Projects pros and cons

The basic premise of the project is as follows;

1. The oscillator is to be controlled through the Flex Kenwood CAT command set through a serial port or a USB port with the appropiate hardware. ASCII commands would be sent to the controller asking for status or asking that a new frequency be set.
2. Be able to use pushbuttons, joystics, encoders and similar means to manually control the oscillator.
3. Be able to permanently contain settings for band limits and have control of relays to activate and select front end filters.
4. When project is finalized, create PCB using SMT components to contain entire system, OSC+Band Filters in a small "black box" componet to be used in multiple projects.

Since this would be a not too tiny yet not too massive project that can be done in stages I will be using it as a test vehicle to play with different CPU chips that I'm not too familiar with. So what are the CPU choices their advantages and disadvantages?

PicAxe specifiucally the 28X1 chip

1. + Inexpensive and easy to work with, available in different sizes to meet project requirements. can be programmed 100K times, has I2C for the 18X and higher chips.
2. + Easy to program, free programming tools available, it uses a built in Basic interpreter, lots of user groups with code and assistance.
3. - Lacks the built in data types to make the large number calculations required to program the SI570. This in itself is a major minus since I'm determined that the PC will not know anything about the device it's controlling, it just sends a frequency command and the controller's job is to figure out how to comply.

PIC controller in the 18F  family, specifically the 18F2550

1. + Low cost ($5), yet very capable 32KB Flash, built in USB port a big plus, 48MHz, EEPROM, I2C, SPI, Serial
2. + Free development software (C and Assembler), inexpensive debug hardware.
3. + Free resident FORTH Compiler/Interpreter available, a big plus since FORTH is ideal for hardware development.
4. + Inexpensive prototype boards available from SparkFun
5. + Chip is available in DIP and QFP versions to make development easier.

AVR ATMEGA series, specifically the ATMEGA168 CPU

1. + Low cost ($3.5) very capable, 16KB Flash, EEProm, I2C, SPI, 20MHz, Serial
2. + Free development software (C and Assembler) plus Pascal, and Basic, inexpensive debugging tools.
3. + Possible free FORTH resident Compiler/Interpreter (I have seen it but I don't remember where)
4. + Inexpensive very capable Basic compilers are available for less than $100
5. + Inexpensive development boards from SparkFun
6. + Chips available in DIP and QFP for easier development.

AVR Butterfly

1. + Same advantages as the AVR ATMEGA chips mentioned above
2. + Built in display and input buttons including a joystick and serial port to program and use in controlling the device.
3. + Inexpensive prototype board available from Smiley's to hold the Butterfly, add voltage regulation, and have a decent breadboard area.

MSP430 family specifically the 430F1213 and the 430F1611 which is way overkill but I have other interest in using this chip.

1. + Parts are inexpensive, more powerful it's 16 bits instead of previous 8 bit CPUs, 48KB Flash, EEProm, 10K RAM, I2C, SPI, Serial
2. + Free development software (C Assembler), inexpensive debugging tools.
3. + Free Forth compilers available, runs Forth very efficiently due to it's 16 nature and it CPU architecture.
4. + I own a copy of the ImageCraft Professional C compiler for the MSP430 family.
5. + Even though DIP packaging is not available SparkFun has inexpensive header boards.
6. + Inexpensive development boards available from SparkFun
7. + I'm more familiarwith it than any other choice, this CPU has the most easy to work with instruction set, I love it.
8. + I need to finish my optimizing Forth compiler for this CPU. (Fun Factor)

Which one I will use first? I'm not sure but I lean towards the Butterfly, since it will benefit a lot more people, but then there are probably a lot of people working on similar projects with the Butterfly already. The one I find more interesting is the PIC 18F2550, I have never worked with a PIC before and I'm wondering how easy it is to work with.

During the Holiday time off I'm trying to decide so I have been installing the different development systems, and I will be connecting my CPU prototypes to the PC and see which one is interesting and fun.

1. First Project, low cost stand alone unit.

This will be a project that I outlined in an email to the SoftRock group. This is to be a low cost controller for the SI570 that requires no PC. Initial plans is to store a table of frequencies, the CPU will calculate the parameters for the SI570 itself.

It will in it's final form have the following features;

1. 16 bands and 4 LEDs to display in binary which band is active
2. Up to 64 frequencies in a band, with 6 LED bars to show the selection in binary.
3. Use of EEProm to remember last selection after powering the unit off.
4. Frequency stored as 4 bytes as the number of hertz, and bandpass filter information
5. Optional output pins to activate band relays.
6. Optional crystal with frequency trimmer instead of cheap resonator to have an accurate oscillator for counter feature.
7. Optional connection to a 2x16 display to display frequency.
8. Optional frequency counter mode if connected to a 2x16 display
9. Optional self calibrating, will read output of SI570 and adjust a SI570 compensating number to have the frequency be more accurate, no display required, but will require trimmed crystal.
10. Possibly bigger ATMega chip to accomodate optional features.
11. 2 to 4 buttons to select frequencies and modes

If options raise the price too much then I would implement options 6 and 9 foremost and the others as resources permit.

It consist initially of the following hardware;

1. ATMega88 CPU (28pin PDIP), later maybe switch to a ATMega48hich cost $1.60 in Qty 25 if it will fit in 4K of flash
2. A LED bar 10 segment bar graph for the display, 4 bars for the band information, 6 bars for the frequency in the band in binary form.
3. Optional 2x16 display.
4. A SI570 chip + miscellaneous components.

01/05/2008

I started this morning putting the hardware together and will try to communicate to it later today. I'm having problems with the development system not taking to the board, most likely it's a wiring problem, will continue tomorrow when I'm not so tired.

01/06/2008

After beating my head against a wall and getting nowhere, I could not find anything wrong with the CPU board  I decided to try a development board from Sparkfun so I can sort out where the problems are. Well, I can't talk to it either. I then tried a different compiler and IDE and it could not talk to the chip either so it seems I'm having a problem with my serial port. Everything else seems OK so I will have to dig out a laptop or something that has some serial ports. The PC I used is the only one that has a serial port. Iwas having problems prior to this talking to a PicAxe so now I know why. I will try a USB to serial adapter and see if that works(they rarely do for programming chips).

01/07/2007

It started acting up again then the PC would not boot turns out the serial port is fine it's the PC going of the deep end of ocean. It stopped booting but worse it had a BIOS error saying the Plug and Play could not enumerate the devices, sounds like a root cause. I changed some settings in the Bios and the error went away, but the PC would still not boot except in safe mode. Afer re-installing the Video, USB, and chipset drivers the PC settled down and seems fine.

Went back to working on the hardware and it's working fine I could download shoftware and run programs. Tomorrow is back to writing the software.

01/19/08

I have been pretty busy at work and had little time but some progress has been made;

I started evaluating different software development systems for the AVR chips not just for this project but it's something I needed to do anyway. There sure are a lot available. But I'm homming in on a few;

1. Bascom-AVR Basic compiler, it's reasonable ($99) generates fast code, but it's a little code hungry, it's binaries are a little bigger that the others, besides the price it has a lot of plusses. I supports a lot of data types including Double floats, has a library of tools to work with all sorts of devices, and it interfaces to PonyProg real well. Some advanced libraries are extra but reasonable.
2. MikroPascal Pascal compiler, cost about $150, generates the fastest and smallest code with the exception of MikroBasic compiler which generates a tiny bit smaller code. It has a very extensive library of routines to use the hardware features, no Double floats. Has Units for organizing the code.
3. MikroBasic same as above except it's Basic instead of Pascal
4. AVRco Pascal compiler, outrageous price ($650) but the demo version will do 8K of code, you can do an awful lot in that space,  generates slighly bigger code than the Bascom-AVR compiler, buying it is out of the question. It's library is huge, and even covers a lot of external chips. A huge plus is that it generates code with debugging information for the AVR studio debuggers.
5. GCC C development system, it's free, it will work in Windows or Linux, generates decent code, and debug information for the hardware debuggers. It has a big minus, it's C++ which I totally dislike and try to stay away from as much as possible( too many years trying to find pointer errors) . I have not tested this one yet.

All these development systems are either free of have functional demos that will allow enough code to give it a serious workout, so soon I will settle on one of them. In the meantime, systems 1 thru 4 have I2C libraries and in all four I was able to talk to the SI570 chip rather easily. They all have interfaces to serial ports, I2C, keypads, LCD displays, even Ethernet controllers so down the line I will see how that works out.

Yesterday I received a ICSP Programmer/JTag ICE debugger/DebugWire debugger and have been playing with it, I finally have it figured out after a lot of false starts and headaches, I even thought for a while that I had killed a couple of chips, it was the operator who was the troublemaker instead. I bought a second Programmer/Debugger (AVR Dragon) that can handle  all the programming and debugging schemes that the AVR chips use, but I need to buy some cables before I can use it. I also received some tiny AVR chips for use in the final project, they are cheap ($1.60), and in 14 pin DIP packages but have enough resources to do some serious applications.

Slow progress so far because I been playing with the toys instead of doing anything useful but all the pieces are falling into place, I'm starting to dissect how to calculate the parameters for the SI570 without using huge numbers But I'm having a lot of fun while doing it, that is except when I had problems with the debugger, that was no fun, but all is well now.

Useful links;

www.mcselec.com/index.php

www.e-lab.de/AVRco/index_en.html

www.mikroe.com/en/compilers/mikropascal/avr/

www.mikroe.com/en/compilers/mikrobasic/avr/

www.lancos.com/prog.html

2. Butterfly with computer serial RS232 port control or manual control

This will be a SI570 project using the Atmel Butterfly to control and display the frequency of a SI570, using the built in display and buttons.

The goals of the project are;

1. Control of SI570 using the serial RS232 port of the Butterfly
2. Respond to the Flex Kenwood command set to set/read frequencies and control of external band-pass filters
3. Optional self calibration mode if Butterfly hardware will allow it.
4. Generic use of serial ports to allow use of USB to serial adapter.

Initial project hardware will consist of;

1. Atmel Butterfly
2. Smiley Butterfly  carrier
3. Miscellaneous parts, and band pass filter relays.

This project is on standby until project #1 is well on it's way.

3. USB computer controlled SI570 with frequency counter and general purpose frequency source.

This is my final goal for this project series, consisting of a CPU controlling a SI570 and getting it's command from a PC via a USB port. This project is not just for a softrock, but will allow it to be used as a general purpose frequency generator from 10KHz to about 150 MHz

Project goals;

1. USB communitation to PC
2. Set and read frequencies from PC
3. Accurate crystal clock to allow frequency measurements.
4. Frequency counter mode.
5. General purpose frequency generator mode. 10KHz - 150MHz
6. Self calibration mode.
7. Optional frequency multiplier for use as a SoftRock clock(4X) of a generat purpose Frequency generator(1X).
8. Controlled  divide by circuits (/1000) to allow lower than 10MHz clocks.

Hardware required;

1. Higher capacity CPU, either a 18F2550 or a ATmega16 class CPU with USB capability.
2. Output divider chain and multiplexer to select clock source.
3. The use of high speed logic to handle 150Mhz
4. Ouput to control band-pass filter switching relays.

This project is a ways off.