On Fri, Aug 09, 2002 at 01:46:22AM -0500, Dale Botkin wrote: > On Fri, 9 Aug 2002, Sean Alcorn (SYD) wrote: > > I just continue to wonder one thing: Given the plethora of stable, > supported, capable, commercially available, cost effective PIC > programmers, development environments, boot loaders, etc ad nauseum, does > the world really need Yet Another Pic Starter Kit? But hey, if a group of > list members want to have a crack at it, have a ball. And no, I'm not > asking you specifically, just wondering in general. If it's only a programmer, then the answer is no. A starter kit. Again the answer is no. However we've come to quite a bit of consensus in a very short amount of time. Here my summary: * Sean's factory is going to develop, assemble and package fully assembled units that will be ready to use out of the box. Country specific distributors have already been lined up. * The unit will serve both as a demo/design board and as a programmer. So it's not YAPP (Yet Another PIC Programmer). It will come with a fully hooked up array of peripherals (LCD,LEDs,7 segment, buttons/switches, pots, opamps, IR) along with a breadboard prototyping area and I/O connector so that other items can be added to the design as needed. * The unit will have a mechanism for programming other PIC parts. * Both serial and USB interfaces will be available to connect the unit to the the host machine. * A CD will be delivered with the product that will contain development software and a body of tutorial style exercises that illustrates the theory and practice of developing for the PIC and the common associated periperals that are routinely encountered in common microcontroller projects. * The primary development language will be assembly. Now in my estimation that's quite a lineup and I don't believe I've heard anything but positive responses to the above list in the last week or so. I want to point out that this unit, which I've been calling the PICLIST Designer of late, is something different that what most everyone else is offering. It has significant value. Its demo board setup means that the initial hardware setup and testing has been completely abstracted out for both the novice PIC user and IMHO more importantly for the intermediate and advanced developer. Just plug it in and get started. No programmer to build. No target board to layout and test. All the common I/O periperals right on hand with standard routines to jumpstart a project. Any project specific addons can be either quickly dropped onto the breadboard or attached via the I/O connector if it's somthing more significant. It's elegant and useful at all levels of development. It's a box that I would have wanted to buy when I was first starting out with PICs . It's a box that I'd still want to buy and use now (almost 8 years later). Now I do admit that there are still some points of contention. But compared to the big list they are relatively minor: * One of the biggest remaining points is whether we view the designer as primarily a programmer that can be used as a target or as a target demo board that also happens to program. I readily admit that I fall into the latter camp. I feel that programming is only required when one is ready to take the final project off of the Designer and onto its own board which can be wired well after the initial project has started. Because of this very secondary nature of programming IMHO, I personally feel that the Designer only needs to offer a ICSP header that the chip for the final project can be programmed from. My design rule of thumb is "Simulate what you can, Emulate what you cannot." That's one of the reasons I absolutely love the current version of gpsim. It allows you to load virtual hardware modules (LCD, LED, serial) that you can interact with during simulation without having any real hardware at all. You can program with the simulator as the target. The Designer extends this model into hardware allowing the designer to utilize all of the onboard hardware for project development. Again it's the target, not a programmer for the target. Only once the project is fully realized does an actual target chip on an actual target board need to be programmed. And if the design is correct this step will have a very small cycle interval. * The next point is related to the first. If the Unit is a programmer and not the target then there will be a chip that does the programming as well as a target chip that is programmed. Much like Wouter's WISP628 or the venerable PICSTART. It then brings in all of the issues about the connectivity of the programmer chip to the target chip: ICSP, ZIFs, how many sockets of which types, which I/O is available, unavailable, or restricted, etc. However if the Designer is the target then all of those issues disappear. Since it is the target only it needs to be programmed. Since it is the target there only needs to be one chip. Since it is the target there aren't any issues about connecting it to the target... because it is the target! The only one minor niggling point is that the target must be self programmable which leads to the next point.... * Bootloading. First the cost which Brenden has pointed out: It costs program memory on the chip. Probably somewhere between 512 to 1K instructions. However I have well documented the advantages you get with it. BTW Brenden if you do actually run out of memory and bump against the bootloader, it's possible to switch to ICSP programming of the final target, thereby getting that last sliver of program memory back. And the one final point: * USB. I think we all agree that we need it. It is kind of attached to target vs. programmer issue, because if the Designer the target then in theory the target must incorporate USB. But as I've pointed out in other posts, USB can easily be treated as a separate interface requiring a separate chip for the interface. So we frontend the USB post with a 16C765 or a FT232.. type part. But it's getting close to the nuts and bolts: which LCD? which ports are assigned to which peripherals? which opamps for the A/D and PWM interfaces? Which types of connectors for the ports? How many tie points on the breadboard? This is all stuff that is already in agreement so we can start specifying these items. It's closer than you think. BAJ -- http://www.piclist.com hint: To leave the PICList mailto:piclist-unsubscribe-request@mitvma.mit.edu