Microchip's PIC32 : comments needed

Hello,

I am interested in many ways by MCHP's new family, as it seems to adress many of my requirements. However, this is a brand new thing (understand : "preliminary") and although MIPS has been around for ages, the PIC32 is not yet in full production. Errata and the likes will certainly come in wagons, this chips is more complex than what i have seen from this MCHP.

So i turn to this group and ask you all, in all honesty :

  • Have you read the PIC32's documentation, and/or used it (either the plug-in module or the starter kit), and what do you think ?
  • How do you compare this first iteration to other CPU/MCU you already know well ?
  • What are the strengths and weaknesses ?
  • What do you think about the development tools ? (for example, i have no XP or Vista box, and buying a 0 computer just for using a kit is a bit... tough)

I have a lot of ideas where the PIC32 fits neatly in embedded projects, I hope to get rid of my barebones and PC104 boards... but this is not going to happen in one day. Particularly, prices and costs are displaced where i did not expect them, and where they annoy me most :-(

Happy new year all, YG

Reply to
whygee
Loading thread data ...

I've got the starter kit, it works OK and is a good tool for evaluating the PIC32. It has an expansion connector that could be used for interfacing to external hardware, and one or two people are designing boards for it.

It isn't as easy to use as, say, the NXP LPC2000 ARM chips, but it has some advantages over it.

Leon

Reply to
Leon

$800? I am using an old HP Pentium III (900MHz Pentium III) laptop I bought off eBay for $70. It came with a licensed copy of Windows 2000, which is adequate for the task.

I do not as yet have a compelling application for this particular chip, I have experimented only for the sake of learning a bit about it. I'll consider it for new designs, but likely will continue to use ARM.

Reply to
larwe

Well, I am currently using a P3 laptop (500 MHz) here, under Debian, and the same, but 700MHz version, under w2k, for some electronics stuffs. The sad fact : MCHP's MPLAB9 does not work/install anymore on w2k (i use the earlier MPLAB7.60 now). And guess what... MPLAB8 is the first version that supports PIC32. I'm screwed and angry. Same with the FPGA proto board that i am currently ordering, and this one is going to suck more RAM than i can afford (after buying the Actel kit, which recommends 2BG, while my best laptop has 256MMB...).

Well... I also just got a PicKit2 and this sucker's driver installer wanted w2k SP4 (i'm somewhere in SP3). fortunately, this idiot just looks at the.... InternetExplorer version ! I fortunately figured that editing a stupid key in the registry allowed it to install (from version 5.00.0xyz to 5.01, and it works ok). I guess that it's not going to work with MPLAB8.

Oh, and i'm fed up with having to "patch" a proprietary system every time other proprietary (free as in free beer) SW have to be installed. And my computers are getting old. I have almost nothing < y2k :-/

What did you try with it ? What succeeded and what failed ? have you found something particularly good or bad ?

I have looked at ARMs in the last years but have never found something that suits me, or easy-to-obatin chips. I can play with someone else's ARM9 proto board, but the environment (Debian Linux) makes it useless for my application (hard real time stuffs, where 99% of the CPU is needed).

With the new PIC32 family, i have the hope that it starts anew, cleanly and that i'll be able to keep up in the future... ARMs are already so developped, and have so many versions, that i can't follow. Finally, being an old MIPS enthusiast, i can't stay calm :-)

YG

Reply to
whygee

Microsofts dominance has never been just about them making money. They would have never been allowed to be where they are now if that were so. They are all about the public having access only to controlled "computers" (some might prefer the phrase "computerised TV sets").

If someone wants to sell me a chip which comes only with a wintel system attached to it to be usable, well, he fails. There still are other viable options.

Microchip going MIPS will make it a lot more difficult to those "other viable options", of course - this core choice puts Microchip out of the "automatically discarded" MCU manufacturer group. It will depend on how many people will be willing to give up some of their toolchains to surrender to what they are allowed to have and use under Vista, hopefully I will not be the only one not needing that to do development work...

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

formatting link

------------------------------------------------------

formatting link

he

supports PIC32.

rd

d.

t suits

rd,

llow.

Reply to
Didi

This is a good example why Linux is not a good development tool.

Those who preach the benifits of Linux, fail to understand that the WinTel enviroment is just easier to use and program in.

I install a application in WinTel and it is ready to use. Icon on the desktop, examples ready to run. Without ever seeing it before.

Linux requires you to have a tutor on hand to tell you where the program got loaded to. Someone to tell you where the example file are located. Someone to tell you where the linker files are located.

There is not one editor to use. So someone need to tell you where an editor is located, or which one to load and you start at the top all over again.

I for one is sick of the M$ tax, but Linux will not ( does not want to as far as I can tell) be as simple to use.

donald

Reply to
donald

By definition, anyone who is doing embedded development is at least a bit ahead of the "I need an icon to double-click!" mindset. If not, perhaps you're in the wrong profession.

Reply to
larwe

I am not the only one with this need, or Linux would have more acceptance.

But, its does not, enough said.

donald

Reply to
donald

Maybe, but only a fool makes more work for themselves than need be done=20 to complete a task. If you aren't smart enough to surround yourself=20 with the tools and knowledge necessary to efficiently do your job,=20 perhaps you are in the wrong profession.

Reply to
James Beck

Donald made a valid point:

The goal of a project is accomplishing something that works. The learning curve only adds to the overhead.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

formatting link

Reply to
Vladimir Vassilevsky

he

Oh wow, I didn't know that. I use V8.

rd

2GB is a bare minimum for doing big FPGA development, and a fast multicore processor is advisable. A full build on a big project (I was working with Virtex4) can take a day on a slower computer.

SP4 is a rollup update, it can be downloaded from MS.

Nothing failed. Nothing is particularly bad. The problem is that nothing is particularly good either; cost, MIPS/mW, peripherals. There's absolutely nothing in the range that would justify me moving my code from existing ARM platforms. Same problem I have with AVR32; I read the product announcement and my response is "so what?".

For a brand new design I would consider PIC32 alongside the other alternatives, but frankly uChip's chance of a design win is low unless they compete very favorably on price. And the problem with that is that most of the other ARM chips we already buy in huge quantities, so uChip would have to match the 10^6 price on an order that's maybe 10^4 pieces.

Reply to
larwe

Well, be it known that I do not in fact do most of my dev work in Linux (any more). This is principally because I haven't developed for embedded Linux much in the past 6~12 months. When Linux of any flavor is the target OS, I find it definitely the path of least resistance to use Linux as the host OS.

But even within Linux, there are many simple-to-use tools. I use EAGLE for my PCB/schematic work, and this is 100% identical Windows/Linux/ MacOS. Rowley's IDE is likewise identical Windows/Linux. Xilinx ISE etc...

Simply the lack of a Windows logo in the taskbar is not sufficient reason to declare an operating system impossibly difficult to use. When I'm working with parts that have good Linux tool support, I prefer to use Linux.

Reply to
larwe

Would it be that dificult to create an environment under Linux that acts like some of the WinTel development environments ??

Again, why don't the Linux developers just try it ?

Every few months I try to get a development system running with Linux.

This holiday I tried again.

Could not do it by myself.

Please help me and the others who want to leave M$ behind.

I would even pay for it. A Linux environment has got to be cheaper that then the $1000- $2000 IDE environments out there.

Start with a FREE compiler and add the install tools that will make it a no brainer.

I just don't understand why it has not already happened. ( unless the linux people just don't want too )

donald

Reply to
donald

The only person with financial incentive to do it is the chip vendor.

It's very difficult for enthusiasts to develop such software, even if they have the desire, because of proprietary secret debugging protocols not disclosed by the mfr (and covered by NDA so not implementable in open-source software).

Reply to
larwe

You are still doing it the hard way. Instead of downloading anything, grab the remote control of an oldfashioned TV, sit down and enjoy.

No need to use or understand anything beyond the few keys on the remote control.

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

formatting link

------------------------------------------------------

formatting link

Reply to
Didi

Well that sounds like a cop out.

Ok, yes there needs to be a hardware debugger. Yes, the chip vendors do keep it secret.

But to install an IDE and compiler is harder than it looks.

So, do I hear you saying the Linux people are out of line complaining about WinTel. :-)

donald

Reply to
donald

Linux is very handy once you got used to it and configured everything for yourself. Until then it is pain and suffer. M$ is vice versa: the deeper you are getting, the more annoying and restrictive it is.

It is getting there. The Ubuntu is almost as user friendly as Windows 95 used to be. Wait for another 10 years :)

And why exactly do you want to leave M$?

There are no wonders. Everyone gets what he paid for.

Quite many of the commercial packages are based on the GCC core with some modifications, wrapped into the good looking IDE.

It has happened. You just don't see the origin under the upper layers.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

formatting link

Reply to
Vladimir Vassilevsky

I don't disagree with you on any particular point. I was just playing devil's advocate. We use a couple of flavors of Linux here and I like the fact that it can=20 run VERY well on limited resources. That being said, simply the addition of a Windows logo doesn't make an=20 OS any more easier to use either. Most of what people consider the=20 "OS" these days isn't. I think the "Desktop" look and feel is now=20 mistakenly referred to as the OS. At least Linux still differentiates=20 its' OS from the GUI in its' general terminology.

Jim

Reply to
James Beck

I don't really know what you're talking about here. Installing, say, Eclipse, is a one-step operation. Installing a compiler depends on whether you have to build it (which I usually do) or if a prebuilt version is available.

But it's all useless unless there is a way to get the object code into the target, which is the really difficult step.

Reply to
larwe

... snip ...

Why in heavens name do people want an IDE? One window in which to run the editor, another in which to run make and the test runs, add others as needed. All free, all controllable.

--
 Chuck F (cbfalconer at maineline dot net)
   
   Try the download section.
Reply to
CBFalconer

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.