FrankT wrote: > Anybody know the differences [between the ICEPIC and PIC-Master > emulators] besides the huuuuge price difference(x8) ? Does anybody > worked with them both? Comments please. Frank: The following is a copy of a message I posted here about 15 months ago. Two things to note: 1. I incorrectly identified the manufacturer of teh ICEPIC as "RF Technologies". They're actually called "RF Solutions". 2. In the time since I wrote the review, Microchip has expressed their intent to support the ICEPIC emulator in their MPLAB environment. As yet, though, there's no firm date for this. Ok... Here's the message: ------- Forwarded Message Follows ------ To: PICLIST@MITVMA.MIT.EDU Subject: ICEPIC Emulator Evaluation Date sent: Thu, 1 Feb 1996 02:04:20 -0800 Dudes: Some time ago, in response to a question posted here, I mentioned that I was evaluating RF Technologies' "ICEPIC" emulator and would post a brief report when I was done. Before I do, I should mention that my company has no involvement, financial or otherwise, with RF Technologies. I also have no personal relationship with the ICEPIC's creator, Amar Palacherla or with his partner. Amar designed Microchip's PIC-Master emulator, which I use every day and believe to be the world's best emulator for ANY microcontroller, but on the other hand, I once had to spend hours debugging some 16C71 code he wrote for a client of mine. Therefore, my personal feelings for him are absolutely neutral, and in no way influenced my report. Ok... The ICEPIC is based on the same bondout chips used by Microchip in their PIC-Master, and its software has a look and feel similar to that of Microchip's MPLAB software. Unlike the PIC-Master, ICEPIC has no hardware-trace capability. It also has no external breakpoint input, nor does it have a "trigger" output or the 8 generic inputs PIC-Master provides. It also uses a serial interface as opposed to the PIC-Master's plug-in ISA card. The serial interface is both good and bad. It's great for laptop users, anyone who's run out of slots in his PC, or people who need to share the emulator among multiple computers. However, serial interfaces are pretty slow; even at 57,600 baud, step-by-step "animation" of one's program (with register-refresh on) is pretty slow. The ICEPIC comes in a rather cheap-looking box (which is ok, I guess, since it costs only $700 versus $2300 for the PIC-Master). The "personality modules" that configure the emulator for a particular PIC fit inside the case, which is fastened with two screws. It's a little tedious to switch personality modules (or even change clock speeds; the oscillator's inside the box, too), but enclosing the whole emulator DOES make it less likely that you'll accidentally short the thing out (a real possibility with the PIC-Master's fully-exposed probes). PIC-Master features that are missing from the ICEPIC (these aren't bugs; they're inherent in the ICEPIC's design): No external clock capability. Incomplete emulation of Power-On Reset. No Power-Up Timer, no Oscillator Startup Timer, and the watchdog timer doesn't work in single-step mode. Weak source-level debugging. ICEPIC displays source code in an integrated editor very similar to the WIndows Notepad. While the emulator talks to the editor just fine (updating the cursor position to match the Program Counter, etc.), there's no communication BACK from the source-code window; you can't set breakpoints, change the PC, etc., without leaving the editor window and selecting the Program Memory window. The most important PIC-Master feature that COULD be added to ICEPIC, but isn't: Robust breakpoint capability. Currently, breakpoints are allowed only on "Address Match". Even in "Animate" (continuous, automatic single-step) mode, there's no way to break on Register Contents or Boolean Expressions. Breakpoints also can't be saved between sessions, they can't be set by specifying ranges, they can't be individually turned on and off, etc. All in all, ICEPIC handles breakpoints pretty poorly. Aside from the above-mentioned design issues, there are a huge number of bugs in the current version (1.52) of the ICEPIC software. Most of these are of the "garbage in, garbage out" variety; there's EXTREMELY little error-checking of user input. For instance, ICEPIC (which expects decimal values to be entered normally and hexadecimal numbers to be entered with the "0x" prefix) accepts as legal numeric input "0xA", "0A", "0QWERTY", and "0x!@#$%^&*()". Only the first is interpreted properly as decimal 10; the next two are interpreted as 0 and the last was variously interpreted as decimal 88, 17, etc. No error messages are displayed when entering meaningless data like this, nor are errors generated when you try to directly modify the Program Memory with impossible instructions like (when emulating the 16C5x) "MOVF TRISA,W". There's also a big problem with the user interface. Most of the windows are WAY too big; as I wrote in the report to my client, the ICEPIC screen looks as though it were designed by Fisher-Price. Additionally, there are numerous interface quirks that quickly become annoying. If any of you have used the PIC-Master for more than a year or two, you'll remember that when it was introduced, its character was similar to ICEPIC's: Great hardware with only mediocre software. True, ICEPIC's "mediocre" is a step or two below the original PIC-Master's, but RF Technologies doesn't have Microchip's resources, so I guess that's to be expected. I believe that the software issues I mentioned are likely to be corrected soon. Once they are, ICEPIC will be a pretty cool product. Personally, I'll keep my PIC-Masters -- I use the hardware trace and the external probes a LOT -- but I'd have no problem recommending the ICEPIC to anyone looking for a reasonably-priced emulator that supports the entire PIC16/17 range of products. Even with its current software, the ICEPIC isn't ALL that bad... The software is sometimes irritating, but the hardware is solid and (aside from the few design issues mentioned above) emulation is completely accurate. ------- End of Forwarded Message ------ -Andy === Andrew Warren - fastfwd@ix.netcom.com === Fast Forward Engineering, Vista, California === http://www.geocities.com/SiliconValley/2499