Post by thread:[PIC] PIC24 to support QVGA and USB OTG

Good news. But it does not follow ARM arch. Therefore ppl. from ARM dev may shy away from the learning curve. John --- Xiaofan Chen <spam_OUTxiaofancTakeThisOuTgmail.com> wrote: > 21icbbs.com/club/bbs/bbsview.asp?essenceid=7575 {Quote hidden}> (In Chinese). > > The poster (Martin) is from Microchip China. > > Simple translation: > > "Microchip will focus on PIC24 16-bit MCU in the > future. Microchip will > come out 20+ models of PIC24 (including PIC24F and > PIC24H) with > Flash up to 256KB and RAM up to 16KB and many > peripherals. > > In the future, PIC24 will integrate QVGA display > module and USB OTG > (mini host). More importantly, the price will be > very competitive as > ARM7 based MCU". > > --

On 3/27/06, Gökhan SEVER <.....gstr2005KILLspam@spam@gmail.com> wrote: > Why does the Microchip team insist to use ARM cores' in their MCU's? > > Do you think they can attract current ARM users with their architecture? > Microchip team will not use ARM core. I think it makes business sense for them. Even Atmel is now developing AVR32 to compete with ARM9. Companies need to pay ARM license fees to develop ARM compatible MCU, and I guess the licensing fee will be much higher than to use 8051. ;-) Let's wait and see if Microchip will be sucessful in the 16bit market or not. Take note that ARM7 based architecture is not at all the market leader now in the 16bit MCU market. Renesas/Infineon/Freescale/... all have bigger market share. Regards, Xiaofan

On Mar 26, 2006, at 6:26 PM, Xiaofan Chen wrote: > Let's wait and see if Microchip will be sucessful in the > 16bit market or not. I haven't quite managed to understand how the ARM chips manage to be so cheap. I mean, forget the CPU; microchip currently charges much more for the memory array(s) than ARM vendors are. I really have a hard time imagining the ARM taking over much of the "deep embedded" space (where the finished device doesn't at all resemble a computer.) The chips are too ... inconvenient. Yeah, the ARM might take over from bigger 16bit chips; things like MP3 plays and so on where the CPU power is worth the extra infrastructure... OTOH, I don't know whether one needs a PIC24 or AVR32 there either. I can't see 8-bit controllers disappearing... BillW

> I really have a hard time imagining the ARM taking over much of > the "deep embedded" space (where the finished device doesn't at > all resemble a computer.) The chips are too ... inconvenient. what is inconvenient? packaging, voltage levels, architecture? > Yeah, the ARM might take over from bigger 16bit chips; things > like MP3 plays and so on where the CPU power is worth the extra > infrastructure... > OTOH, I don't know whether one needs a PIC24 or AVR32 there either. > I can't see 8-bit controllers disappearing... It's not one or the other, there is a broad spectrum of (deeply) embedded applications. When chip price (and maybe long-term availability!) is important IMHO PIC-12 and PIC-14 are here to stay. I am not so sure about pic-18, pic-24, and mega-AVR, who are in the same price region as low-end ARMs and can't deliver the same horsepower. Think of a low-end ARM as an SX at 60 MHz but with 32-bit instructions, more RAM, no banking, lots of peripherals, lots of development tools, JTAG debugging interface, build-in bootloader, etc etc. Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu

On 3/27/06, Wouter van Ooijen <.....wouterKILLspam.....voti.nl> wrote: > > I really have a hard time imagining the ARM taking over much of > > the "deep embedded" space (where the finished device doesn't at > > all resemble a computer.) The chips are too ... inconvenient. > > what is inconvenient? packaging, voltage levels, architecture? I've just bought an Olime LPC-P2148 board. I will say the support from chip vendors are not so good. The packaging is not so friendly for hobbyists. The Philips flashing tool is of beta quality. Philips does not provide the compiler or even assembler. Instead, people have to use the free GCC or the limited Keil/IAR tools. There are also so many instructions for the ARM compared to PIC even it is RISC... There are no cheap but "official" development tools like MPLAB ICD2. Instead people have to use expensive tools or cheap tools like wiggler which is not officially supported. I really doubt the hype around ARM taking over the high-end 8-bit and 16bit market... Regards, Xiaofan

> > what is inconvenient? packaging, voltage levels, architecture? > > The packaging is not so friendly for hobbyists. I agree > The Philips flashing tool is of beta quality. is it? I've had no problems with it > Philips does not provide the compiler or even assembler. But you can get a GCC on every corner of the street (oh well, every corener of the internet), for free. C and Assembler (and real C++!, and if you spend some time on it: Java, Ada, Fortran, etc). > Instead, people have to use the free GCC or > the limited Keil/IAR tools. So compared to PIC18: - (free + supported + limited) tools are available - (free + sunsupported + unlimited) (==gcc) available for ARM but not for pic-16 > There are also so many instructions for the ARM > compared to PIC even it is RISC... Learning the instruction set is indeed a bit easier for pic14 that for ARM. But the added paging and baking puts the score even, and realy using pic14 is definitely more difficult than ARM. I dunno about pic18 (I have not done any pic18 asm classes), I think the number of instructions are about the same with the ARM having more variations. Which makes it a bit harder to learn but easier to use. > There are no cheap but "official" > development tools like MPLAB ICD2. AFAIK McGregor and a lot of others make debugging tools and support them. The ARM world is multi-vendor for both chips and tooling, so you won't find a single official toolset from for instance Philips. I am not sure that is a disadvantage. > Instead people have to use expensive > tools or cheap tools like wiggler which is not officially supported. Are the expensive tools so much more expensive than an ICD2? So far my classes have done fine with our wiggler-clones (we have had the occasional hiccup, but so have ICD2 users). And a wiggler (a JTAG debugger interface, capabilities compareable to an ICD2 but parallel port based) is roughly as complicated as a PIC parallel port LVP programmer. > I really doubt the hype around ARM taking over the high-end 8-bit and > 16bit market... we will see. my one main problem is the lack of any DIP chips. I guess I could make a fortune selling adapter boards :) Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu

On Mar 27, 2006, at 4:47 AM, Xiaofan Chen wrote: > I will say the support from chip vendors are not so good. It seems (without having used any) that it's comparable to many. PIC and AVR seem to be exceptionally well provided for compared to most microcontrollers. > The packaging is not so friendly for hobbyists That's for sure. Not only are there no DIP packages, but things go straight to beginner-unsolderable fine-pitch stuff. Not even PLCC or SOICs... Homemade PCBs are out. Assembly by software engineers (or university EEs) (ie people without practice) is out. And I find the multiple-supplies required for most chips to be very daunting as well. It's depressing how fast the costs accumulate for such 'cheap' processors. Professional PCB. Multiple regulators... The "mythical" $4 60Mips processor suddenly becomes a $60 module... > The Philips flashing tool is of beta quality. I don't usually expect tools from semiconductor manufacturers to be of the highest quality. I wonder how the Motorola/Metroworks acquisition has worked out from most people's view? I remember a lot of uncertainty around when that first happened. Not to mention complaining from the 3rd party compiler/tool vendors. One wants the manufacturer to provide usable tools, but they should be so overwhelmingly wonderful as to force other tool vendors out of business! Shared architectures like ARM, 8051, MIPs, etc, must have additional business and philosophical issues. I know how careful Intel is to keep their own compiler SEPARATE from their support of open source tools, for example. A real pain... > Philips does not provide the compiler or even assembler. > people have to use the free GCC or the limited Keil/IAR tools. > There are also so many instructions for the ARM compared to PIC > even it is RISC... Huh. Doesn't this make it more like the 8051 that you like? An architecture so prolific that tools come from ... everywhere? And the ARM is a true RISC; you're not supposed to program it in assembler; you're supposed to program in a HLL like C. Hmm. I've always thought of C programmability as being a big advantage; I only feel vaguely uneasy when the assembly language becomes nearly unusable. But perhaps... Is there a segment of the hardware-oriented community that actually feels like assembler is easier to program in than C? That's a bit frightening, but I can see how it could be true... And isn't GCC an advantage rather than a disadvantage? Oh. I guess it IS pretty primitive compared to most IDEs. People just aren't used to CLI-based toolsets any more. You practically need to be a unix sysadmin to do anything. Maybe... (Surely there's at least one project to supply an microcontroller-oriented IDE for gcc/etc. Other than emacs. (Hmm. I worked with someone recently who couldn't use EMACS without the mouse button pseudo-GUI turned on. A weird feeling.)) BillW

On Mon, 27 Mar 2006, Xiaofan Chen wrote: > On 3/27/06, Wouter van Ooijen <EraseMEwouterspam_OUTTakeThisOuTvoti.nl> wrote: >>> Take note that ARM7 based architecture is not at all the >>> market leader now >>> in the 16bit MCU market. Renesas/Infineon/Freescale/... all >>> have bigger market share. >> >> Which is not surprising because ARM is a 32bit architecture.... > > I know that but the low end ARM7TDMI based chips are targeting the > 16bit market and even higher end 8-bit market. And I think they are > not that popular in the business world yet compared to the incumbent. No, the 16 and 8 bitters are trying to encroach on 32 bit territory because of the need for speed and networking. They will loose. The need for speed and networking is greater than the need for 8 bitters with paging and banking shenanigans. In 5 years every respectable digital watch and tv remote control will be networked, probably wirelessly. Peter

On Mar 27, 2006, at 4:14 PM, Peter wrote: > No, the 16 and 8 bitters are trying to encroach on 32 bit territory > because of the need for speed and networking. They will loose. The > need for speed and networking is greater than the need for 8 > bitters with paging and banking shenanigans. I think the "need for speed" is over-rated. But you have a good point about networking. Manufacturers seem to get a bit depressed when the peripherals start occupying more chip real-estate than the CPU and memory :-) It'll be interesting to see how the current crop of ethernet-enabled 8bit chips do in the market (along with things like microchip's device that enables ethernet on 8bit chips...) OTOH, the low-end ARM chips we're talking about don't specifically include networking support beyond having enough flash/ram to make a network stack "easy" compared to most 8bit chips... BillW

> (about 8 / 32 bitters) > I think the "need for speed" is over-rated. For some applications this is sheer nonsense (try doing a decent MPEG decoder on an 8-bit PIC). Although you might need more than just speed (for instance a lot of easily addressed RAM). But there are of course some other applications for which this is perfectly true. I don't think the low-end 8-bit world will be going away. I don't see Philips introducing a 6-pin ARMs for $0.50 real soon now. I think there would be a big market for 4-bit chips if they were available for $0.10 each... Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu

So What {Quote hidden}> > On Mon, 27 Mar 2006, Xiaofan Chen wrote: > >> On 3/27/06, Wouter van Ooijen <wouterspam_OUTvoti.nl> wrote: >>>> Take note that ARM7 based architecture is not at all the >>>> market leader now >>>> in the 16bit MCU market. Renesas/Infineon/Freescale/... all >>>> have bigger market share. >>> >>> Which is not surprising because ARM is a 32bit architecture.... >> >> I know that but the low end ARM7TDMI based chips are targeting the >> 16bit market and even higher end 8-bit market. And I think they are >> not that popular in the business world yet compared to the incumbent. > > No, the 16 and 8 bitters are trying to encroach on 32 bit territory > because of the need for speed and networking. They will loose. The need > for speed and networking is greater than the need for 8 bitters with > paging and banking shenanigans. In 5 years every respectable digital > watch and tv remote control will be networked, probably wirelessly. > > Peter > --