a hobby class on microcontrollers

I don't know if it was already mentioned in this thread:

People who don't have much computer experience sometimes have problems with complicated toolchains and pure C, so the Arduino project might be easier for them. Writing your own user friendly IDE would be much work.

I don't know Arduino myself, because I like full control and low-level access to microcontrollers, even in assembler, but if your goal is something like a pottery class for people who have never done something pottery-like before, it might be better to concentrate on the projects (blinking LEDs, simple robots with sensors) instead of the development environment, because only strange people like programmers are interested in the technical aspect of programming, IDEs, assembler etc. :-) most people are just interested in the results and don't want to learn many low-level things.

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de

The problem is creating something to capture their interests.

We, programmers, are interested in blinking LEDs because it's the easiest way to verify program logics. Blinking LEDs might not be so interesting for none programmers.

Not easy to do with low cost.

Oh. I see. I probably wasn't following, then. Yes, I need to solve this.

One possibility is to use USART peripheral on the micro side tied to a USB/RS-232 adapter that ties into a PC and then using any terminal program I choose to talk. I just use dumb terminal code and the HID driver in Windows. Another might be to use a micro chip that includes the USB interface and necessary software to appear as a slave serial port device and avoid entirely the extra adapter. Again using that dumb terminal program under Windows or Linux or whatever. Still another might be to adapt something like the ZipIt2, which has a 312MHz XScale PXA270, a color, backlit TFT LCD with

320x240 TFT LCD and more color range than I care about, 802.11b/g, 8Mb flash, 32 Mb sdram, stereo DAC audio, a qwerty keyboard, status LEDs, clam shell case, joystick rocker, micro sd card slot, etc. I've read these can be had in 1's for about $35. Runs Linux. This may work out, as well.

Nearing optimal would be something quite similar to the $4.30 TI board but supporting a DIP packaged micro with somewhat more memory than the value line carries, so far, _if_ that board which plugs into a USB slot _also_ had the ability to act as a HID serial port device to the DIP's USART pins AND a JTAG debugger, both as appropriate.

Something very close to this already exists in the USB programming stub found in the TI Chronos watch kit (for which I paid $24, shipped.) This device includes the USB slave device and a separate CPU dedicated to supporting JTAG and serial port via a tiny 6-pin header. 4 pins are for JTAG and the outside 2 are for Rx and Tx. When the device is plugged into the USB port of a PC, it sets up as BOTH a JTAG downloader/debugger AND a COM port on the PC at the same time. Anything that works with the JTAG and the Rx/Tx pins can then be plugged into the 6-pin header, programmed, debugged, and communicated with via full duplex serial port communications programs. That could simply be a specially wired ZIF socket, I suppose.

I'm still agog over thinking this through. Lots to consider, yet.

Jon

Jon's first cost estimates seemed to push down toward the single-digit dollars area, and Arduino costs more than that. My guesstimate would be around $20..30 per unit per person. Once you pay that, you get a system that can be programmed to run stand-alone, a USB programming interface, a simple language and programming framework, and versions of the usual peripherals -- ADC, PWM, digital I/O, UART serial. Hooking up to lash-ups of outside equipment is what they're for.

You can use the Arduino hardware outside the framework, but at that point it's cheaper to attach an ATmega_8 to a board and do precisely what you want for yourself.

Mel.

I have been looking at the Arduino for a community ed class. I figure by the time I get an Arduino, large breadboard, a handful of leds, pot (for the adc), and some sensors it will total in the $50-$60 range - quite a bit above Jon's hope. But with the hardware and the development system it would be a simple and powerful system. For standalone projects they could use the Arduino for development and then get a cheaper board to plug the programmed chip into.

The Arduino tool chain is actually has a preprocessor that targets the AVR GCC tool chain. The sketch "language" is a subset of C,C++ that includes a number of predefined functions to drive the on chip peripherals.

More advanced students can write C and C++ code that will be passed on through to GCC/LD. In fact I have included my own C++ classes as part of the sketch source and called them from the loop code.

If you look at the code for the predefined functions you will find that much of it is C++ classes for the peripherals with simple interface member functions so that the end user is completely shielded from the details.

Jim, are you sure about that? If I get what Jon is after (without having followed the complete thread) he wants to get people interested in that thing. "Click here choose that" is nothing new for most people. I suppose all of us have started with non-source level debuggers, memory dumps etc... If Jon manages to get some people understand *that*, it will be a huge success on his part. Once past this everyone will understand the rest alone if interested.

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

------------------------------------------------------

You are right, looks like it is not really a mass product, so the boards are expensive. Of course, you could buy an ATmega328 in DIP package, 25 pieces for $2.87:

a breadboard for $5,70 (maybe less expensive, if you buy 25, too)

and finally an USB to RS232 TTL level converter for $1,50:

Ok, that's $10,07 and you need additional parts for interesting projects, and wires and at least one ATmega programmer to program the Arduino bootloader, but nearly in the single digit area.

Maybe the MSP430 launchpad for $4.30 is better, which include the USB connection and the USB cable:

and there is already lots of sample code available. The only drawback would be the slightly more difficult development environment. You could solder some headers to it, for using it with a solderless breadboard, if you don't want your class to solder.

If you think soldering is part of the fun, it could be less expensive.

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de

I was thinking of something like PWM output, RBG LED and a table tennis ball, or mounting it on a fast rotating base :-)

You need some fantasy. Something classical: Two motors and two LDRs is all you need for some fun:

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de

The launchpad is so cheap because TI is subsidizing the "demo tool" to generate sales.

If you would approach TI with an idea to "get budding engineers started" you may get them to design a board for you.

I am sure TI is selling 10s of thousands of these boards. A price break you will never see building it yourself.

TI also gets a tax break for advertising.

Using this board as a standard is a little short sighted.

hamilton

Does any one have one of these boards handy ?

Looking at the pics on the web site, I see four chips next to the mini-USB connector.

What are these chips ?

This board is no $4.30, even if its is a "430".

h

There is a schematic on the webpage: MSP430F16x for the emulator and the programmer interface, TUSB3410VF for the USB to RS232-TTL interface, TPD2E001 for ESD protection on the USB port, TPS77301 LDO power supply and CAT24F32UI, an EEPROM. Except the EEPROM, all chips are from TI, so this could be very cheap for TI.

But you are right, it doesn't look optimized for price (why not using a microcontroller with integrated USB?) and maybe manufacturing costs are higher (the "4.30" as a price is a nice marketing gag). But if they produce

10s of thousands of it, maybe they can sell it at least without loss.

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de

Loss does not matter, its a marketing deduction.

But it might be close.

h

Probably so.

Worth trying. However, I think I need to have a much clearer and much more fully fleshed out "business plan" to present, then. I don't need to convince me. But I will need to convince them if I'm asking them to invest their money.

Still, I like the thought. I will keep it near the top, for a while, and may wind up giving it a shot.

Indeed. Speaking of which, I _really_ like the USB stick (actually, there are two different ones but I'm speaking of one of those two) that comes with the Chronos watch. It includes a power/gnd + 2-wire JTAG debugger (4 pins) but also another 2 pins for Rx and Tx... 6 pin connector. This comes up under Windows as both a supported JTAG debugger and flash programmer as well as a COM port. With suitable cable and connectors, this provides something the students can take home and use on their own. Since I bought that Chronos for $24 shipped, and since almost the same thing comes on their F2013 demo unit which at times sells for $9.99 shipped, I might be able to worm those out of them as solo pieces I can use. Might be worth a conversation, anyway.

Better, in some cases I'm thinking about, I can provide them with a 501(c)3 for charitable deductions (if I can convince the board here to go for it or otherwise form up a separate one just for this purpose.) That's work, too. But with incentive on both sides, it's a crossable bridge.

Well, one does need to put one foot in front of another and make some mistakes. I have never once seen a 5-year plan for a new company that came anywhere close to where they were 5 years later. Or even 3. Doesn't mean there aren't some very experienced people adressing themselves to thoroughly mature markets where everything is pretty much known. But I haven't ever worked in those areas and, frankly, that doesn't apply here at all. In this sphere, the chances of me knowing most of what this will look like 3 years from now, assuming there is anything at all, is.. low. So short-sighted is about the only way to go. It's sparsely mapped waters.

I hope for a usable lookout tower, manned as best I can, and a lot of unplanned course corrections along the way, I think. Unless I can convince others to sit down with me and add their own experience and good knowledge to identify the 'sizzle' for this as well as the core meat in it and we form the right combination at the right time and somehow also get lucky. Since I'm one person and control no others at all, all I can really be sure of are my own hours.

Good thoughts about contacting TI, though. I will put that on my list and follow through when the moment is right.

Jon

I see a number of important elements here

• engage the student

• Allow different learning rates

• Do not hold back the smarter ones.

and yes, that makes source-level step/watch type operation, important. This was what made TurboPascal so popular.

The point I was trying to make, is if the low cost HW allows this, do not side-step it by language choice.

With modern Syntax highlighting editors, and close to the same appearance in Debug, novice users do not need memory dumps ;)

I did see a nice terminal/message-to-pc Pathway in that low cost TI system, which is a great way to keep the HW costs down. Users can see real run-time actions, coming back on the screen - of course, some Flash/Klunk stuff is still needed, so they know the PC can be disconnected!.

-jg

We can let the students control some GP I/O ports via USB without programming.

I have posted the Window Apps wdip.exe (talking to com6:), Inf file for the virtual com port and the USB firmware. Wdip connects to the virtual com port and monitor/control 10 GP I/O bits. You can click on the second rows of buttons to toggle input/output direction. For output ports, click on the third rows of buttons to turn it on and off.

My virtual com port is com6:. I still have to work on auto-detecting the com port. For now, i will just post wdip4 (com4:), wdip5, wdip6, etc.

And yes, i know, i have to line up the buttons.

The board looks nice and is in the price range of the OP. But the demo application doesn't sound like an interesting project for students. Maybe enhance it with a simple BASIC interpreter on PC and on the microcontroller?

BTW: the virtual COM port can cause problems in Windows. I've used for a customer project LUFA and a generic HID device implementation, which works on Windows without the need to install INF files or other drivers.

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de

n

Yes, that's the plan. He will do one for the micro and I might (if I have time) do one on the PC. I might do a 6502 ICE on the PC, since the 6502 instruction set is in the public domain (I think).

s

It occasionally causes the usb stack to get struck. Fortunately, it can recover by unplugging/plugging the board without rebooting window. We might switch to win32-libusb eventually.

The real status of the 6502 instruction set is a little fuzzy. The Western Design Center have made some patent related claims that have now probably expired

I have seen a lot of 6502 based processors pop up in asia in some cases with opcodes swapped around for all the instructions except for branches.

Regards

walter..

-- Walter Banks Byte Craft Limited

Yes, indeed. We can order 6502 look alike embedded core in ASIC. They don't work truly like the 6502, but close enough. In fact, we need to build the ICE to target an OTP ASIC.