gnupic: Re: [gnupic] Learning PIC on the job
Subject:
RE: [gnupic] Learning PIC on the job
From:
Chen Xiao Fan ####@####.####
Date:
16 Aug 2005 04:36:33 +0100
Message-Id: <3B8AEFFADD3DD4118F8100508BACEC2C07F7747B@spex>
As you know I disagree with you on the choice of chips and
programmers for new beginners. :)
Never mind, let me explain in more details.
1) Chips of choice:
The new trend is that all the new 16F chips are standard flash
and does not support bootloader. They are supposed to be the
main stream PIC16F chips in the near future. In terms of
price, they will be the cheapest PICs. So I will recommend
the new 8/14/20 pins of cheap PICs to new beginners than
big 16F88 or 16F87xA. Paging is a pain. Of course, 18F is
a good choice as well if price is not a problem. dsPICs may
be too new to use.
2) Porgrammer of choice:
I will recommend PICkit 1 as the choice for absolute beginners
and Wisp 628 (or PICkit 2) as the choice for the other beginners.
ICD2 may be for those who want to work with dsPICs but the
problem is that it does not support right now.
3) Why I am against the use of JDM/Tait?
I am not against the idea of using bootloader if it is supported
by the chip used. I like to use bootloaders if possible, e.g.
for 18F USB and dsPICs. I am against the idea of using JDM/Tait or
similar. A newbie will be much more likely to fry a raw PIC with
JDM/Tait (even the PC ports) because of the inadequate protection.
Unfortunately for legacy reasons, there are too much obsolete
information on the Internet for 16F84 and JDM. I read too much
reports of difficulities with JDM type of programmers in various
forums.
4) To build or to buy?
Expensive or cheap is subjective. If time is more expensive, then
PICkit 2 or the fully assembled Wisp628 may be well worth the money.
If people think PICkit 2 or fully assembled Wisp628 is too expensive,
they can build one as well since all the information are there.
For me this is not a chicken and egg problem. A lot of the people
has mentioned this, why not leave the programmer developing to
those experts or pros? Why burden newbies with the task of building
a programmer (sometimes frustrating)? Of course it they consider it as
a good exercise, I have no objections. Then build a proper one.
5) Basic Stamp
If one can afford to use a Basic Stamp, why not buy a
PICkit 1/2/Wisp628? I do not know much about Basic Stamp but I think
it is only popular because of the HLL. I even think that it is again
a legacy product like 16F84 IMHO. Still I have no experience and I can
be totally wrong here.
6) "I beg your pardon, how to mimic the desktop computer?"
By the way, I do not understand that to mimic the way how a desktop
computer works to the low end embed PIC world be appropriate. But I do not
know what exactly do you mean by "mimic". One does not need to know
how a Pentium works before doing programming. One does need to know
how a PIC works before doing programming, even only for blinking a LED.
7) This is just my two cents.
I am not a pro myself. So far I've only worked with 12F/16F
(all <=2k words so no paging) at work. I am only starting to mess
with 18F4550 USB and dsPICs from a hobbyist's point of view
quite recently. So maybe my opinions are more appropriate for new
beginners than a pro. still this is only my two cents.
Regards,
Xiaofan
-----Original Message-----
From: Byron A Jeff ####@####.####
Sent: Tuesday, August 16, 2005 9:58 AM
To: ####@####.####
Subject: Re: [gnupic] Learning PIC on the job
As you know I disagree. I think it's wrong to peg newbies as dumb or stupid
simply because they have a lack of knowledge on a subject. Simply explaing
to
them in clear language what to do and most of them get it.
For example traditional programming vs. bootloaders. Simple analogy:
Imagine
if every time you wanted to install a new program on your computer you had
to
open the case, remove the hard disk, put the hard disk in a special machine
to load the software, then reverse the process. Further imagine that you had
to do this each and every time you wanted to change the software on your
computer. This is analogous to the traditional programming process for
a microcontroller.
A bootloader works like a BIOS. It's a program that's in your computer that
can load other programs. Now PIC chips come to you completely blank. So
you use the traditional programmer to put the bootloader onto the chip.
From then on the chip can load programs for itself.
Each method has its advantages and disadvantages. Traditional programmers
work with blank chips. Traditional programmers also use no resources on
the chip such as memory or I/O pins. Bootloaders can give you a lot of
flexibility. There are bootloaders that wiggle the MCLR pin in a special
way to program the chip for example. However by definition they must occupy
some memory space in order to operate. Also you have the chicken and egg
problem of getting the bootloader onto the chip in the first place.
But at the end of the day bootloaders mimic the software installation
processes of the computers we use every day. And frankly mirroring
a known process is simpler for a new user than an unfamiliar one.
BTW why do you think that a newbie is any more likely to fry a raw PIC as
opposed to a bootloaded one? The fact of the matter is that lots and lots of
newbies got started using the Basic Stamp, which had a serial EEPROM
bootloader coupled with a simple HLL interpreter.
>
> I will recommend Wisp 628 and PICkit 1 now for beginners.
So you have to get two programmers? Or is it one or the other?
> Later I will recommend PIckit 2 after we get it to work with
> Linux.
And so a newbie will have to plunk down $50 just to get started?
And for this you get 1 part (16F690), a pot, some LEDs and
a programmer? Seems pricey to me to pay for convenience. Then on
top of that any Linux person will have to wait until Mchip gets around
to releasing the protocol to talk to the programmer.
Hmmm.
BAJ
