How did kids in the old days learn without an oscilloscope?

[zapta]

just because in the old days people learned electronics without an oscilloscope does not mean you shoul ddo that now. yes it can be done , but why deny yourself the insigt that the avaialble machinery can give you ?

How did people get to work before there were cars ?
how did people eat soup before there were spoons ?
how did people stay warm before they developed the knowledge to make, maintain and transport fire ?

That's a red herring. I don't think anybody objected for using spoons or oscilloscopes.

[CraigHB]

I started with electronics about ten years of age in 1970.  I never even used a scope until I got into the electrical engineering school in college.  There's a lot you can learn with only a multi-meter and a breadboard.  For simple circuits that teach basic concepts a scope is not required.  However once you get into projects that perform some useful task it's hard to get by without one.  I used an analog scope for the longest time and finally got my first DSO a few years ago.  Now I wonder how I got along without it.  I think you just do the best you can with what you have.  There's a number of test instruments I wish I had but don't due to cost.  I manage to get by without them.

[David Hess]

Hands up if you are old enough to have used an oscilloscope with a 'stability' control!

I had one of those Heathkit oscilloscopes when I was like 12 but I did not really understand how to use it.

Two hands up if you *still* have an oscilloscope with a stability control and have used it!  (Tektronix 545)

When I was a senior in high school an uncle gave me a Tek 212.  it was later stolen and I was scopeless until this year when I found two different Tektronix 2200 series for $40 and $50 so now I use one and my son uses one.

SO: Uh, whose oscilloscope is that in the living room?
Me: Mine.  Tektronix 2232.  The oscilloscope I've always wanted and now I have it.  I rule!

[retrolefty]

Back then, as now, one can learn lots about electronics by starting in the DC domain. Just a simple multimeter can and does go a long way in learning and even building simple projects. After that a scope does help immensely in working in the AC/RF domain.

[rstofer]

It wasn't until the very late '60s that it became possible for hobbyists to experiment with digital systems.  I'm not sure when RTL was introduced to the hobbyist but I ran across it in 1969.  Prior to that time, flops were built up from transistors and there was some kind of limit to the types of projects hobbyists could build.  Perhaps a CPU was out of the question.

Once we started playing with logic, the requirement for tools increased (in my view) in that a logic probe was a minimal requirement and a scope wasn't far behind.  By '75 or so, microcomputers were becoming available at the hobby level and the speeds were so fast that a logic probe didn't provide much information.  Sure, a signal blipped but when?  About '79 or so, I built a floppy disk controller for my Altair 8800 using a Western Digital 1771 controller chip.  That project REQUIRED a scope!  And, in my view, digital projects have ever since.

Some time around '75, I built a Pong game from the original logic diagrams.  That project used 66 integrated circuits (more or less) and the wire-wrapping was considerable.  I think I got by with an old Dumont scope and a logic probe.  I'd have to wait 41 years to get my first DSO...

http://hackaday.com/2012/12/22/fabricating-hardware-from-the-original-arcade-pong-schematics/

Heathkit projects could be built without test equipment and, where testing was required, the device itself became the test equipment.  I think that was the case for the color television kit.  I never built one.  I just built an audio amplifier (late '50s) and the power supply, VTVM and VOM in the late '60s.

[tggzzz]

It wasn't until the very late '60s that it became possible for hobbyists to experiment with digital systems.  I'm not sure when RTL was introduced to the hobbyist but I ran across it in 1969.  Prior to that time, flops were built up from transistors and there was some kind of limit to the types of projects hobbyists could build.  Perhaps a CPU was out of the question.

Tell me about it I built a multivibrator 7-bit counter out of transistors, resistors and capacitors that I had liberated from random sources. It worked, and swiched my radio off after the relevant number of minutes.

Once we started playing with logic, the requirement for tools increased (in my view) in that a logic probe was a minimal requirement and a scope wasn't far behind.  By '75 or so, microcomputers were becoming available at the hobby level and the speeds were so fast that a logic probe didn't provide much information.  Sure, a signal blipped but when?  About '79 or so, I built a floppy disk controller for my Altair 8800 using a Western Digital 1771 controller chip.  That project REQUIRED a scope! 

in '76 I built a 6800 (inspired by the Altair 8080) with 128 bytes of RAM without a scope. The PCB construction technique was primitive (sticky-backed plastic used as etch resist!), but it worked. It was very successful in helping me get my first job.

I concur w.r.t. logic probes; they tell a signal changed, but so can an analogue multimeter

And, in my view, digital projects have ever since.

For checking signal integrity, yes. For checking embedded system program function printf() and LEDs are still very useful, and a multimeter can be used to determine things such as mean spare time in a processing loop.

Some time around '75, I built a Pong game from the original logic diagrams.  That project used 66 integrated circuits (more or less) and the wire-wrapping was considerable.  I think I got by with an old Dumont scope and a logic probe.  I'd have to wait 45 years to get my first DSO...

I built a glass TTY display controller using wirewrap, without a scope.

Heathkit projects could be built without test equipment and, where testing was required, the device itself became the test equipment.  I think that was the case for the color television kit.  I never built one.  I just built an audio amplifier (late '50s) and the power supply, VTVM and VOM in the late '60s.

Good example.

[rstofer]

For checking signal integrity, yes. For checking embedded system program function printf() and LEDs are still very useful, and a multimeter can be used to determine things such as mean spare time in a processing loop.

I use printf() {or equivalent} all the time.  The first thing I do on any microcontroller project is get the UART running.  Toss in a couple of binary to hex display functions and I'm pretty much good to go.

LEDs are useful as long as they are on long enough to see and at least one of my multimeters does %on time.  It's a matter of knowing what you have and knowing how to use it.

I have only had my DS1054Z for a few weeks but it makes all previous techniques obsolete (well, printf() is still useful).  Things are so much easier with a modern scope!  And I do know how I got along without it!

[med6753]

I was lucky. Before my Father joined IBM in 1955 he owned a TV repair shop. When I expressed an interest in electronics about 1965 or so he got the Heathkit O-10 scope and V-5 VTVM out of storage and fixed them up. So I had a scope right from the get go and no idea how to use it. But I soon learned.

I still have the Heath V-5 VTVM but the scope went to my younger brother.

[tggzzz]

Things are so much easier with a modern scope!  And I do know how I got along without it!

Just so.

My principal (and principle) objection is to people that think (and worse claim) that you absolutely must have the latest and greatest of everything.

[tggzzz]

I almost never used debugging features. My first thing to do is to get LED to blink, then UART. For some chips without float printf support, I will also port a float to char* function.

Back in 1981, I reused the compiler's printf() by replacing the putc() with something useful on that machine. Since it was an embedded Z80, the application didn't use floating point and it wasn't in the binary. I don't recall changing the printf() function, so I don't remember how the "%f" was omitted.

To measure CPU usage and interrupt performance, a scope with some IO toggling is enough. Profilers are just bloated marketing features IMHO.

Generic profilers have their places, but in shipping systems I've used custom performance profiling techniques to get other companies in trouble ("Here's what I sent you, and you took this long to reply; that's out of spec").

[CatalinaWOW]

just because in the old days people learned electronics without an oscilloscope does not mean you shoul ddo that now. yes it can be done , but why deny yourself the insigt that the avaialble machinery can give you ?

How did people get to work before there were cars ?
how did people eat soup before there were spoons ?
how did people stay warm before they developed the knowledge to make, maintain and transport fire ?

There is merit in learning to crawl before learning to walk, and to walk before learning to run.  An oscilloscope is a wonderful tool, but unless you understand why those wiggly lines are doing what they are doing it is useless.  Depending on your level of knowledge and skill the lines on the scope may represent some aspect of the real world, or they may be totally bogus garbage.  Even when the lines are an accurate representation of reality there is plenty of misdirection possible.  Is that noise an artifact of the measurement system (a real thing, but often not relevant to what you are trying to do) or is it actually coming from your circuit?  Once all that is sorted out, what do you do with what is left. 

Some folks can learn very well looking at the wiggly lines.  Other brains are constructed differently.  And some don't realize they don't know everything.  It is sometimes hard for those who have mastered a subject to understand the confusion and pitfalls for the beginner.  One advantage of learning without an oscilloscope (at least initially) is that it usually forces you to take smaller bites out of the problems.

[rstofer]

And then there is the CP/M incantation of DEBUG.  I remember many wasted hours single-stepping through somebody else's code.  What a colossal waste of time!

Core dumps are another nightmare especially if the underlying CPU lacked a stack and self-modifying code was 'normal'.  The IBM 1130 comes to mind...

How about single-stepping an FPGA CPU?  It was easier with a 32 bit logic analyzer.  Since Digilent has gone all-in on PMOD connectors and eliminated the large headers, this has been getting more difficult.  One thing a 32 bit analyzer does is make you wish for a 64 bit analyzer.  This is particularly true if the FSA is using one-hot encoding.  It takes a fair number of bits to figure out what state you're in!

Regrettably, printf() doesn't work well for FPGA projects.  Or it's just a whole lot more difficult!

[rstofer]

just because in the old days people learned electronics without an oscilloscope does not mean you shoul ddo that now. yes it can be done , but why deny yourself the insigt that the avaialble machinery can give you ?

Keep in mind the OP is in Malaysia. Asian wage standard is about 0.05x to 0.5x of western standard, depending on countries. When I was a kid, I dream of having my scope for years, and I didn't get mine till I was 19, when I got my Owon scope which I kept till this day. I didn't get a decent western brand scope till I turned 23.
Not sure how much money average Malaysians make, but in China, for a normal, common family, we make ~$1k per month per family, and in China you have to pay averagely ~$200k for an apartment, which means, for a young couple, even if their parents paid half of their house for them, they still have to work hard and save every penny in their first 20 years after marrying to pay for their house.
Spending $400 on a hobby toy is way to much for average Chinese young man, and I do not think in Malaysia this pressure is much low either.

Money is always an issue!  I kind of like the Maker movement where tools and facilities are shared among the members.

I also agree that you can get along without one and, if forced, that can go on forever.  There's always a way!

There's a lot of electronics manufacturing in Malaysia, there must be a market for used test equipment.  Singapore is another hub of electronics manufacturing with lots of small companies.

[eugenenine]

Hands up if you are old enough to have used an oscilloscope with a 'stability' control!

I had one of those Heathkit oscilloscopes when I was like 12 but I did not really understand how to use it.

Two hands up if you *still* have an oscilloscope with a stability control and have used it!  (Tektronix 545)

When I was a senior in high school an uncle gave me a Tek 212.  it was later stolen and I was scopeless until this year when I found two different Tektronix 2200 series for $40 and $50 so now I use one and my son uses one.

SO: Uh, whose oscilloscope is that in the living room?
Me: Mine.  Tektronix 2232.  The oscilloscope I've always wanted and now I have it.  I rule!

"his" scope is in the dining room.  I was going to sell the one when I got the better one but it seems its been claimed.

[mikeselectricstuff]

And then there is the CP/M incantation of DEBUG.  I remember many wasted hours single-stepping through somebody else's code.  What a colossal waste of time!

Core dumps are another nightmare especially if the underlying CPU lacked a stack and self-modifying code was 'normal'.  The IBM 1130 comes to mind...

How about single-stepping an FPGA CPU?  It was easier with a 32 bit logic analyzer.  Since Digilent has gone all-in on PMOD connectors and eliminated the large headers, this has been getting more difficult.  One thing a 32 bit analyzer does is make you wish for a 64 bit analyzer.  This is particularly true if the FSA is using one-hot encoding.  It takes a fair number of bits to figure out what state you're in!

Regrettably, printf() doesn't work well for FPGA projects.  Or it's just a whole lot more difficult!

Nice thing about FPGAs is you can build whatever test gear you need right on the chip!

[coppice]

And then there is the CP/M incantation of DEBUG.  I remember many wasted hours single-stepping through somebody else's code.  What a colossal waste of time!

Core dumps are another nightmare especially if the underlying CPU lacked a stack and self-modifying code was 'normal'.  The IBM 1130 comes to mind...

How about single-stepping an FPGA CPU?  It was easier with a 32 bit logic analyzer.  Since Digilent has gone all-in on PMOD connectors and eliminated the large headers, this has been getting more difficult.  One thing a 32 bit analyzer does is make you wish for a 64 bit analyzer.  This is particularly true if the FSA is using one-hot encoding.  It takes a fair number of bits to figure out what state you're in!

Regrettably, printf() doesn't work well for FPGA projects.  Or it's just a whole lot more difficult!

Nice thing about FPGAs is you can build whatever test gear you need right on the chip!

When you are trying to monitor something that's going on deep inside a chip, you really don't have any choice. 

[rstofer]

Nice thing about FPGAs is you can build whatever test gear you need right on the chip!

To some extent that is true.  The problem with trying to stuff a logic analyzer inside my CPU project was that I had already used up all of the BlockRAM.  I did set output flags that would indicate where I was in the FSA because encoding or decoding the state was a PITA.  I also carried 32 bits of Debug bus through most of the logic modules.  I just commented it out when I was done with a module.

The Integrated Logic Analyzer implemented in the Xilinx Vivado toolchain is another nice improvement.  I have only done a couple of demo projects with it but I look forward to using it for real some day soon.

I guess if it got down to sound card or nothing, I might reconsider my opposition.  It's better than nothing even if it is severely bandwidth limited, only AC coupled and poses some risk to the PC.  I suspect that a lot of people have junk PCs around and almost anything will be able to keep up with a sound card.

There isn't a lot of difference between a sampling scope and a logic analyzer.  Given that, I might take an FPGA and a ADC and build up some kind of scope.  Or maybe I would just use one of the ARM processors.  I could probably sample at multiple MHz so the bandwidth might not be too limiting.  There are a number of these projects on the Internet.

Maybe it's too much work and the sound card takes on a new appeal.

[eugenenine]

So if t your using an FPGA to make some test equipment and you make some test equipment inside the fpga that your using to make your test equipment in the first place is that sort of like when you use a vacuum to vacuum your vacuum and it implodes inside itsself?

[vk6zgo]

So I built a kit multimeter and a kit oscilloscope. I still use the same oscilloscope today.

Do you mean the scope, along with all electrolytic caps, as well as tube filament and phosphor, worked for 54+ years???

Not particularly uncommon---If 'scopes are for home use,they aren't left on 24/7 like many used in commercial service ,so the filament hours are reduced radically.
There is really nothing to go wrong with the phosphor,apart from "burn-in",& again that is unlikely in home use.

A lot of old gear has had the electrolytics & a few other parts changed throughout the years,but they are still substantially the same instrument.

[Howardlong]

As others have suggested, you can work out a hell of a lot with a moving coil multimeter, even a crappy 1k ohm per volt like I had. Supplement that with an LED and resistor, and you can debug a lot of digital stuff. Add a monostable to catch pulses (logic probe). We did a lot of make do and mend back in the 70s as teenagers.

With RF we used wavemeters with different coils for HF and very low VHF. For higher frequencies it was "tuning for maximum smoke".

I didn't get my own scope until I was in my mid 20s, although I'd begged and borrowed on the infrequent opportunities that turned up. Having very occasional access to a delayed timebase scope was a real eyeopener, but learning how to configure the CRO's dual timebase up properly I remember was a steep learning curve as a teenager. Even now I have to think a bit to get some of the more esoteric delayed timebase modes of a 2465B to work when I get it out, and that was my main daily driver scope for some years.

In real terms the cost of a decent scope has never been lower, but equally I'm also sure the functionality of scopes isn't even touched by their owners.

[vk6zgo]

Kids built projects out of magazines,
The better ones like "Radio Television & Hobbies" (formerly "Radio & Hobbies".later "Electronics Australia") gave circuit operation descriptions.
RTV & H also had a lot of basic theory articles.

We also voraciously read books --ARRL Handbooks,RSGB  Manuals,& so on.
Some authors of Electronics books had great names,like Fink,Chinn,Scroggie & other whimsical sounding monikers.

There was usually a older friend or relative into such things who would loan,or gift older versions.

We made Radios---- hobbyists commonly had a better understanding of Superheterodyne principles than some folks asking (& answering) questions on this forum.
Audio amplifiers were also common projects,back when HIFI was a "Geek" hobby.

Many people of that period had the ability to "visualise" the operation of circuitry---an ability which seems far less evident these days.

Electronics was BIG--enthusiasts usually found ,if not Apprenticeships,counrter & store jobs with  local Electronics shops.

To increase their knowledge,many attended Technical College night school classes.

My first "hands-on" experience with an Oscilloscope was  at around 13 when my brother brought a little 2" AWA 'scope home from work.
It was fascinating injecting random mains hum into the vertical input & watching the "wiggly lines".

Next time was at Perth Tech School,where they had  Cossor 'scopes.
When I got a real Technical job (at 22),I was amazed by the wondrous Tek 545 & similar,but didn't get to play with them.
That job fell through,& in the next we had  5" AWA & EMI 'scopes,with fairly poor bandwidth,& a 5MHZ Serviscope.
We hardly ever used them.

Then I was sent to a TV Transmitter site with a 545B which was used regularly.
What a nightmare learning how to use the Delayed timebase !

My first private 'scope was when I was about 29----just a collectors item,really,as it was a 2" Philips which looked almost identical to that little AWA of so many years before.

My first really useable  private 'scope was a BWD which I bought secondhand in the late 1980s.

[electrolust]

I came up after punch cards, but I learnt computer programming literally with pen and paper.  Whenever I could get at the actual computer I'd have to key everything in and see if it worked.  When you're young you have the luxury of time on your hands so it's doable.

It's not the same as "before the scope even existed" but the method is similar.  You start at first principles and take your time, work out the math, etc.  Not like we can do today which is just to be lazy on the basics, cobble something together and let the tools tell us where to go next.

[Audioguru]

Money is always an issue!

When I was a kid money grew on trees and was everywhere. Today there is also a lot of money around but I can't reach most of it.

[joeqsmith]

I was given my first oscilloscope when I was too young.  It was an old Dumont 100KHz tube unit.   I experimented more with my first VTVM than I did the scope because I had some idea how to read the meter.  I had made up some sort of light bulb brightness control that was sensitive to sound.  Louder would make it brighter and I could sort of get that but that squiggly line just didn't make sense to me.  I was playing around some with digital gates and had a little probe with LEDs in it that I used rather than the scope.     Later I bought a booklet from Radio shack (which I still have) that showed how to use the scope.   

I bet those $20 kits would far exceed what I started with! 

[ebclr]

We have oscilloscope in the old days