Controller vs. Electronics 2

[tulum]

Hey folks,

Simple question that I can't seemed to find a rule of thumb for...

When does it become worthwhile to use a microcontroller for a project vs. a typical electronic / TTL / CMOS circuit?

For instance if you want to make a horn circuit, or a temperature sensor is it more productive and cost efficient to use a microcontroller or just lay it out using caps, resistors, transistors, etc?

In industrial controls is there a downside to using a PIC? Are they considered as failsafe as an electronic circuit?

Anythoughts from experienced users would be appreciated.

Regards,
TULUM

 

[BobM3]

Microcontrollers are nice if you need to make changes. New code is much easier than replacing/adding chips. Adding functuality in the future can also can be easier with a microcontroller. They take up less space than a bunch of discrete chips.

I'm a big fan if you can't already tell. I've used them for years and had no problems. Give them a good power supply and protect them from transients.

 

[tulum]

Thanks bob..

What about cost... I looked at pricing on the web, but it seems to vary... is the software for programming the chips expensive... I heard some software is included in the chip price and some you buy separate.... I also heard the software costs you an arm and a leg? what do you use?

 

[itsmoked]

Strictly speaking micros are much cheaper than jungle logic especially since you can buy 5pin micros now. However if you have no development equipment the initial outlay and learning curve will make developing with them much higher.

On the third hand, your enterprise will probably tank because you will be competing with scads of companies willing to make that minor,(in the big picture), initial investment.

One of the biggest challenges in technical logic designs is in actually nailing all the requirements before starting the design. It can almost never be done! This is where using a micro is leaps and bounds beyond jungle logic.

You finish a jungle logic design, buy parts, assemble, tweak, and show the prototype and 999 times out of 998 times they will say, "Wow, Great! Can you make it go faster or slower?", or "can you make it wait a bit before starting?".

All these changes are easy with a micro and the jungle logic developer is left looking weak, stoopid, emasculated. While the micro developer is the knight in shining armor. (Well something like that anyway.)

If you are going to develop modern products you need to use micros.

You should consider a micro course offered by companies who make micros. I am most familiar with Microchip and know they have great courses all over. The courses are very informal, packed with info, and cost like a $100. You often get tools, development boards, manuals, and LUNCH!! for that. But mostly you learn a lot in 5-6 hours. Other companies, no doubt, also do this. Check out their sites. Atmel, Microchip, and Zilog.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[IRstuff]

Obviously, you are paying extra for both the flexibility and the excess capacity, for which you are trading speed and programming.

As a general rule, it may be more reasonable to prototype using a micro and then decide whether a discrete logic implementation is required for either cost, performance, or size.

TTFN

FAQ731-376

 

[MacGyverS2000]

How fast do you need it? How inexpensive does it need to be? What environment will it appear in, and if it's harsh, is it properly protected? What experience do you have in designing with one over the other?

For more than a couple of simultaneous and/or complicated functions, micros are typically the way to go. If you have a ton of I/O that does simple tasks, often times TTL/CMOS is the way to go. Both can be made robust in an industrial environment, you just have to know how.

If you know how to program on embedded systems, micros can make a full system in as little as an hour or two... if you don't, it could take you months. For low quantity, TTL/CMOS may be the least expensive in man hours (especially if you don't know how to program), but kill you in the long haul for large production runs.

So many ways to answer, but the answers depend on what you're starting with and where you intend to go.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[richs]

Hiya-

Voting along with the rest, micros have some significant advantages.

Programming time (the time to develop the program) might
exceed the development time for descretes, and debugging might
take longer, hoever the flexability of the devices make it
an invaluable tool in ones kit.

Indeed there are classes, however, there are some fine
tutorials on the web. The one that I started with is:

http://www.winpicprog.co.uk/pic_tutorial.htm

It is now somewhat dated using the 16F628 rather than (in
my opinion 16F648A), but still going with this tutorial
might be all that you need. My first programmer was home
built, also copied from the web. I have since "graduated"
to an ICD2 and PICKIT2 programmers.

Both assemblers and compilers are available for most of
the micros free.

The downsides:
1. Investment in software programming, both tools and
techniques.

2. Latency, both polling and interrupt service routines.

3. Instruction execution time.

4. Choosing the right part. With the Microchip (pic) parts
there seem to be hundreds to choose from. I've picked
3 in the 16F line (12F675, 16F648A, 16F877A) and going
to be using 16F886s in the near future. Same is true
with the ATMEL chips. There are so many configurations
with such a different mix of functionality that I found
it easier to limit my knowledge to 3 levels and learn
them well. Now, I can certianly cross to another part,
and the changes will be minor, but I'm only learning the
changes, not a whole different set.

5. Investment of a programmer. However, if you really want
you can roll your own. I have bought some from MPJA
that were under $20.00, parts included.

Going through the tutorial, you will start off with blinking
lights and work your way up. Instead of using the prototype
boards that Nigel uses, I used one of those prototype boards
where one pushes wires into the holes. For the tutorial,
the "shakey" connections are usually good enough.

Hope this little missive helps.

Oh, and as one wag mentioned:

"Why use a 555 when there are PICs!"

Cheers,

Rich S.