by on Avr Development

I like wordpress.

It allows me to do much of what I post on the web without having to look at the underlying html and still letting me at the html. In fact I use WordPress to to post on Dorkbots Drupal pages. It is easier than hand rolling html and the new wordpress saves your drafts. This is no small issue: As I was reminded at 2:30 monday morning as Drupal timed out the session that I was writing on and ate my post. Between that and the Eagle files I was working with I lost most of sunday nights sleep. The other issue is portability.  So last night I ran up mysql, unpacked the latest wordpress into my home directory and reconfigured the apache daemon that comes with leopard.

I plan to get more done and loose less sleep.

by on Avr Development

This has been reposted on dorkbotpdx.org

A few weeks ago I bought the last of the Really Bare Bones Arduino rev A boards from Brian at Wulfden (http://wulfden.org/freeduino/freeduino.shtml) I now have enough boards for one more workshop and then we have to reevaluate the boards which are avaliable. Pictured below is a finished board from the first “Arduino Cult Induction Workshop”

In addition to letting me clean out his new old stock Brian threw in one of the “Rev B” boards so I could check it out. Unfortunately I don’t like it.

The new board is almost 3/4″ of an inch longer than the original. Most of the new space is dedicated to room for a power jack and a regulator. Things that I never use. The design is supposed to allow you to cut the power jack off as well as the regulator. Notice in the finished board above that their is space for a big old reset switch. When the original Arduino came out you had to reset the board to get it into the bootloader. However in the past year or so all the new boards have been using the dtr signal run through a capacitor to pull the reset line so the switch is wasted space. The original board made up for this space by adding ample ground and supply connections. The new board adds another 1/8″ and gets rid of these.

As I was grumbling about this Mark Gross suggested that I just roll my own.

So we are back to the drawing board. Above is my pen markup of the changes that I wanted to see done to the original Really Bare Bones Arduino and below is the draft of the rework to scale with the RBBB rev B.

Now comes the fun part The cost of goods sold and a business case.

by on Dorkbot


(this is part of an ongoing DorkbotPDX workshop series being sponsored by TempusDictum, Inc.)

A few weeks ago I bought the last of the Really Bare Bones Arduino rev A boards from Brian at Wulfden (http://wulfden.org/freeduino/freeduino.shtml) I now have enough boards for one more workshop and then we have to reevaluate the boards which are avaliable. Pictured below is a finished board from the first “Arduino Cult Induction Workshop”

In addition to letting me clean out his new old stock Brian threw in one of the “Rev B” boards so I could check it out. Unfortunately I don’t like it.

The new board is almost 3/4″ of an inch longer than the original. Most of the new space is dedicated to room for a power jack and a regulator. Things that I never use. The design is supposed to allow you to cut the power jack off as well as the regulator. Notice in the finished board above that their is space for a big old reset switch. When the original Arduino came out you had to reset the board to get it into the bootloader. However in the past year or so all the new boards have been using the dtr signal run through a capacitor to pull the reset line so the switch is wasted space. The original board made up for this space by adding ample ground and supply connections. The new board adds another 1/8″ and gets rid of these.

As I was grumbling about this Mark Gross suggested that I just roll my own.

So we are back to the drawing board. Above is my pen markup of the changes that I wanted to see done to the original Really Bare Bones Arduino and below is the draft of the rework to scale with the RBBB rev B.

Now comes the fun part (not) The cost of goods sold and a business case.

by on Dorkbot

Note: This work is ongoing see the Benito info page for current info…

There were a few lessons that I learned at the Arduino Cult induction workshop that I put together this month. One of which was that I needed to simplify my programmer design on the cable end and not wait until I had a full blown product. Revisiting the original I first revised the ftdi boards to use a pinout compatible with the programming end of the RBBA (really bare bones arduino). Then I went back to the AT90USB162 based programmer modified the schematic to reduce the parts count.

Then I made it fit into a similar profile.


Link to positive image at 600 percent

Then I put together a parts manifest at q25 and found that in spite of the increased parts count it is actually cheaper than the ftdi boards.

Index Quantity Part Number Description Customer Reference Backorder Quantity Unit Price
USD Extended Price
USD
1 50 RHM10KARCT-ND RES 10K OHM 1/8W 5% 0805 SMD 0 0.02340 $1.17
2 25 AT90USB162-16AURCT-ND IC AVR MCU 16K FLASH 32TQFP 0 3.15000 $78.75
3 25 631-1099-ND CRYSTAL 8.0 MHZ SERIES 0 0.48900 $12.23
4 30 PCC220CNCT-ND CAP 22PF 50V CERM CHIP 0805 SMD 0 0.06900 $2.07
5 200 RHM220ACT-ND RES 220 OHM 1/8W 5% 0805 SMD 0 0.02340 $4.68
6 50 RHM22ACT-ND RES 22 OHM 1/8W 5% 0805 SMD 0 0.04080 $2.04
7 100 399-1168-1-ND CAP .10UF 25V CERAMIC X7R 0805 0 0.02670 $2.67
8 50 399-1284-1-ND CAP 1.0UF 16V CERAMIC X7R 0805 0 0.09500 $4.75
9 30 475-1401-ND LED 3MM 570NM GREEN DIFF RADIAL 0 0.05600 $1.68
10 1 AT43301-SU-ND IC USB HUB CTRLR 4PORT 24SOIC 0 1.83000 $1.83
11 25 609-1039-ND CONN RCPT USB TYPE B R/A PCB 0 0.54200 $13.55
Subtotal $125.42

The eagle files are attached below.

Firmware is also attached.
This requires Dean Camera’s MyUSB library (>1.4.1 along with a recent avr-gcc toolchain and dfu-programmer)
To build it; Create a directory in your MyUSB tree on the same level as the Demo directory named Projects.
Unpack the code into that directory, make and make program.

by on Avr Development

There were a few lessons that I learned at the Arduino Cult induction workshop that I put together this month. One of which was that I needed to simplify my programmer design on the cable end and not wait until I had a full blown product. Revisiting the original I first revised the ftdi boards to use a pinout compatible with the programming end of the rbba. Then I went back to the AT90USB162 based programmer modified the schematic to reduce the parts count.

Then I made it fit into a similar profile.

Then I put together a parts manifest at q25 and found that in spite of the increased parts count it is actually cheaper than the ftdi boards.

Index Quantity Part Number Description Customer Reference Backorder Quantity Unit Price
USD Extended Price
USD
1 50 RHM10KARCT-ND RES 10K OHM 1/8W 5% 0805 SMD 0 0.02340 $1.17
2 25 AT90USB162-16AURCT-ND IC AVR MCU 16K FLASH 32TQFP 0 3.15000 $78.75
3 25 631-1099-ND CRYSTAL 8.0 MHZ SERIES 0 0.48900 $12.23
4 30 PCC220CNCT-ND CAP 22PF 50V CERM CHIP 0805 SMD 0 0.06900 $2.07
5 200 RHM220ACT-ND RES 220 OHM 1/8W 5% 0805 SMD 0 0.02340 $4.68
6 50 RHM22ACT-ND RES 22 OHM 1/8W 5% 0805 SMD 0 0.04080 $2.04
7 100 399-1168-1-ND CAP .10UF 25V CERAMIC X7R 0805 0 0.02670 $2.67
8 50 399-1284-1-ND CAP 1.0UF 16V CERAMIC X7R 0805 0 0.09500 $4.75
9 30 475-1401-ND LED 3MM 570NM GREEN DIFF RADIAL 0 0.05600 $1.68
10 1 AT43301-SU-ND IC USB HUB CTRLR 4PORT 24SOIC 0 1.83000 $1.83
11 25 609-1039-ND CONN RCPT USB TYPE B R/A PCB 0 0.54200 $13.55
Subtotal $125.42

by on Dorkbot

First of all I was really happy that so many people showed up and got as far into their construction as they did.

27 people and only one dead board.

I now know that getting all of that done in under 3 hours was a little ambitious. (The pacing was initially based on the photo session at http://www.flickr.com/photos/7175086@N05/sets/72157604783725339/ which took 20 minutes from start to finish)

For those of you who were unable to finish the cable Jacob Hedwig has provided a clear picture of a working cable at

Arduino Cult Induction

Pin 1 is on the bottom (Red wire in this picture) on both ends.

Pins -> USB
1 -> 2 (Rx)
2 -> 1 (Tx)
3 -> 7 (+5V)
4 -> 4 (Reset)
5 -> 3 (Ground)

The important thing is to count up from the side of the cable with the leds on it.

I am looking to have a second session for software as soon as I can get a room scheduled (this sunday or some sunday soon). In the mean time you can download the software (http://www.arduino.cc/en/Main/Software) and poke around at the main site (http://www.arduino.cc/). I will also be at the bi-weekly dorkbot meeting at the Lucky Lab NW tomorrow (May 12, 19th and Quimby) between 7 and 10 and will bring a soldering iron if anyone gets stuck or wants to finish up their board.

Feel free to email me as well.

by on Dorkbot

Note: This work is ongoing see the Benito info page for current info…

Introduction.

The Benito Board is an at90USB162 board intended for use in programming and communicating with microcontrollers which have serial based bootloaders.

Among those are the Phillips lpc21xx series ARM chips, the Dallas Semiconductor, DS500x family and Atmels using any number of STK500 compatible bootloaders.

In the case of the Atmel and Phillips chips, a reset pulse is required to put the device into programming mode. A logical trigger for this pulse is the DTR signal which is pulled low when the computer begins to talk to a given device.

Background

Recently on avrfreaks.net….

Smileymicros wrote:

Now that they have the AT90USB82 for $2,17 each per hundred at DigiKey, I’m forced to reconsider. Why the f*ck would I continue using the FTDI part which costs $3.89 each per hundred?Smiley

It was not the first time I had had this question posed. Paul Stoffregen was talking about the new at90usb82/162 a few months back; its cheaper than the ftdi232 and has at least 2 driverless public usb-serial implementations: Dean Camera’s and Atmel’s. The best part is that it has a built in usb based bootloader and its a fully programmable AVR microcontroller. I thought it would be perfect to replace the ftdi ft232 series chips and to build an stk500 programmer with built in usb to boot. It seemed like the thing to do for many reasons.

Hardware

The sample code from both Atmel and MyUSB are based mostly on the AT90USBKey so I started with the schematic from that and cut out everything that I didn’t need. (a careful rereading of the datasheet says I am missing a 1uf cap here can you find it?).

A little hard to read

Click for larger image.

Compared to the the ftdi boards I have been working on the parts count is pretty high but I am banking on flexibility to make it worth while.

board positive

Click for link to positive at 600%

For the prototype I used my usual 1.5 sided board technique where the ground plain is brought to the top of the board and the alignment can be off by a couple of mm. The leds are some super bright ones in a plcc4 package that I have several hundred of. The pads in the eagle library for them are HUGE. I turned over the eagle files to Monty Goodson of bittybot.com who has lots of experience at getting things to be small so hopefully the final board will be a lot smaller.

In the mean time feel free to use the Positive image above to create your own board

The Programming Header

At the moment I have a very AVR centric view of the universe. For that reason the programming header that made the most sense was a combination of a serial connection with the 6 pin isp connection. For target boards based on other processors such as the DS5000 and the LP21xxx the SPI pins can be repurposed and the software adjusted accordingly.

Using the built-in boot-loader

When you get power, usb, crystal, reset and hwb wired you can enter the chips usb boot-loader by holding HWB low during reset. If you have a mac like me and look look at the usb section of your system profiler you should see the following:

device AT90USB162
DFU:Version:    0.00
Bus Power (mA):    500
Speed:    Up to 12 Mb/sec
Manufacturer:    ATMEL
Product ID:    0x2ffa
Serial Number:    1.0.5
Vendor ID:    0x03eb

You can now program the chip using Atmel’s FLIP utility if you are running windblows (sic) or Weston T. Schmidt’s dfu-programmer on sourceforge http://dfu-programmer.sourceforge.net/ DFU stands for Device Firmware Update.

Running dfu-programmer looks like this.

$ dfu-programmer at90usb162 erase
$ dfu-programmer at90usb162 flash --debug 20 USBtoSerial.a90

target: at90usb162
chip_id: 0x2ffa
vendor_id: 0x03eb
command: flash
quiet: false
debug: 20
device_type: AVR
------ command specific below ------
validate: true
hex file: USBtoSerial.a90
Validating...3182 bytes used (25.90%)

$ dfu-programmer at90usb162 start

I made the .a90 file from my modifications to Dean Camera’s sample code (to be described later) with the following voodoo.

avr-objcopy -R .eeprom -O ihex USBtoSerial.elf USBtoSerial.a90

Now on top of the blinking lights on the board an entirely different device is attatched to my usb port.



The “Driverless” Serial Device.

The USB standard is a lot about throwing several thousand dollars down so you can play as a member. Fortunately for the rest of us there are a couple of generalized devices that have generic definitions. one is the HID (human interface device) and the other is a CDC communications device under the CDC a class of devices is defined which looks like a modem (ACM). If a device says its one of these and acts accordingly most well developed operating systems will load a generic driver for them and just work. (and then there is windows which still requires a .inf file to load the generic driver).

Atmel’s Code Samples.

As much as I love Atmel’s processors it is a constant source of irritation that their best tools are only made available on the windows operating system. It was nice to see that the sample code for Atmel was available for avr-gcc and I ran it up only to find myself picking through 5 layers of slashes “”. This is not just a preference, it is bad coding. (to quote an old co-worker of mine “Its Rubbish — Bin it!)

So I ran it up made it work and started looking at alternatives.

This lead me to Dean Camera’s MyUSB library. Though it is still in development it is much easier to understand and work with than Atmel’s Code.

crtusb162.o

When compiling Atmel’s CDC code I wound up with a linker error complaining that crtusb162.o did not exist After recommending AVRMackPack for nearly 6 months I thought I had found my first bug. After rebuilding my own version of the tool chain and digging around I found that it is a known bug and should be fixed in future releases. In the mean time the fix is easy. Either copy it from the avr3 directory into your source or link the avr3 directory to avr35 where the linker should find it.

#cd  /usr/local/AVRMacPack/avr-4/lib/
#ln -s avr3 avr35

Building the Firmware with MyUSB

The firmware linked below under resources implements a USB serial port which produces a 2ms *reset pulse instead of DTR. It is based on the USBtoSerial Demo from version 1.3.2 of the MyUSB library. To build it unpack the MyUSB library into your work area. Then under the MyUSB library directory create a directory called Projects and unpack the Benito firmware into the Projects directory. If you have xcode (v>2.4) and AvrMacPack you can open the project and build it. Other platforms will have to adjust the Makefile to match their environment. the “make program” target defined in the make file uses the dfu-programmer utility so you can program the device using its built in boot-loader.

Results:

By using Dean Camera’s MyUSB library I was able to create a serial programmer which plugs into Linux and OSX and it “just works”. I was then able to produce the reset pulse in firmware and it continued to “Just Work”. You can see it above plugged into a modern device RBBA board running Limor Fried’s AdaBoot. I was able to do this in a very short amount of time and while the hardware is slightly more complicated than the ft232 boards the cost at 100 boards is actually cheaper.

Moving Forward

In the photo below you can see my prototype board (both sides) next to an ft232rl based programmer with a .1uf capacitor “auto reset hack”. The board on the bottom shows a programming connection between an at90USB82 and a phillips Arm chip.

On the Atmel side the next step is to implement an STK500 based programmer using the SPI port on the at90USB162. I expect that this will be a fairly easy adaptation however first I may revisit the DS500x chips that I have and figure out the best way to switch between the programming and communication modes. From there it would also be a no brainer to adapt the firmware for the lpc21xx chip pictured above.

Resources


						

by on Avr Development

It seems like all I do nowadays is to try to decipher other peoples code.

Which make a body feel a lot like a teething toddler.

One of my projects is an an avr programmer based on the new AT90USB  using Dean Camera’s MyUSB library. In it there are two examples using the driverless serial class (CDC/ACM). The examples are deceptively straightforward.

I was able to bring up a functioning serial port which works on OSX, Windows, and Linux in a very short time. The problem was that I needed to implement the DTR and what I eventually want to wind up with is a reset pulse where the DTR pin used to be.

My present understanding of the way that the USB CDC Modem class handles RTS and DTR is that the host computer sends a control packet to the device telling it the state of those lines. This is described in  the document entitled “Universal Serial Bus Class Definitions for Communication Devices”
http://www.usb.org/developers/devclass_docs/usbcdc11.pdf. and should be handled in the USBtoSerial code in the file USBtoSerial.c

After much wrestling to understand Deans implementation of “handlers” I came up with the following fragment of code  which turns on led3 when the dtr line should be pulled low.

#define SET_CONTROL_LINE_STATE_RTS_MASK 0x0002
#define SET_CONTROL_LINE_STATE_DTR_MASK 0x0001
EVENT_HANDLER(USB_UnhandledControlPacket)
{
uint8_t* LineCodingData = (uint8_t*)&LineCoding;
uint16_t wValue;
//Endpoint_Ignore_Word(); <- cant do this
wValue = Endpoint_Read_Word_LE();
/* Process CDC specific control requests */
switch (Request)
{
 case ....
 .... break;
 case SET_CONTROL_LINE_STATE:
 if (RequestType == (REQDIR_HOSTTODEVICE | REQTYPE_CLASS | REQREC_INTERFACE))
 {
  if (wValue & SET_CONTROL_LINE_STATE_DTR_MASK) {
    LEDs_TurnOnLEDs(LEDS_LED3); //handle DTR
  } else {
    LEDs_TurnOffLEDs(LEDS_LED3);
  }
  if (wValue & SET_CONTROL_LINE_STATE_RTS_MASK) {
     LEDs_TurnOnLEDs(LEDS_LED1); //handle RTS
  } else {
     LEDs_TurnOffLEDs(LEDS_LED1);
  }
  Endpoint_ClearSetupReceived();
  Endpoint_Setup_In_Clear();
  while (!(Endpoint_Setup_In_IsReady()));
 }
 break;
}

Two problems I had with grocking the original code were the presence of the mystery variables “Request” and “RequestType”. These variables are created by the macro EVENT_HANDLER(USB_UnhandledControlPacket). As this is a new library the documentation doesnt cover the nuances of the public interfaces build into it.

The other was finding the data. I never did really figure out how I was supposed to figure this out but what I did find in the file DevChapter9.c was a comment

Endpoint_Ignore_Word(); // Ignore unused Value word

Which I realized was the wValue word from the usb documentation.

Now I need to figure out enough about MyUSBs task manager to make a pulse and let the blinking lights stay on long enough to see them when sending and receiving data.

by on Dorkbot

I sent my son to his mom’s with one of the USB Serial boards I built along with a breadboard, an RBBA a pile of resistors and an led array. I thought about this arrangement and decided that I would try to build a breadboard attached programmer that would not be dangling off the edge of the board and in the way.

I am wondering if this isn’t the way to go with the seminar kits.

Board Positive (link to rev A board at 600%)

I am considering a rev B. The tx and rx are reversed and I think I should put 3 pin headers on the power rails. I am also thinking about putting the 3.3v on the inner rail and the 5V on the outer rail on the positive side.

I also made a couple of these but I am missing something and this one doesn’t work (yet)