"Without a microprocessor..."

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
This is not a negative comment on any post or anyone's strategy, but an observation.

In the past, you might have had to justify using a processor to solve a problem. Domain specific ICs, discrete logic, even discrete transistors would have made more sense on the grounds of ease, complexity, and cost.

But today, I find the reverse. When someone says, "I want to do this without a microprocessor, I really need to hear the rationale for NOT using one before I feel I can help, since there is a very good chance they are confused about micros or avoiding learning what they consider "complicated" while it is actually the simplest and most effective answer in most cases today that seem to need one.

Obviously, proper analysis and development of a solution is necessary, but we all use mental shortcuts for practical reasons, and today, anything that interacts with sensors and appears to make decisions, well, why NOT a micro? Then we can figure out an alternative...
 

Alec_t

Joined Sep 17, 2013
10,906
well, why NOT a micro?
Of course micros offer many advantages and for commercial products would normally be the first choice, but for educational reasons it's important to have a good understanding of the function/purpose of basic discrete electronic components. The best way of obtaining that understanding includes a hands on approach, hence why so many forum posts are about projects which don't involve micros. Even where a micro is involved, peripheral circuits will often require the use of discrete components.
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
Of course micros offer many advantages and for commercial products would normally be the first choice, but for educational reasons it's important to have a good understanding of the function/purpose of basic discrete electronic components. The best way of obtaining that understanding includes a hands on approach, hence why so many forum posts are about projects which don't involve micros. Even where a micro is involved, peripheral circuits will often require the use of discrete components.
Yes! "I want to learn about discrete components" is a perfectly good rationale, so we agree on that. Also, even if you are going to use a micro it will inevitably require some discrete circuitry, so you learn that too.

Much of the time, though, it seems the opposite. "I don't really know either thing but micros seem more complicated, so how do I do this practical thing while avoiding the best solution".
 

LesJones

Joined Jan 8, 2017
2,497
Hi Yaakov,
I was just about to post a reply totally agreeing with your comment but after reading Alec's I can see both sides. I was just looking for two recent posts (One for a very small unit that started buzzing one minute after being powered on and another that was required to produce 4 pulses when triggered) that were perfect examples to support your post. I did not even think about the learning the basics reasoning as I have progressed from valves (Tubes.) to transistors to integrated circuits to microprocessors over the years.

Les.
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
Hi Yaakov,
I was just about to post a reply totally agreeing with your comment but after reading Alec's I can see both sides. I was just looking for two recent posts (One for a very small unit that started buzzing one minute after being powered on and another that was required to produce 4 pulses when triggered) that were perfect examples to support your post. I did not even think about the learning the basics reasoning as I have progressed from valves (Tubes.) to transistors to integrated circuits to microprocessors over the years.

Les.
I agree with Alec, you know. if someone says they want to learn about discrete components, that’s a good rationale. It’s the people who, in ignorance of one or both make a judgment they should avoid something actually better, that have turned the previous thing—looking for an excuse to use a micro—on its head.

15 years ago it would have been easy to write an inverted version of this post, questioning people with a practical, not educational, purpose trying to shoehorn a micro into a project that a 555, for example, would fit nicely just to use one,
 
Last edited:

danadak

Joined Mar 10, 2018
3,891
Of course micros offer many advantages and for commercial products would normally be the first choice, but for educational reasons it's important to have a good understanding of the function/purpose of basic discrete electronic components. The best way of obtaining that understanding includes a hands on approach, hence why so many forum posts are about projects which don't involve micros. Even where a micro is involved, peripheral circuits will often require the use of discrete components.
There are micros out there that are best of both worlds, can give you a fabric to experiment with
basic logic, and a ARM core (if you choose to use it) to move on to a micro solution. Or better than that
use both discrete logic and micro. All onchip and routable, both onchip and to pins for external interface.

Whats to not like about that ?


Regards, Dana.
 

Delta prime

Joined Nov 15, 2019
164
In my travels to be a good successful research scientist electrical engineer or electronic technician one must acquire skills that are not taught in undergraduate lectures or labs they must become familiar with a wide range of commercially-produced equipment and learn to use it effectively in addition they must be able to design apparatus that cannot be purchased and must build it is well.To use micro-p or not to use micro-p you will use and learn both eventually.
 

dendad

Joined Feb 20, 2016
3,316
Just about everything I make now has a micro. But, just a micro is mostly uesless. There is always extra electronics. If you just want to flash an LED, a 555 may be great.
My home made Christmas lights, made many years ago, has 24 LEDs, 6 resistors and 6 caps (not counting the power supply) and the "brain" is a single 74HC14 inverter. It goes very well. But if I wanted a variable LED display, a micro is the first choice.
In my experiance, a micro is the best way to go as changes are almost always wanted later. It is so much easier to change the code!
The argument that you do not learn electronics by using a micro, like an Arduino is a bit biased I think.
Over the many years I have been playing with electronics, I have used all starting with those glowing thingos in glass, up to them with thousands of transistors inside. There is room for all, but I do believe most digital things are best suited to have some sort of micro buiried down deep inside.
Saying that, I really like those digital clocks made with transistors and diodes!
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
In my travels to be a good successful research scientist electrical engineer or electronic technician one must acquire skills that are not taught in undergraduate lectures or labs they must become familiar with a wide range of commercially-produced equipment and learn to use it effectively in addition they must be able to design apparatus that cannot be purchased and must build it is well.To use micro-p or not to use micro-p you will use and learn both eventually.
It’s important to understand the context of my post:

1) It assumes that most practical projects today that could use a micro probably should, which I stand by.

2) That a valid rationale for not doing so certainly changes 1, above.

3) That a very large percentage of requests for X without a microprecess or don’t have a good rationale, usually based on ignorance. (e.g.: too complicated, hard to implement, expensive, hard to learn, etc.)

All practical projects will require learning discrete components and some basics of circuits for interfacing the µP to the world, and not knowing how to program is a major skills gap in the current world.
 

tindel

Joined Sep 16, 2012
704
I used to get this question from my colleagues regarding my home projects - “Why would you NOT use a micro controller?”. Lots of reasons are possible.
  • I work in a discipline where most (but not all) solutions are still not embedded.
  • Some industries still require hardware based solutions for things for reliability purposes.
  • As a home project I want to do things however I damn well please.
  • Micros often simplify things too much for educational purposes
  • Solving problems in the analog domain is still a valuable skill to have to meet some specifications.
  • I’ve been in academia now for a few months after a 12 year hiatus in industry. I’m learning that most of the students have very little understanding of what is happening in the micro within the analog domain that makes the micro work. They just hook things up with no regard for the system or the impedances etc.
  • Why add firmware - yet another skill - if it’s not required?
  • I could go on.
I’m not disagreeing with you as I also believe embedded system design is also a required tool these days for any engineer or technician worth their weight, but there are still legitimate reasons to keep solutions in hardware.
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
I used to get this question from my colleagues regarding my home projects - “Why would you NOT use a micro controller?”. Lots of reasons are possible.
  • I work in a discipline where most (but not all) solutions are still not embedded.
  • Some industries still require hardware based solutions for things for reliability purposes.
  • As a home project I want to do things however I damn well please.
  • Micros often simplify things too much for educational purposes
  • Solving problems in the analog domain is still a valuable skill to have to meet some specifications.
  • I’ve been in academia now for a few months after a 12 year hiatus in industry. I’m learning that most of the students have very little understanding of what is happening in the micro within the analog domain that makes the micro work. They just hook things up with no regard for the system or the impedances etc.
  • Why add firmware - yet another skill - if it’s not required?
  • I could go on.
I’m not disagreeing with you as I also believe embedded system design is also a required tool these days for any engineer or technician worth their weight, but there are still legitimate reasons to keep solutions in hardware.
Yes, that’s a good list of reasons you might not want to use a µP, but you did see that my point is actually that in the current environment, your list is one of exceptions, not the norm, and that they are not because on ignorance but because of knowledge.

So, when someone says, “How do I do X without a µP?”, where X is a task most easily and appropriately done with one, the first question I have is “why no µP?”

If they respond intelligently and with some reasonable answer (any of yours, or others) all is well, but if they respond with something mythical (cost, power, complexity, etc.) then that needs to be addressed before a solution, in my opinion.
 

tindel

Joined Sep 16, 2012
704
Yes, that’s a good list of reasons you might not want to use a µP, but you did see that my point is actually that in the current environment, your list is one of exceptions, not the norm, and that they are not because on ignorance but because of knowledge.

So, when someone says, “How do I do X without a µP?”, where X is a task most easily and appropriately done with one, the first question I have is “why no µP?”

If they respond intelligently and with some reasonable answer (any of yours, or others) all is well, but if they respond with something mythical (cost, power, complexity, etc.) then that needs to be addressed before a solution, in my opinion.
I personally think you are splitting hairs.

I generally don’t understand all of an OPs requirements or motivations for any particular requirement to use a micro or not, and it usually doesn’t matter. I answer questions as I wish and as I have time on topics that interest me. I try not to set their requirements for them. After all, I’m not on their staff, or in their course, or at their home workbench.
 

BobTPH

Joined Jun 5, 2013
2,255
In my experiance, a micro is the best way to go as changes are almost always wanted later. It is so much easier to change the code!
This is, I think is the most overlooked reason for using a micro.

You might think, “why use a micro when I can do it with one quad nand gate?” The reason is, when you realize later that you want slightly different behavior, you will end up throwing away your design and starting again from scratch.

With a micro, it is usually just a mod of the firmware, which a competent programmer can do in less time than it takes to heat up the soldering iron.

Bob
 

visionofast

Joined Oct 17, 2018
68
usually for expensive instruments like medical ones, It must as repairable and replacable as possible regarding to expenses and support, so it's to be based on basic logic parts instead of a uc with an unknown firmware.
plus more stabality in many cases.
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
I personally think you are splitting hairs.

I generally don’t understand all of an OPs requirements or motivations for any particular requirement to use a micro or not, and it usually doesn’t matter. I answer questions as I wish and as I have time on topics that interest me. I try not to set their requirements for them. After all, I’m not on their staff, or in their course, or at their home workbench.
I am not sure why you think I am making rules for you. You can do whatever you'd like, I am not interested in demanding anything of you. But, my participation here is to help, when I can, and that's the context of my original post.

In my career as a consultant, I learned it was necessary to clarify the roles and responsibilities of my consultancy and the client at the outset of the engagement. In particular it was essential to make it clear to the client that they were the expert in their problem and were hiring us as experts in a solution to it.

To that end, our discovery phase was completely without reference to any solution the client had in mind or might have attempted to implement, because not being experts in solutions to the problem, their attempts had failed or fallen short—or they wouldn't have been engaging us.

Many, many consulting engagements are to "finish" a project started by the client, making their ill-conceived solution solve a problem that it can't, or will be a poor and fragile one.

To that end, I perceive my input here as being similar. To the extent the TS provides input that I can see is a valid rationale for some constraint or set of constraints, I can work to help them because that is part of the problem space. In other words, when the constraints really are part of the problem, and not imposed by the notional solution, I can help.

On the other hand, if I ignore the (ignorant) arbitrary constraints (and the could be in many forms, there are other common examples here) I don't really think I am "helping" anyone. Things will not end well, at least not for the proposed solution and the best outcome will be jettisoning the original approach and doing something different. So, I'd prefer they did that at the beginning, notwithstanding the somewhat valuable lesson of such a failure.

In any case, you be you, I wouldn't want it any other way, and I am a bit mystified by that part of your response.
 

dl324

Joined Mar 30, 2015
10,003
I have a toolbox with many tools My job as a craftsman is to choose the tool most appropriate for a job. When the only tool you have in your toolbox is a hammer, you treat everything like it's a nail.

Microprocessors use the logic blocks that you claim are made obsolete by microprocessors. Someone needs to know how to design those blocks.

People who start using microprocessors, and frown on their less integrated relatives, may never fully understand how a microprocessor works or appreciate the skill it takes to design and manufacture one.

I've known microprocessor design engineers who referenced CMOS databooks for circuit ideas so learning the inner workings of MSI integrated circuits isn't wasted effort.
 

Thread Starter

Yaakov

Joined Jan 27, 2019
1,612
I have a toolbox with many tools My job as a craftsman is to choose the tool most appropriate for a job. When the only tool you have in your toolbox is a hammer, you treat everything like it's a nail.

Microprocessors use the logic blocks that you claim are made obsolete by microprocessors. Someone needs to know how to design those blocks.

People who start using microprocessors, and frown on their less integrated relatives, may never fully understand how a microprocessor works or appreciate the skill it takes to design and manufacture one.

I've known microprocessor design engineers who referenced CMOS databooks for circuit ideas so learning the inner workings of MSI integrated circuits isn't wasted effort.
This isn't really a response to what I said, though. I never said discrete logic was obsolete, or that people shouldn't know about it. I didn't say people shouldn't learn how to use domain specific ICs, or simpler components.

I narrowly said that when a person presents with a problem they want to solve and opaquely constrains the solution so that it can't be a microprocessor, I need to understand why that is before I feel I can help them.

Your reasons for knowing how to use simpler components are good, and if they say "I want to learn how discrete logic works" than that explains the constraint. But too often they don't have that rationale and instead stand your point on its head and say they don't have the time to learn micros, or it is too hard, yet they don't know enough to make that judgment or enough to solve the problem without the micro, so for me it seems the most help I can be is to make it clear to them they are wrong about that and get them on the right track.

It has always been my approach to learn complex subjects iteratively, starting with practical applications and learning fundamentals along with those practical successes. Photography is an example. There are many very successful photographers who are not very good at the technical side, but in their evolution as photographers, they've learned the fundamentals of exposure, focus, and the like once they saw the utility of knowing them.

In any case, please don't imagine I believe that discrete logic is obsolete, or that things like the 555 have no application because we have Pics. That's not it at all.
 

MrChips

Joined Oct 2, 2009
20,215
One of the "improvements" that irks me, for a long time the ARRL handbook featured a morse code practice oscillator that consisted of two transistors, a resistor and capacitor, which I still possess. I use it to test PNP and NPN transistors, condition of batteries, and if I need to, demonstrate morse code signalling. That has since been replaced with a 555-Timer circuit... aargh!
 
Top