you're on. An ATmega168 may have all of them, while an ATTiny45 only has a couple.

SLEEP_MODE_IDLE
SLEEP_MODE_ADC
SLEEP_MODE_PWR_DOWN
SLEEP_MODE_PWR_SAVE
SLEEP_MODE_STANDBY
SLEEP_MODE_EXT_STANDBY

While the processor is asleep, the main program won't be running. The only way to wake it up is via an interrupt, and sometimes only an external INT0 interrupt (depending on the sleep mode).

To set the sleep desired sleep mode, put this in your code:
set_sleep_mode(SLEEP_MODE_IDLE);

and to actually put it to sleep:
sleep_mode();

Make sure you have an interrupt enabled, or the processor will never wake up.

Power Reduction Mode:
In addition to putting the whole thing to sleep, you can also turn off parts of the chip that you aren't using by using the chip's Power Reduction Manager.

Like, if you're not using the ADC, shut it off:
power_adc_disable();

Or if your project doesn't need to do any serial communication:
power_usi_disable();

For maximum power reduction, just run this:
power_all_disable()

then enable what you want you actually need.

The 5VP1Rev1

Laen — Thu, 30 Oct 2008 11:45:07 -0500

My first etched board! Thanks, feurig!

It's a DC-DC step-up circuit that takes one or two AA batteries and spits out 5V on the other side.

Making PCBs

Laen — Thu, 16 Oct 2008 11:58:18 -0500

I'm on a quest to make some PCBs. I was going to wait for ~~Portland~~Beaverton TechShop to open, but it's clear now that that's going to be a ways away.

I need boards to etch:
* Copper Clad Boards ("FR4", says drtwist) (1.6 mm thick (35um copper), says the riccibitti link below)
* Surplus Gizmos has tons of scrap copper clad. It's also available very cheaply on ebay.

I ended up ordering 50 sheets of 1/2oz copper clad FR-4, 0.060 thick, 6"x11" from "abcfab" on eBay.

Special paper for transfering my design onto the boards (not strictly necessary, but may provide higher quality transfers)
* Toner Transfer Paper (Digikey 182-1003-ND)
* Green TRF (182-1021-ND) - improves quality of etching by binding to the toner.

I ended up using magazine paper. It works really well, and is practically free, unlike toner transfer paper which is about $1 a sheet. I haven't found the green TRF to be necessary.

Chemicals to etch with:
* Chemicals (Ferric Chloride or Ammonium Persulphate, says drtwist)
** http://www.k9spud.com/wiki/PCB:Etchants
** http://www.opencircuits.com/Chemical_Etchants
* I'll be getting some Sodium Persulfate from Travis.

I ended up using the muriatic acid + hydrogen peroxide version. (http://www.instructables.com/id/Stop-using-Ferric-Chloride-etchant!--A-better-etc/)

Chemical resistant gloves. Latex? Rubber dish gloves?
* Rubber dish gloves with the "water catcher" fold on the end.

I bought a box of Nitrile gloves.

Tinner? Flux?
* "liquid tin" from MG Chemicals is available from Norvac (from drtwist)
* Tinner probably isn't necessary with liquid flux.

Liquid tin makes the boards pretty and keeps the copper from oxidizing, which it _will_ do.

A drill press to drill holes (or I could just go 100% SMD)
* Dremel tool (already have it)
* Dremel Rotary Tool Work Station (order from Amazon) http://www.amazon.com/Dremel-220-01-Rotary-Tool-Station/dp/B00068P48O
* Drill bits. Either micro-carbide (brittle, but works great with a dremel and FR-4 board), or steel (heats up and distorts. Can rip up traces on the board.) Harbor Freight in Portland seems to be a good place to get them. http://www.harborfreight.com/cpi/ctaf/Displayitem.taf?itemnumber=34640

Snips to cut the board with. FR4 dust is carcinogenic, so you want to make as little of it as possible.

An iron or laminator. I bought and modified a GBC-H212 laminator.

The Method:
* http://www.riccibitti.com/pcb/pcb.htm

The Evil Benito Mouse

Laen — Fri, 03 Oct 2008 19:00:45 -0500

To take advantage of the Benito's small size, I've zapped mine with a piece of software that, when plugged into a computer, causes the mouse to jump about randomly.

The size of the jitter can be adjusted by setting the "STEP" constant. 10 seems to be a pretty good number to make it seem like there's something wrong with your mouse. Anything larger, and it's pretty obvious that you're being messed with.

Wii Matrix Board

Laen — Tue, 13 May 2008 20:21:41 -0500

For dorkbotpdx last night, I made this little toy:

It's an 8x8 red+green LED Matrix, driven by two shift registers and a Bare Bones Board Freeduino by Modern Device Company.

Before the meeting I just had it flashing fun patterns based on odd bit-math equations. During the meeting, I was urged to add some code to make it controllable by the Wii Nunchuk. So, I wrote a simple program that takes the X and Y of accelerometer and plots it on the matrix.

I drove the LED Matrix with three shift registers: One 74HC595 8-bit register gave power to the rows of the display, and one Maxim-IC MAX6969 sank current for the columns to the display. To turn on row 1, column one, you shift "B10000000" into the row register, and "B10000000" into the column register:

B 1 0 0 0 0 0 0 0 1 x o o o o o o o 0 o o o o o o o o 0 o o o o o o o o 0 o o o o o o o o 0 o o o o o o o o 0 o o o o o o o o 0 o o o o o o o o 0 o o o o o o o o

Altogether, it was about 30 lines of code to drive the matrix with some fun patterns.

Also seen at:
my blog