Learning PIC - where to start?

[Jon Kirwan]

I second your opinion. PIC is for perverts. Experience with PIC is brain
damaging. PIC forces the programmers to bend over mentally. After they
broke, they can see only PICs and think only in the PIC manner.

That's an excessive bit of bigotry if I've seen it. I don't expect to
change your idiocy here, but for others I'll write a little just to
compensate for your lacks.

I use quite a wide variety of processors, spanning from DSPs from TI
and Analog Devices to the Am2900 bit slice processor family, as well
as doing chipset testing for the P2 family working at Intel. 8-bits
from 8051's, 8080s, 8085s, Z80, 6800, 65C02, 65C816, AVR, and so on.
Yes, PIC included. RISC, including from 1986 with MIPS R2000, then
M88k and Dec Alpha. Etc. Earliest experiences were with the PDP-8,
PDP-10, PDP-11, IBM 1130 (assembly and FORTRAN II), HP 2000 and 21MX,
and 8080. PIC is not how I think, despite having used them for some
percentage of projects since when Microchip finally decided they
wanted to sell them into smaller markets than rice cookers and stopped
requiring non-disclosures.

One of the things that has marked the company, Microchip, is their
dedicated support for their development tools and chips, even to small
manufacturers. Unlike some companies (Atmel is included here),
Microchip doesn't insist on knowing exactly how many chips of theirs
you will sell "real soon" before they will supply you with early
samples. When I have so much as a slightly flakey power switch on one
of their programming tools -- and I'm thinking right now of an
experience I had a few years back with a ProMate II which they no
longer even sold at the them, having replaced it -- and they sent me
another one without any question, with a return posted box included.
This hasn't happened just once, but on at least three occasions with
old tools they no longer sell. They still support them, so far as I
can tell, forever. And without question or quibble. On compiler
support, I was put through (despite being only a small purchaser of
their parts) directly to the compiler team and they corrected a
problem we had within a week of my presenting sufficient details so
that they understood the issue (compiler-generated statics in the
presence of interrupts.) That I was allowed to make that connection
is pretty good, considering their size.

The actual instruction set is not much on the business radar screen,
so far as I'm concerned. It's about a lot of other things, but I have
a hard time understanding why an instruction set makes you say such
ignorant things about a product line and people who might dare to use
them.

No, I don't work for Microchip and never have. Never derived a single
penny from direct or indirect contracts, so far as I'm aware. However,
my experiences are mostly in the 1000-5000/year product area and in
that arena they back you up pretty good. As a hobbyist, no so bad,
either.

Although I mentioned Atmel in a negative context above (you can look
up the detailed experiences I've reported, if you care), I like their
parts just fine and I've used them in products with excellent success
(I've also reported those positive experiences, as well -- look them
up.) I also like the 8051 family. And I like pretty much everything
I see, if it fits the application space and if the business side is
right.

I see no reason to denigrate folks like you have. But I suppose that
is more your problem than anyone else's. I just figured I shouldn't
allow your bigotry to go unchallenged.

Jon

 

[Jon Kirwan]

no longer even sold at the them

should be: "no longer even sold at the time"

(spell checker and my too-quick acceptance.)

Jon

 

[krw]

))))))

In fact, I am absolutely indifferent to what CPU, OS, etc. to use. There
is only one criteria: profits/costs. It is such a pleasure to tease the
zealots and enthusiasts of all kinds and observe them nervously
defending their toys and beliefs!

If something is deformed from the beginning and/or difficult to learn,
the spent effort make it seem more valuable to you. This is a natural
human reaction.

You are spoiled. I never ever had any need for the vendor support.
Except for some very specific questions that are not addressed in the
datasheets. Never I begged for the free samples and evaluation boards. I
used to build my own tools and write the software utilities; this is how
the experience and the self reliance is developed.

I thought your "one criteria" was profits/cost. Certainly good vendor
support helps costs and reinventing wheels hurts profits. Which is
it?

<snip>

 

[Jon Kirwan]

))))))

I guess you are proud of it.

In fact, I am absolutely indifferent to what CPU, OS, etc. to use. There
is only one criteria: profits/costs.

I don't find much to disagree with here.

It is such a pleasure to tease the
zealots and enthusiasts of all kinds and observe them nervously
defending their toys and beliefs!

I enjoy my work, no question. But I'm no zealot for specific
processors. I find I enjoy working with all of them that meet the
needs of my clients.

If something is deformed from the beginning and/or difficult to learn,
the spent effort make it seem more valuable to you.

Well, this is just you putting words into my mouth that I never spoke.

This is a natural human reaction.

Yes, putting words into other peoples' mouths happens far too often.

For the matter of pissing contest, I mention the CPUs which I have NOT
used yet: Z8, NEC, PIC32, M16C, PSOC, MSP430, ST9, ?oldFire. May be I
have missed something. Yes, PIC excluded. The latest experiences are
with BlackFin-527, the earliest with IBM 360.

I didn't list everything, either.

You are spoiled. I never ever had any need for the vendor support.

Oh, that's bullshit.

Except for some very specific questions that are not addressed in the
datasheets.

Ah. So I agree with you, then.

Never I begged for the free samples and evaluation boards.

I notice that you are changing the subject, completely. Which is
disingenuous and probably due to the fact that you were caught being a
bigot. Oh, well.

Taking you on this change of subject, though, sampling is important
not because it might be free.. but because it can provide a part that
is a better fit for the application space with some lead time before
production. I don't bother with samples in the case where it is in
production, already. What's kind of funny here is that you just bring
this up for no reason and make up your own strawmen to knock down.
Must be a reason why.

I
used to build my own tools and write the software utilities; this is how
the experience and the self reliance is developed.

Heck, I have done much the same, including linkers, assemblers, and
compilers where appropriate. I don't consider this necessarily a mark
of pride -- it's just that I do what is necessary, when necessary, and
it has been, on occasion. I've lived long enough, I guess. I
consider myself self-reliant, as well. So?

The worst problem of Atmel is that they change the product lines at
will. Even if they offer the newer replacement for the legacy chip, it
is not 100% compatible. So the code has to be modified and everything
has to be checked again.

I have several problems with the lack of fit of their business model
with the business models of clients I've done more of my work with.
But I don't feel upset about it. It's just a difference in goals.

So you feel yourself obligated to MicroChip, do you? As if they did you
a favor? Making you feel obligated is #1 trick of marketing.

No. Again a strawman.

The PICs below PIC24 are the hassle from the programmer point of view.
PIC24 and PIC32 are sensible.

I like them, too.

But of course, I see that while you earlier wrote above that all you
care about is "profits/costs" here you limit yourself. There are
applications where the "PICs below PIC24" are worth the hassle exactly
in the way you seem to earlier describe. I guess you just avoid those
projects, which is fine with me.

Oh... AVR does have problems which certainly show up when the quantities
exceed 10k or so. Here is one: the flashloader of the AVR has to be in
the same address space as the program. So a random glitch can activate a
flashloader and corrupt the code. There are the other problems as well.

From my experience, Motorola/Freescale CPUs are the best from the
technical, programming and manufacturing standpoints, but they are not
very cheap, and they are not for beginners and amateurs.

Strawman. Besides, you aren't talking to a beginner or amateur. 35
years doing stuff like this, now.

The 8051 is the ugly core. Only PIC can be worse

In its time, it fit a lot of application spaces as there wasn't all
that much going 'round and choices were limited. These days, things
are a little different. But SiLabs has a few parts, for example, that
offer unique combinations that may fit some application spaces better
than others. They have a place. Obviously, your experiences are
sadly too limited in that regard. Perhaps because you self-limit
them.

Zealots defending their beliefs. That's exacly my point )))))

Well, you didn't successfully make any points. And you are still a
bigot.

Jon

 

[Jon Kirwan]

I thought your "one criteria" was profits/cost. Certainly good vendor
support helps costs and reinventing wheels hurts profits. Which is
it?

<snip>

Bingo!

Jon

 

[Jon Kirwan]

<snip>
From my experience, Motorola/Freescale CPUs are the best from the
technical, programming and manufacturing standpoints, but they are not
very cheap, and they are not for beginners and amateurs.
<snip>

A real craftsman uses tools appropriate for the application and works
to expand the scope/range of those tools and necessary experiences
with them, so as to be better able to bring more options to the table
for clients.

To be proud of being limited is what struck me funny about your
comment above It reminded me of an phrase I once heard -- "To a man
with a chainsaw, everything looks like a tree."

Jon

 

[Vladimir Vassilevsky]

A real craftsman uses tools appropriate for the application and works
to expand the scope/range of those tools and necessary experiences
with them, so as to be better able to bring more options to the table
for clients.

To be proud of being limited is what struck me funny about your
comment above It reminded me of an phrase I once heard -- "To a man
with a chainsaw, everything looks like a tree."

Wow! Tons of truisms, irrelevant BS and personal insults. This is how
zealots defend their beliefs. I used to think we talk about the technology.

"It is dangerous to separate tiger cub from his mother, and a man from
his delusions" (c) LaRochefocault (or Kipling ?)

OK, I have to return back to work now. Just another ~20k lines project.
For AVR, BTW. Enjoy your Sunday


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

 

[Paul E. Schoen]

And to a man with a hammer, everything looks like a thumb

Wow! Tons of truisms, irrelevant BS and personal insults. This is how
zealots defend their beliefs. I used to think we talk about the
technology.

"It is dangerous to separate tiger cub from his mother, and a man from
his delusions" (c) LaRochefocault (or Kipling ?)

OK, I have to return back to work now. Just another ~20k lines project.
For AVR, BTW. Enjoy your Sunday

I'll just throw in my own opinion that the Microchip family seems a good
place for someone like the OP to start. My earliest projects were with the
8085 and then the Z80, which I really liked, and then we made a product
with a Z-world SmartCore which uses a Z180. The earliest projects were all
assembly language, and the newer ones used various amounts of C.

I also used some Signetics PLDs for fairly simple applications, but for
some more recent projects a microcontroller seemed more appropriate. At
that time (about 10 years ago) I looked at such things as the Basic Stamp,
and eventually decided to try the Microchip PIC16C63. It was difficult to
learn, especially with respect to bank switching and Read/Modify/Write
problems, but I found their tech support very helpful, and their parts
suited my needs at the time. I invested a lot of time and money learning
their products, and that is probably the main reason I did not really
consider using other parts. My projects are very low volume, so the
development effort is a major part of the cost, and the cost of the ICs are
just about insignificant.

It would be helpful for anyone considering the Microchip products to read
their forums at http://forum.microchip.com, where you can learn from the
experiences of others, as well as receive timely advice from many experts.
I'm sure other companies have similar discussions, and each product family
has its own quirks, which are often disclosed more candidly in forums than
on corporate data sheets and errata.

Also I would like to praise Microchip for their lifetime support of their
development tools, which they have replaced even when damaged by my own
fault. They have been good about samples, too, but recently they are
charging about $8.00 per order for handling and shipping, so sometimes it
is more expedient to just order a few parts from Mouser or Digi-Key. I also
have enjoyed two of the Microchip MASTERs conferences in Arizona, where you
get intensive hands-on training in classes that are tailored to your
specific needs or interests.

Paul
www.pstech-inc.com

 

[Jon Kirwan]

Wow! Tons of truisms, irrelevant BS and personal insults. This is how
zealots defend their beliefs. I used to think we talk about the technology.

You were the __very__ one starting with insults, not talking about the
technical aspects but actually denigrating a class of individuals. Now
you dare to call the kettle black. It makes it all the funnier, in a
sad way.

Especially since you seem to imagine that every solution seems to see
freescale in it. Are their any 6-pin varieties? No. There are
applications where that is appropriate as SOT23-6 and there are
processors that fit that niche. I find your narrow and bigoted
attitude the problem. You want to turn this around and claim that it
was me handing out the insults, when all you need to do is look at
yourself for that. You project onto others what is really just you.

You need a mirror.

"It is dangerous to separate tiger cub from his mother, and a man from
his delusions" (c) LaRochefocault (or Kipling ?)

OK, I have to return back to work now. Just another ~20k lines project.
For AVR, BTW. Enjoy your Sunday

I'm actually working on mine, like you. I'm glad to see that I won't
have to hear much more from you on this subject, today.

But if and when I see it again, I will remind you and others of this
conversation and your initial insults to start it, if motivated to do
so.

By the way, I'm NOT working on a project using "PICs below PIC24." I'm
currently working with an NXP. Not that it matters. But I figure
your bigotry reaches into deciding what I must be doing and it's worth
saying your prejudices aren't always correct.

Jon

 

[Jon Kirwan]

And to a man with a hammer, everything looks like a thumb

Yes, but I liked the way I heard it better! More picturesque, in a
way. And I've "kind of" seen it in action in real life, too, using a
chainsaw for chisel-work. An art form of its own, I suppose.

Jon

 

[Paul E. Schoen]

Yes, but I liked the way I heard it better! More picturesque, in a
way. And I've "kind of" seen it in action in real life, too, using a
chainsaw for chisel-work. An art form of its own, I suppose.

I knew an old guy who specialized in chainsaw art, and he made some pretty
cool things from big logs. I think he preferred Paulownia trees.
http://en.wikipedia.org/wiki/Paulownia

Paul

 

[Nobody]

No, the point is that the best way to learn about pic is to forget about
the PIC and learn about a proper processor, at which point you'll know
better.

Where's the fun in that? I'm already familiar with "proper" processors.

FWIW, so far as the "brain damage" is concerned, bank-switching seems to
be the least of it. I've found the trickiest part being getting my head
around all of the logic inversions in conditionals, i.e. skipping the next
instruction if the condition holds, but the next instruction is a goto
which skips the "meat", and the condition is testing the Z flag (so zero
is true) or testing a pin wired to a switch where zero => pressed.

 

[Rich Grise]

Odd for an improper processor it seems to be pretty popular.

Microsoft crap is also popular.

Ergo: "popular" != "good". ;-)

Cheers!
Rich

 

[Rich Grise]

If you are heading to Atmel anyway, why 8051 instead of Avr?

In my case, because I've used them before. ;-)

Cheers!
Rich

 

[JeffM]

[...]my boss[...]wasn't happy -- he really, really wanted a "built from
scratch" solution, being untrusting of the TCP/IP stack in Linux, the
model-class USB driver someone else would have written, etc.

Does your boss have the technical chops to question those
--or has he been listening to the bleating crowd
about *that software built by amateurs*?

We told him that, sure, that was doable, but
you're probably now talking at least a year of development
rather than a couple of months. That pretty much dissuaded him
from continuing the discussion...

I was reading the Slashdot coverage of the Terry Childs case
(the Frisco sysadmin who's in jail for doing his job too well).
One guy said that when his (MBA) boss
wanted to be in the loop on tech stuff,
the techie buried him in paperwork
(like opposing council in lawsuits).
Pretty soon the boss just said "You handle it".

 

[JeffM]

Microsoft crap is also popular.
Ergo: "popular" != "good". ;-)

Widely used != Popular
Sworn at != Sworn by

 

[JeffM]

Seems to me he's in jail for deciding to play politics at work... and losing.

....if that's what you want to call following standard security
protocols.

I mean, I agree the charges (particularly the modem stuff)
appear largely trumped up (by managers who got pissed off
and are looking to punish the [guy])
-- and aren't going to hold up in court --,

The boss is going to come out looking like an incompetent weasal.

but if your boss stands there
and tells you to hand over the network passwords,
sure, you can certainly mention that doing so
with strangers in the room is a massive security risk
and is he really sure he stills wants to do this?,

Not setting up escrow accounts
in case your techie gets has a heart attack, or gets hit by a bus,
or just decides that you're a jerk he doesn't want to work for any
more
is *your* fault, boss; don't try to put this on somebody else.

but if he says 'yes,' well, if you then still don't hand over the goods,
I think it's pretty clear your job is over.

....and as a former employee,
what obligation does he have to cooperate AFTER that?
Again, the boss sounds like he should have a job
that involves asking "Do you want fries with that?".