PSOC No overclocking.
The Cypress PSOC's clock is set internally, so it is not so easy to tweak.
Super X3. 3600rpm. Sheridan 6"x24" Lathe + more. THREE ways to fix things: The RIGHT way, the OTHER way, and maybe YOUR way, which is possibly a FASTER WRONG WAY!
A family that doesn't get the recognition is the Philips ARM7 21XX series. 128 bit internal acrhitecture, up to 70mhz internal clock rates, robust internal periperials, embedded boot loader, ICE and RT debugger and reasonable prices. At the low ent the LPC2101 can be bought in the $2 range.
Phil, Still too many interests, too many projects, and not enough time!!!!!!!!
Vist my websites - http://pminmo.com & http://millpcbs.com
and LPC2101 can already do a lot even at the entry level.
Price wide, the PICs did the best job.
Something like multiple quadrature encoders at that rate, I'd start taking a serious look at CPLD's.. Or even FPGA's. But talk about added complexity. YEEEehaw!Originally Posted by neilw20
48MHz becomes 24MHz. sure, and you don't need a crystal for most apps.
I know some things get well and truly divided down. Some instructions can be 13 cycles. But it sure is easy to use. Once you set something going it runs in the background, often in all in hardware with little or no interrupt suport required.
If you want a PWM (or 2), or a D-A it just works. Write a new value to a register and you are done.
Development system is tiny. Signals don't get corrupted by long wiring (unless you add heaps your self)
I gladly paid for a good C compiler. Libraries give instant results.
I am not a biased one eyed supporter. It just works for me.
Naturally, if you wanted multiple interrupt driven quadrature encoder support at 200Khz/channel you would not choose this chip. What would you use for that?
depending on your location, PICs sometimes tend to be priced high and are not the best in high volume pricing. a more powerful AVR would cost around 1/4 the price of a PIC here though PICs and Z80s are more known here for hobbyists.Price wide, the PICs did the best job.
Philips ARM7 are priced at the fraction of some of the low power PICs yet is more powerful, a lot more powerful.
How hard to get samples of those two?
Try to get some AVR and NXP sample before, never get one. Microchip is so convinent to get samples here in the USA.
Wondering how easily you can get AVR samples.
My experience in low volume AVR's cost less than PIC's taking in account features, speed, I/O. My #1 beef with AVR's is the gap between 8 pin models and 18 pin models. Would love to see more choices around 14 pin units.
Phil, Still too many interests, too many projects, and not enough time!!!!!!!!
Vist my websites - http://pminmo.com & http://millpcbs.com
Maybe they don't count as samples, but I've got a few different chips when I bought the STK500 development system a couple of years ago, something they called AVR samples pack.Try to get some AVR and NXP sample before, never get one. Microchip is so convinent to get samples here in the USA.
Wondering how easily you can get AVR samples.
I've gotted AVR samples a number of times, but you have to go through a Sales Rep and they hand deliver. So for a pure hobbiest that wants some to play with, probably not.
Phil, Still too many interests, too many projects, and not enough time!!!!!!!!
Vist my websites - http://pminmo.com & http://millpcbs.com
Microchip no longer sends samples to my country, I think for around 3 to 5 years now. But you can get samples from Sales Rep.
I think TI's the best in sending away samples, caters even to hobbyists. Never received a non-working sample from TI. I received one chip from Dallas where I can not load the chip using parallel programming methods but the bootstrap loader, as well as the chip, is fine. Another one from Atmel (though from Sales Rep) where location 000Ch always contained FFh and can never be changed.
If the Sales Rep can not give samples there's always RS and Farnell here. Entry level NXP ARM7 is 1/4 to 1/3 the price of PIC16F877 considering the ARM7 can get on par with DSPs in some applications.
When you work for a research and development group and you are as well a hobbyist, you get the chance to have better samples.
Having friends in the industry may also give you good samples.Got two on NXP.
Overall, I think PIC get to be well known because of the number of development tools (DIY) available for hobbyists and the number of website. There are also these things for AVR (even a GNU C Compiler for AVR) but there are less websites (but there's AVRfreaks). SDCC works good for me for PIC (for all the hobby projects I do) but I rewrite some part of the Assembly Language then reassemble. I also write timing critical program in Assembly (except Switch statements, lazy to type so much sequence in assembly
Pic and Mikroelectronica
Mikro lab or Mikroelectronica this is an awesome system they have C as well as basic compliers I believe up to so many lines, but for the hobby user it is more than enough, also there is the Pic-pro or Pic-basic these are also great they start at 100.00 dollars as a matter of fact I have a legal copy of Pic basic if you want it we can work something out I upgraded to Pro, however I recommend Mikroelectronica, they have development systems such as the Mikro Lab and you can plug up a 2 or 4 line display, I love Microchip products, but they do have limitations as was noted earlier, they do not cross platform as was stated about the other chips the instruction set does vary with their chips, but I love them they are so easy to use, also being concerned with instruction set is not as critical if you are not using ASM, if you are then it is an issue, you do need to have a basic understanding for debugging even if you are only using a C or Basic complier. I am not able to comment of Atmel I do not use them I do however believe that I have a programmer and development system somewhere, I will look if I do I will send it to you free.
Also note that the spelling for Mikroelectronica is correct it is a Russian company they have a vendor in the USA, his name I believe is Don, nice guy.
I use CCS C Compiler for PIC (but I don't normally use their built in commands) on the company. For hobby I am using SDCC, though yes there are a lot of compilers that are good enough for hobby projects, just not complicated ones.
For instruction set, the AVR has a lot more instructions to learn than the PIC. Never yet did I do a program for AVR in ASM, just understand them. I used to write ASM programs for PIC until I came to know of all the good compilers where a lot of writing was compressed.
AVR is much faster than PIC16,18, here are some instructions in AVR that perform 2 operations like
movw r17:16,r1:r0
copy in one instruction 2 registers (moves 16bit data)
and assembly language is much simpler then PICs asm, because of multiple working registers (32) where you can store variables and perform between them operations, pic has only one Working registers and this is only bad thing that PIC architecture has, and it is really hard to code pic compared to AVR, I think that because of hard PiC asm language all PIC coders use C language for coding.
I started to learn coding with PICs and then tried out AVR and I found that AVR code looks 2x smaller compared to pic code becouse you don't have to preload working register with new value to make some operation and in this case one AVR instruction is2 PIC isntructions. so AVR performs 2x more operations compared to PIC in this particular case.
if we speak about price then LPC2101 beats all 8bit and 16bit microcontrollers like AVR,PICs.
So for motion controller I would choose LPC2101 + some CPLD of fpga, or just FPGA(as I already did some time ago).
Excuse me for stepping in on this matter, but myself am entering the world of "Microprocessor" programming, and according to what i have understood up until now, The title "Microchip vs Atmel" doesn't seem correct.
Microchip produces Microprocessors based on PIC "arquitecture", while Atmel makes them based on AVR, 8051 and ARM7 "arquitecture".
You are comparing PIC vs what from Atmel? AVR arquitecture?
Nowadays i have been trying to decide on what should i learn to program, on what "arquitecture" i should deploy my time, and start from there.
What "arquitecture" do you guys prefer?
8051?
AVR?
PIC?
ARM7?
Acording to this:
http://www.8052.com/forum/read.phtml?id=149579
ARM7 is the way to go, the new technologically advanced "arquitecture", the way of the future. What do you think?
PS: i typed "arquitecture" so many times it makes my eyes cry.
Dude, "architecture."
As for which to learn, wellllll.. it entirely depends on what you want to do, what your requirements are and how you prioritize them.
For the PIC, there are a LOT of free tools, and many free language implementations to choose from. The hardware is relatively inexpensive and the devices tend to be robust and easy to work with.
Is the architecture of the processor important to you? in other words, how easy it is to understand.. then the PIC might not be the best, as there are certain issues, especially if you program in assembly language, that you have to be aware of such as not trying to create a jump table that crosses a page boundary. Many a brave soul has been fornicated upon a mightily by that little gotcha. The limitation of working registers is also a bit of a pain (And i REALLY REALLY hate how they try to pass off a little sliding memory window as a 'robust register set')
The Motorola (freescale now) processors are good to work with, like the 68hc11, but they are costly to work with.
The luminary microprocessors use the Mortex M3 core which is a sub set of the arm7.. (9?) core and is optimized for the embedded universe, and motor control in particular. The ST micro's, (ST32's) are also pretty nice for three phase motor control.
So choosing one really depends on.. you. My only *REAL* advice on this, is.. don't fall into a "religion". In other words.. don't get to thinking that one particular CPU is the end all do all super cpu that is better than anything else. They all have their pro's and con's, just like our elected officials.
Horsedorf
The comparison is just like when you're buying a computer part like which is better AMD or Intel where we know that both of them have their own way to make their processors.
If you're using high level languages like C or C++ for programming these things the architecture won't matter. If you're doing it on assembly then you really have to know the architecture, and like what horsedorf said, there's a lot of people who fall into that page boundary crossing thing.
The comparison is much like performance-wise. The AVRs can do an instruction every cycle while the PICs needs 4 cycles to do an instruction. 1Mhz for AVR would be almost the same for a 4Mhz PIC (and if you're concerned with EMI, the higher the clock rate the higher the chances of making unnecessary emissions).
As you go along, you will know what MCU fits your needs, their pros and cons. You'll not always going to use one chip for all your needs. I am still studying ARM7 as of the moment but I choose from 4 MCUs for my needs, TI's MSP430, Microchip's PIC, Atmel's AVR, and the old 8051. A lot of my simple projects (like lighting and simple on/off control) are done with 8051 using SDCC. A little more complex projects are done with PICs. Then the most complex projects I use MSP430s and AVRs. MSP430 operates on 3.6V, favoring portable applications so it's not readily suitable for 12V 5V sources common to all my projects. Though the 5V to 3.6V regulator would be tiny, the fact that the chips are surface mounted makes it harder for playing around like the DIP ones (it is my tendency to do a lot of program for my hardware and used to swap chips for socketed DIPs, I can't do this on the MSP430s, it takes time to reprogram them
For enthusiasts, the PICs can overclock a lot.
perhaps AVR. 8051 and ARM7 is not exclusive to Atmel.
ARM7 is definitely one of the most powerful processors out there. most networking equipments use them, a lot of mobile devices use them, and a lot of complex industrial electronics use them. but certainly, an engineer who maximizes the balance of cost, productivity, efficiency would not use them on all but the most complex designs.
Welcome to the world of microcontrollers and embedded development.
You have a good point, when developing hardware that will be sold tens of thousands in quantities, like major routers, switches, etc.
I intent to develop alone (or die trying to) boards that do specific stuff.
The chip cost is not the main problem (unless its way more expensive),
the ease of program developing and not being crippled by hardware is.
Hardware should be plenty powerfull and adaptive to lots of situations.
Hard stuff!
Thank you, i enjoy being uneased
When I start a project, I begin by mentally building it up on top of hardware, as soon as I have my hardware planned, and blocks of it tested, I begin to develop firmware, after having decided which chip will basically fill out my needs.
For me, it is important that running out of memory on that chip does not mean having to drastically change the hardware platform. Having access to different, progressively bigger capability chips from the same manufacturer, with similar internal structure and somehow "compatible" assembly code, makes my life easier when I need to add a feature later on the design chain. Pin-out compatibility is also a plus.
I only code in assembly the critical parts of the software, I don't like to re-start a design on a new platform and having to begin coding assembly from zero.
Kreutz
Reply with Quote
