Coding Tool Goes with the Flow

The Flowcode 4 software from UK-based Matrix Multimedia simplifies programming tasks by substituting drag-and-drop flow-chart elements for C-language code statements. This approach works well and can get engineers, teachers and students off to a quick start with microcontrollers. So far, Matrix Multimedia offers Multiprogrammer boards for MCUs in the Microchip PIC and dsPIC families, and in the Atmel AVR and ARM AT91SAM7S families. The company also supplies a large variety of E-Blocks hardware modules that attach to the MCU boards. I enjoyed working with the Flowcode 4 software and PIC-based hardware the company provided. This software and add-on hardware allows people to create useful microcontroller programs and devices without forcing them to learn C-language syntax; at least to start. They can do productive work right away.

The Flowcode software installed easily from a CD-ROM, and Matrix Multimedia responded quickly by e-mail with an installation code that let me run the Professional version of the software. The accompanying PIC Multiprogrammer board took more time than expected to figure out because the board's power connector doesn't indicate either the necessary voltage or polarity. I discovered that polarity doesn't matter and the board needs 13.5V from the power cube, which a voltage-output selector switch let me choose. People should not have to dig deeply for this type of basic setup information. They want to start quickly without fear of blowing out the programmer board with an incorrect voltage or polarity.

The PIC Multiprogrammer came with an 18-pin PIC16F88 installed and it can also accept 8-, 14-, 20-, 28-, or 40-pin DIP PIC MCUs. I connected the board to my lab PC and installed the needed USB drivers. After that, I followed instructions in the "Getting Started Guide" to create a simple program.

The user interface displays a layout with a BEGIN and an END block already in place. When you drag a function block and drop it between blocks it inserts itself in into the program flow. An exercise in the "Getting Started Guide" explains how to select the MCU you have in a socket, enter the MCU's clock frequency, select an output port, set an output pin to a logic 1, monitor the state of that pin on the screen, and download the program to the MCU chip. These steps went smoothly. The Multiprogrammer board routes all MCU I/O ports to 9-pin connectors and the board's two 40-pin male headers give you all of an MCU's signals.

Even without further instructions I quickly set up an 8-bit binary counter in an endless loop and introduced a short delay so I could see the count on an E-Blocks 8-LED board. Programming can't get any simpler. At first I tried to use the statement MY_OUTPUT++ to increment the count, but Flowcode didn't recognize that as a valid math operation. Remember, you don't use the C-language syntax at the flow-chart level. The statement MY_OUTPUT = MY_OUTPUT + 1 did the trick. The Flowcode building blocks let people write C-language code if they wish, but they put that code in special blocks created just for that purpose. Flowcode comes with many blocks ready to control MCU UARTs, external LCDs, and other add-on devices and circuits.

The flow-chart approach to programming makes sense. Often I mistakenly type a semicolon after a C-language if-else statement and it causes debugging problems, even though I know where to look for my typing mistakes. With Flowcode, I move a conditional-branch "diamond" into the program flow and think about the condition to test, not my typing skills. You can still make mistakes in program flow, use of variables, and an overall algorithm, but Flowcode takes care of the syntax and simplifies moving blocks. That's a time saver.

Flowcode lets you examine the C and assembly-language code it produces. The C code has comments, but they do not carry over to the assembly-language listings. You can add comments to Flowcode diagrams, though.

While looking through the C-language counter-example code I noticed the MY_OUTPUT variable has changed to FCV_MY_CODE, which I assumed meant a Flowcode variable (FCV). I tried to create a C-language block but couldn't get it to work, probably because I didn't properly define my variables via a Variables window I discovered after working with Flowcode for a while. It might take me a bit more time to master declaring and using variables in the Flowcode programming environment.

Next I imported a sample project that used the PIC UART in a loop-back test. The resulting C code seems to cover every UART configuration as well as several compiler types and options. Thankfully all the pages of UART C code turn into only about 20 assembly-language statements. I had no idea what bit rate the UART used until I measured it (9,600 bits/sec), but the chief technical officer at Flowcode got me squared away. When you choose a function, such as a UART, flowcode graphically represents that function at the bottom of the screen. Chick on that graphic and you get to a window that lets you set parameters such as a UART's bit rate. That's a nice way to change attributes without messing up Flowcode's behind-the-scenes C functions.

While working with the UART on the PIC16F88 chip I found several discrepancies between the MCU pin-function diagram in the PIC Multiprogrammer data sheet and those shown in the Microchip data book. Rely on the Microchip information. Also, the information that came with the UART example did not explain what people should see when running the UART program, so I put together another program and finally understood how to program UART functions. The UART block will transmit or receive a character or a string. But combining those actions in one block increases its C-language complexity.

I downloaded the information in "An Introduction to Microcontroller Programming," offered by Matrix Multimedia in the online-training section of its website. Unfortunately, the free version of this information does not include the lab exercises.

After experimenting with the UART I removed the PIC16F88 chip and substituted a 28-pin PIC16F876A I had on hand. The PIC Multiprogrammer board supplies high-quality sockets, but even so, inserting this large chip took a lot of force. Consider using a zero-insertion-force (ZIF) socket if you plan to program more than one or two MCUs in one session. Although the board provides many sockets, you can only program one MCU at a time.

The PIC16F876A MCU provides more I/O pins than the 16F88, so I could connect an E-Blocks LCD module and drive it from Port B. The display, pushbutton-switch, and other modules require a power connection to the Multiprogrammer board, as shown in the photo. The LCD-control block worked well and I used it to display information from a simple counter program.

When I first tried to program the 16F876A I got a warning message, "Hex file has config2 word, but selected chip does not."  I clicked Yes to continue and later discovered I hadn't reconfigured Flowcode - a simple operation - for the new MCU type. The Flowcode software and E-Blocks hardware provide an excellent way to start programming and create new devices without the headaches of debugging C- or assembly-language code. Engineers can use the Matrix Multimedia multiprogrammer boards or create their own boards and still use the Flowcode 4 software to write programs.