Atmel AVR vs PIC

[DirtyLude]

I need to make an order from Newark. I didn't realize their prices were so good on MCU's. I wonder what their shipping is to Canada.

For an alternative, you guys might want to look at the Luminary Micro (now Texas Instruments) stuff.

LM3S1608 50Mhz Cortex-M3 32bit
128k flash - 32k ram
ADC - I2C - SPI - UARTS
$4.65

or

LM3S811 50Mhz Cortex-M3 32bit
64k flash - 8k ram
ADC - I2C - SPI - UARTS
$3.70


For a pure price/performance comparison it's pretty nice.

I've only used the LM3S811 so far and it's very "8bit like", actually a little too 8bit like for my preference.

EDIT: Sorry, no DIP parts, though.

 

[DirtyLude]

Hmm... Unfortunately Newark doesn't give you a shipping price at check-out. That kinda kills it for me. I don't want to make an experimental order to see how much it's going to cost.

 

[davidbball13]

I hate to say this but I was on this forum over a year ago looking to get into PIC's. But there was not a centralized location for information for bigginers. Don't get me wrong but this place I'm sure is great for trouble shooting with PICs. I got into AVR's early in the summer and have been loving it ever since. I am ALWAYS on avrfreaks.net and #avr on freenode. With about over at least 120 active people on the IRC channel at a time it is extremely easy to get your problem resolved in a matter of time. And Avr Freaks forum is HUGE. There is a very large tutorial section and many many projects.

You can get a AVRISP MKii for 35 bucks for a full usb programmer. And for 50 you can get a Dragon which does almost everything. And the good 'ole stk500 is around 70 bucks.

This again is from the point of view of a 15 year old. I am no 'wizz at 'uC's....(yet. lol).
I just do electrics for fun and learn what I can on my own.




Also I've noticed that many many many people have been steering people away from avr's on this forum and telling new users that they should use PIC's (obliviously not in those exact words).

Both MPLab And AVR Studio are free which is great.
As for pricing you can get a 8 pin DIP avr with 1k of flash and 6 IO's and A/D converter, running at 20MHz (so 20 MIPs). for $1.40 (The attiny13)

and the very popular attiny2313 with 2K of flash and 18 IO's running at 20MHz for $2.20 in small quantity.

I am sure I will work with PIC's in the future. Just to try a new platform and test the waters.

Well my 2 cents have been given...

 

[Mr RB]

Microcontroller performance is not always relevant.

Most applications I make would perform just as well with a PIC/AVR or whatever. So microcontroller choice comes down to;
* cheaper
* easier to obtain
* already having stocks
* versatility, (ie stock one part, use them for lots of stuff)
* ease of development (this is really = experience with that particular part)
* quality of support and info available for that part

I stock PIC 16F and 18F and rarely bother using my 18F parts because for most stuff I can squeeze enough performance from the cheaper 16F parts of which I keep more stocks and will replace those stocks cheaper.

If you are going to make 50 microcontroller gizmo's in a year do you need 50 micros with the most MIPS or 50 micros that are versatile enough and pretty cheap?

 

[Russ Hensel]

The question only appears to make sense. Everyting depends upon best for what. Member of the forum will go on and on each answering their version of what the question means. Answers may differ, but all may be correct.

For example: **broken link removed**

 

[gramo]

Wow, I was still subscribed to this thread from my last post in 2007! May as well join in now that I've come this far

Hi davidbball,

The PIC vs AVR topic is always a hot one to say the least, and there is quite a following for PIC's here at electro-tech. It's not a good thing, nor a bad thing, just that the diversity is slightly biased to a particular product range.

A contributing factor to the bias might just be how popular PIC's are, and how much they are accepted by the tech community. Last year some time, Microchip **broken link removed** that they had shipped their 6 billionth micro-controller. If that isn't an example of how widely popular and used they are, then I don't know what is

My overall decision on which mcu came down to the compiler, and Swordfish is arguably the best structured compiler out there - though is designed for use with the 18F family of PIC's (There has been plenty of talk about 24F and dsPIC support in the future).

Microchip offer sample ordering which costs a flat rate of $7.50 (unless your a valued customer ). You can get up to 6 devices, bringing the cost down to $1.25/chip. Also, a full USB programmer with built in UART Tool, Logic Analyser and ICD for the PIC's can be brought for $35 from microchip

**broken link removed**​

Another point I want to share here is that the 16F and 18F devices are, in terms of technological advances, becoming obsolete. If ATMEL owned the 16F/18F range, they would all have been discontinued alas two or three individual "high-end" popular models. The 24F and dsPIC's are where the comparison should lie with the newer AVR power packages, as they are much closer together on the time scale of production. Further on that point, Microchip has and will not discontinue a product line, hence the overwhelming continual use of the dreaded 16F PIC devices and all of their deficiencies/drawbacks.

A common mistake for many people I see on threads like this is to profile PIC Micro's around there experience with 16F devices, then compare it to AVR's. An image comes to mind;

**broken link removed**​

Anyway, enough ramble out PIC's (see how easy it is to support something you know!!!). AVR's do have some REALLY good aspects. As a RISC and Harvard architecture device, ATMEL are much cleaner by design. Vectored interrupts, Flat memory model (no banking), Software Stack management, True indexing and indirect access and the instructions are executed in one cycle. There's really no excuse for AVR to not have a compiler on par with Swordfish...

All that said, use what works/available/your level of skill, and you will find support in any case. With no bias toward ANY mcu type, there's nothing quite like the feeling of programming your first device and watching that LED flash or "Hello World" displayed on an LCD/UART Tool.

The competitive nature between ATMEL and Microchip will only result in better products for us, the end users. So by all means, use AVR's if you love em, and use PIC's if you prefer

 

[Sceadwian]

Succinctly put Gramo. I'd add more, but you pretty much said all that needs to be said.

 

[Mr RB]

...

The 24F and dsPIC's are where the comparison should lie with the newer AVR power packages, as they are much closer together on the time scale of production.
...

Comparison of what? This is like arguing whether Lamborghini or Ferrari is best when 99% of drivers just need a decent functioning Chevy to get to work, (and the drivers asking the question are beginner drivers).

Have a look on the forum at all the people's requests for help with their microcontroller projects, especially newbies to micros. Just about every project can be made perfectly well with a lowly old PIC 16F, and there is tons of source code and project examples that they can use, and a 16F datasheet is much easier for a beginner to get to grips with than a 24F.

This whole bigger/faster/better argument is just lame. What percentage of the projects you make actually NEED a 24F or equivalent?? It's a very rare situation where only some pros and power-users NEED a top end micro and they are the exact people who DON'T need to ask which micro is best.

 

[gramo]

Hi Mr RB! Bare with me though your comment strays me away from the OP's direction

Just about every project can be made perfectly well with a lowly old PIC 16F, and there is tons of source code and project examples that they can use, and a 16F datasheet is much easier for a beginner to get to grips with than a 24F.

I wouldn't recommend direct entry to 24F micro controllers either, especially given the limited support by higher language compilers, though their day will come.

Perhaps the 18F range would be more suitable rather then the 16F devices. They offer a more conventional approach to programming/development over the 16F's and all of their flaws. The net is still littered with examples of 18F PIC's, but naturally doesn't have the 20+ years exposure time of the 16C/F range.

That opinion is coming from my 16F hobbyist experiences. I now use 18F PIC's, love them. I honestly have not had the need for more capability out of them. I guess it was to my own peril that I started off with the 16F84 thinking it would be *easier* and focused toward beginners given its exposure and ?acceptance? by the tech community...

This whole bigger/faster/better argument is just lame. What percentage of the projects you make actually NEED a 24F or equivalent??

For sure it is, my comment was focused toward those conversions where beginners with 16F's jump ship to a high end ATMEL product. The result is usually a profiled experience of the 16F range which associates all PIC's to a world of confusion and underpowered capability.

 

[Mr RB]

Hi gramo, you've made some good points there. I agree that a 16F84 is getting a bit obsolete now, and 18F PICs can be a good entry point for beginners too. But the 16F still have their place.

Its like the fact that transistors didn't become obsolete just because we have opamps now. There's still times when the small size/cost/simplicity makes a transistor best for the job. I just don't see 16F becoming obsolete any time soon, they are still cheaper for many low end applications and there is a 20 year resource of code out there for people to benefit from.

I'm not entirely happy with the direction of the new high end PICs. There's only so much math capability you need on a microcontroller. A "high end" PIC to me would be something with a fast 12bit ADC on each port pin, not just switching one 10-bit ADC between pins. Or some DACs!! What about an 8pin PIC with multiple 12bit ADCs, 12bit DACs, and some 24 or 32bit timers?? Now THAT would be a step forward, instead of focusing on huge ram and MIPS and math modules. Not everyone wants to make MP3 decoders.

 

[Sceadwian]

A 12 bit ADC, and a DAC would be specialty items. If you need 12 bit resolution you can get it, just use decimation. Need a DAC? You got one, a 1 bit PWM DAC. I would like to see more timers myself. I do think the math modules are really nice, but general purpose ones, not specialized ones.

I still personally think a hobby practical version of the FPSLIC that Ateml makes would be an absolutly incredibly thing. For those that don't know it's a huge chunk of FPGA core with an Atmel AVR micro processor built into the middle of it, so you can write your own peripherals and math functions.

 

[3v0]

The PIC 24H family has 12 bit ADCs but as mentioned there are one or two shared between many channels. I think the dsPICs can be had with even more resolution but am unsure about the number of converters.

These processors may be overkill but I do not think you could buy the external ADC and a processor for the same money.

3v0

 

[Sceadwian]

3v0 It's a micro controller, all you'd need to do is buy the ADC. The processor is already there. I'm sure you could use a lower end PIC and a higher end ADC at lower cost than a 24H or other such high end micro controller.

 

[davidbball13]

Wow, I was still subscribed to this thread from my last post in 2007! May as well join in now that I've come this far

Hi davidbball,

The PIC vs AVR topic is always a hot one to say the least, and there is quite a following for PIC's here at electro-tech. It's not a good thing, nor a bad thing, just that the diversity is slightly biased to a particular product range.
.....................
The competitive nature between ATMEL and Microchip will only result in better products for us, the end users. So by all means, use AVR's if you love em, and use PIC's if you prefer

Hello, I am in no way saying anything wrong with PIC's both MCU's are amazing and both play their part in the MCU market. There is NO such thing as a perfect anything. Nothing is perfect guys and 'gals. My opinions are just truly based on the fact of getting up and running with avr's and starting to get working on them in my free time and ease of use.

Although I did read up a lot on the New PICkit3 when they screwed up on it a ton. there was a youtube video of a guy ranting on it and MicroChip saw it and is addressing the problems.

I know people are talking about MIPS and advantages of each processor at its core but there will be a lot more people at that level if the learning curve is as small as possible.

As in the nontechnical world there will ALWAYS be a Ford guy or a Chevy guy. I am a Toro guy myself for lawn mowers. As my father is a Honda fan. We always argue about the mowers. Even there is not much of a difference at its core. The small things like Personal Pace self propulsion and double blade cutting is what makes the end user happy. Now I see I am getting very very off topic. What were we talking about again?...............


Oh yeah MCU's! Fun stuff they are.
Now this post is going to go no where! This forum is very very biased to the PIC side of the equation. If you were to take this on to a full on internet war it would be ridiculous.
Now many/if not all of you on here use electronics for you're job. In which I am assuming you are getting paid for. If you're boss likes PIC's and you're getting paid. In exchange you may start to like PIC's. His color starts to rub off on you. And if you're using a PIC in the manufacturing aspect is completely different then going onto a hobby forum and talking about aspects that many many hobbies ever get/have to get into.