Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ucbvax!utrcgw.utc.COM!RAYBRO%UTRC From: RAYBRO%UTRC@utrcgw.utc.COM ("William R Brohinsky", ay) Newsgroups: comp.lang.forth Subject: RE: Real Time Forth Message-ID: <8907140336.AA29694@lilac.berkeley.edu> Date: 13 Jul 89 13:22:00 GMT Sender: daemon@ucbvax.BERKELEY.EDU Reply-To: Forth Interest Group International List Organization: The Internet Lines: 80 In his message posted 13jul89 1:11, John Saare writes: > I'm planning the construction of a SMALL engine monitoring >and navigation augmentation system for a homebuilt airplane... > Recently, I've decided to start looking at the project again >with the idea of using as many off the shelf components as possible... > I've been considering using Forth as the development environment >(cheap, easy prototyping, COMPLETE)... >Question: What Forth would you recommend that is possessed of the > following qualities: > - Easy to use in a "rough" hardware environment... > - Not dependent on a supporting O/S... > - ROMable > - Reliable > - Runs on '86 family... > > Any suggestions? I have used the NMIX series of boards from New Micros, Inc. of Dallas Texas with very good results. They are well built, sturdy, complete, and use forth-on-a-chip software. Once code is written and debugged, it can be ROMed, or downloaded each time (which makes the continuous updating of software a lot easier). Whether ROMed or downloaded, the system works identically for testing purposes. These newest boards use the 68HC11 chip. This is NOT an '86 family architecture, I know, but for control applications the '86 family is like swatting flies with a caber- it may be effective if you can aim it, but not easy to use in any hardware environment. (I admit my predjudice against segmented architectures and seperate memory and peripheral busses.) In favor of the 68HC11 chip are 8 built-in a/d's, 5 built-in timers, 2 serial channels (one standard comm type, one "micro-wire" type for inter-processor communications), the ability (with the 68HC24 chip added) to use all of its parallel lines (40 of them, in 5 8-bit ports) if you aren't using the other peripherals (a/d lines take one par line each, serial comm uses 2 (minimally) micro-wire uses 3, timers only take one line each if you use external I/O on them, but they can do internal timing without using up a line). This is a lot in favor of the 'HC11, but MAX-FORTH allows using all of it, pluse the on-board EEPROM (512 to 2K bytes, depending on which 68HC11 version you use). The NMIX-0022 (which I have used) is the most fully-packed of the series, and costs around $350 with MAX-FORTH development software. Other boards are available, and are essentially designed to be target (finished apps) boards: some lack the 'HC24, some come without either HC chip, etc. To develope an application, you use the serial channel from another computer (which acts like a dumb terminal, but with memory and mass storage). Write the code on the host computer in sequential/text/non-block files, and download the code thru the serial channel to the board. The code is ROMable, and is beheaded in the process, which makes relatively secure code. I used the board for a computer-controlled Laser Speckle Photogrammetry system, driving a photographic-glass-plate transport and a 10-J pulsed-ruby laser (nasty, powerful thing) and getting input from an optical one-per-rev device of our design. I charged the laser by hand (there being no other way to do it) and told the NMIX board when to start a sequence by typing on a TRS-80/Model 102. The board then calculated how long before the next one-per-rev pulse to fire the flashlamps so that the Q-switch could be fired from the one-per-rev and we'd get full power. It also gated the one-per-rev pulses so that the laser would not fire prematurely, and operated the plate transport to prepare for the next shot. Before moving the plate, however, the NMIX board would calculate the actual RPM and print this and the run and plate number on the glass photographic plate with a set of HP character leds. I recieved a design award here for the system. Your needs could be met easily with a laptop-pc with one serial port, a comm program and a navigation program, and an NMIX-0022 board with sensors in appropriate places. This would give you an optimum nav interface, while offloading all control functions to the NMIX. The PC would give you the ability to query the NMIX if necessary. New Micros address is New Micros Inc. 1601 Chalk Hill Rd. Dallas, Texas 75212 (214) 339-2204 -raybro Disclaimer: I don't work for or get money from New Micros, all opinions here- in expressed are mine, although others may share them freely, and UTRC and UTC do not endorse anything I say here. -raybro