Bailey O.H.Embedded systems.Desktop integration.2005

DIP switch, or jumpers to select the active interface. If we take this approach we will need a total of four pins for the data and handshake lines and two additional pins for a switch for a total of six pins. This approach does save two processor pins. We could also make the interface software switch selectable through the menu, reducing the total pins needed to four. This is a very good solution but there are two concerns. First, if a default interface isn’t chosen, the board may not communicate at all. Second, the BASIC Stamp interfaces to the board in a physically different way and at a different level then a PC. While an inexperienced user might place a call to customer service, the manual with the board clearly shows how to interface a TTL serial device such as a BASIC Stamp. Those who are not as comfortable with technology will call for support. In the final product, marketing has decided that a switch or automatic method of detection is preferred over a software menu item.

There are benefits and drawbacks to both approaches. If we have enough processor pins available there is no reason to add additional switch circuitry other than for future expansion, which isn’t needed right now. This also provides us the flexibility of using multiple interfaces concurrently for debugging purposes. On the other hand, if we don’t have enough processor pins already, then changing interfaces may be more efficient by using a switch instead of software. If we are switch selectable, we limit our available software expansion capabilities and make adding future interfaces a hardware task. On the other hand, if switch settings are left in a state unknown to the embedded system, we run the risk of disabling all interfaces. Remember, if we have to change hardware, we have to make circuit board changes and inevitably our costs will increase for each change, not to mention that field upgrades will become a requirement.

We have a third alternative that combines both of the above approaches. Since we are using two handshake lines for the RS-232 and USB interfaces, both can use the same four BASIC Stamp I/O pins. This would allow us to have a single hardware switch to select RS-232 or USB and would reduce the number of processor pins required by four pins. Since RS-232 and USB are

Chapter 8

from free to thousands of dollars. There is really something for everyone.

Embedded systems developers have as many choices for Ethernet interfaces as PC users.

Ethernet interfaces operate at 10 to 100 MPS (megabits per second). Assuming a 10-bit frame per character (including overhead), that’s a full 1 to 10 million characters per second. Hook that up to the Internet over a high-speed connection and you are moving at a very fast pace that allows huge blocks of data to be moved very quickly. In fact, combining high-speed Ethernet cards with a broadband connection comes much closer to the speed of a dedicated T1 line than dial-up or DSL provide, and usually for a much lower cost than a T1.

Taking your Ethernet interface from being LAN to WAN accessible is a simple matter of providing a cable or hardware link to the WAN and possibly changing the IP address. To go from WAN to Internet access means adding accessibility to the outside world and possibly making some Ethernet address and port changes. The point is that the basic hardware connection doesn’t change. If you’re connected to a single system and want to connect to the LAN, it’s a simple matter of changing cables and possibly making some IP address changes. Here are the benefits that Ethernet provides for the host:

Cost — Ethernet chips and cards are inexpensive.

Wide platform support — All major host operating systems have built-in support for the protocol.

Expandability — Widening the availability of your device is fast, inexpensive, and easy once the initial development has been completed.

Speed — Ethernet is fast when compared to RS-232.

Provides a many-to-many connection — Ethernet can broadcast messages to a single user or a group at the same time.

Chapter 8

The first is to write, purchase, or obtain a TCP/IP software toolkit, making certain it supports the Ethernet chip or microcontroller you have chosen. Most of the products offered in this category either target specific processors and charge a per-unit license fee or come with generic TCP/IP protocol handlers and drivers that require the developer to customize it to the chosen platform. Again, costs can vary widely using this approach.

The second implementation method is to purchase a packaged Ethernet system that contains the necessary hardware and software. There are several excellent products that provide a standalone Ethernet module designed to be used with embedded systems. Usually these products include an Ethernet controller and processor along with software to customize the module and add application-specific functionality.

Note:

All devices used as controllers in these prepackaged Ethernet kits are intelligent. That is to say they have I/O capability and are programmable just as the BASIC Stamp. How much of this functionality is accessible to the developer is up to the manufacturers of these kits.

The third Ethernet implementation method is to purchase a processor that already has TCP/IP embedded in the silicon or has software support from the manufacturer. More and more processors are offering built-in Ethernet capabilities as you enter the 16and 32-bit microcontroller markets.

Chapter 8

Serial cable

SB72 to Ethernet hub cable

SB72 to computer cable

12-volt power supply

The software on the NetBurner CD contains of the following components:

TCP/IP stack

Web server

Real-time operating system based on uC/OS

GNU C/C++ compiler and linker

GDB debugger (the GNU debugger)

IPSetup configuration utility

Auto-update Flash update utility

User manuals and documentation

Figure 8-13 shows a picture of the full kit.

Chapter 8