Solderless vs. soldering - what a difference (measuring 3V LDO)

[dusan]

I bought ME6209A30M3G from aliexpress and I got underwhelming results when I was testing in on breadboard, but on second measurement I got significantly different results so I soldered everything on perfboard and difference was huge. Here are both measurements, first is on breadboard, second is soldered on perfboard. Just an observation.

[ftg]

Images comparing how they were setup might also offer us clues.
But my guess would be the thinner wires and worse decoupling on the solderless breadboard.
Soldering the connections makes for more reliable and lower resistance connections.
Assuming there was no other difference in the parts used.

[tggzzz]

Yet another example of why solderless breadboards are a good way of misleading yourself[1] and discouraging beginners

Fortunately there are better prototyping techniques, technologies, and components: https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/ It looks like you know some of them, but there might be a few surprises there.

[1] another example: https://entertaininghacks.wordpress.com/2024/03/16/practical-traps-with-a-one-transistor-audio-amplifier-solderless-breadboards-and-oscilloscopes/

[shapirus]

Even with a solderless breadboard... Something must have been assembled VERY badly to cause such a big voltage droop at such a low current.

Solderless breadboards can be surprisingly well performing, if you know what you're doing.

[tggzzz]

Even with a solderless breadboard... Something must have been assembled VERY badly to cause such a big voltage droop at such a low current.

Solderless breadboards can be surprisingly well performing, if you know what you're doing.

A bit of oxide coating here, a leaf spring previously forced apart by a large diameter lead there, add a sprinkling of unnecessary L and C - and soon you have a problem.

The question is not whether solderless breadboards can be good enough, the question is what is the probability they will be good enough.

It takes a lot of experience to guesstimate that, and someone with such experience will probably choose to avoid needing to guesstimate and hope

[shapirus]

A bit of oxide coating here, a leaf spring previously forced apart by a large diameter lead there, add a sprinkling of unnecessary L and C - and soon you have a problem.

The question is not whether solderless breadboards can be good enough, the question is what is the probability they will be good enough.

It takes a lot of experience to guesstimate that, and someone with such experience will probably choose to avoid needing to guesstimate and hope

Yes. My point is that with some experience and understanding of where parasitics will matter most the probability of poor performance can be greatly reduced.

It would be interesting to see how the circuit was actually assembled.

[Peabody]

There is also a big difference in the quality of breadboards.

[tggzzz]

A bit of oxide coating here, a leaf spring previously forced apart by a large diameter lead there, add a sprinkling of unnecessary L and C - and soon you have a problem.

The question is not whether solderless breadboards can be good enough, the question is what is the probability they will be good enough.

It takes a lot of experience to guesstimate that, and someone with such experience will probably choose to avoid needing to guesstimate and hope

Yes. My point is that with some experience and understanding of where parasitics will matter most the probability of poor performance can be greatly reduced.

It would be interesting to see how the circuit was actually assembled.

Interesting, yes - especially if it shows sufficient detail to see dirty leads and the leaf springs.

I don't think the probability can be "greatly" reduced, but then I prefer numbers and dislike adjectives

I have no valid opinion as to whether the circuit was oscillating, but oscillations can cause apparently strange effects.

[tggzzz]

There is also a big difference in the quality of breadboards.

True. They range from new to old, and bad to awful 

[shapirus]

I have no valid opinion as to whether the circuit was oscillating, but oscillations can cause apparently strange effects.

Yes, that's another possibility, which isn't at all unlikely, and that could easily have such effects on the DC voltage measurement. Would be nice to see oscillograms of the output signal as well.

[tggzzz]

I have no valid opinion as to whether the circuit was oscillating, but oscillations can cause apparently strange effects.

Yes, that's another possibility, which isn't at all unlikely, and that could easily have such effects on the DC voltage measurement. Would be nice to see oscillograms of the output signal as well.

Not a 10MHz nor 50MHz scope, as illustrated in the second link I gave

[Peabody]

There is also a big difference in the quality of breadboards.

True. They range from new to old, and bad to awful 

I don't think that's right.  I've had very good performance from this board:

https://www.amazon.com/BB830-Solderless-Plug-BreadBoard-tie-Points/dp/B0040Z4QN8

But I've also had bad luck with other boards, particularly the $3 ones from the Far East.  They just don't make good connections.  I don't understand why that happens.  After all, you are sticking a piece of metal in between two other pieces of metal that are touching, and if I asked you to do that without making a good connection, you'd have trouble doing it.  But the cheap boards are able to accomplish the task.

[tggzzz]

We clearly have different definitions for "good" "bad" and "awful"

Adjectives conceal more than they enlighten Famously, Humpty Dumpty had something to say about the subject.

I dislike relying on anything that depends on what a buyer was able to find in the marketplace on that day. It is too easy for them to buy a good batch last week and a bad batch this week - and we couldn't tell before purchase.

There a lot of science and engineering in the design and manufacture of modern connectors. I remember many truly horrible ones when I was young. Solderless breadboards are a modern variant.

Fortunately solderless breadboards can be easily and cheaply avoided nowadays. Why settle for second third best?

[vk6zgo]

Yet another example of why solderless breadboards are a good way of misleading yourself[1] and discouraging beginners

Fortunately there are better prototyping techniques, technologies, and components: https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/ It looks like you know some of them, but there might be a few surprises there.

[1] another example: https://entertaininghacks.wordpress.com/2024/03/16/practical-traps-with-a-one-transistor-audio-amplifier-solderless-breadboards-and-oscilloscopes/

In the latter example, how come the transistor has no provision for forward bias of the base-emitter junction?
Looking at the simulation's input & output signal, it seems that the former very conveniently has its most negative excursion at around +2.0 volts, neatly disposing of the bias problem, as it seems did the "real world" generator they used.

In most people's "real life" the waveform will either be symmetrical around zero volts, or the most negative excursion would be to zero volts.
In both cases, the substantial distortion of the output waveform would be far more obvious than anything as subtle as oscillation.

[tggzzz]

Yet another example of why solderless breadboards are a good way of misleading yourself[1] and discouraging beginners

Fortunately there are better prototyping techniques, technologies, and components: https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/ It looks like you know some of them, but there might be a few surprises there.

[1] another example: https://entertaininghacks.wordpress.com/2024/03/16/practical-traps-with-a-one-transistor-audio-amplifier-solderless-breadboards-and-oscilloscopes/

In the latter example, how come the transistor has no provision for forward bias of the base-emitter junction?
Looking at the simulation's input & output signal, it seems that the former very conveniently has its most negative excursion at around +2.0 volts, neatly disposing of the bias problem, as it seems did the "real world" generator they used.

In general it is good practice to demonstrate effects using the simplest possible circuit/code/mechanics/etc. That enables people to concentrate on the effect, and not be distracted by extraneous components.

What is regarded as 0V is purely arbitrary. Your point evaporates if a +-7.5V symmetrical supply is used.

The TAoE x-chapters example noted didn't even have a generator: they simply applied a constant voltage from a PSU to the base.

In most people's "real life" the waveform will either be symmetrical around zero volts, or the most negative excursion would be to zero volts.
In both cases, the substantial distortion of the output waveform would be far more obvious than anything as subtle as oscillation.

You are missing the point.

Change the +15V supply to +-7.5V: the surprising misbehaviour will still occur, and there won't be any of the distortion you imagine!

[vk6zgo]

Yet another example of why solderless breadboards are a good way of misleading yourself[1] and discouraging beginners

Fortunately there are better prototyping techniques, technologies, and components: https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/ It looks like you know some of them, but there might be a few surprises there.

[1] another example: https://entertaininghacks.wordpress.com/2024/03/16/practical-traps-with-a-one-transistor-audio-amplifier-solderless-breadboards-and-oscilloscopes/

In the latter example, how come the transistor has no provision for forward bias of the base-emitter junction?
Looking at the simulation's input & output signal, it seems that the former very conveniently has its most negative excursion at around +2.0 volts, neatly disposing of the bias problem, as it seems did the "real world" generator they used.

In general it is good practice to demonstrate effects using the simplest possible circuit/code/mechanics/etc. That enables people to concentrate on the effect, and not be distracted by extraneous components.

What is regarded as 0V is purely arbitrary. Your point evaporates if a +-7.5V symmetrical supply is used.

The TAoE x-chapters example noted didn't even have a generator: they simply applied a constant voltage from a PSU to the base.

In most people's "real life" the waveform will either be symmetrical around zero volts, or the most negative excursion would be to zero volts.
In both cases, the substantial distortion of the output waveform would be far more obvious than anything as subtle as oscillation.

You are missing the point.

Change the +15V supply to +-7.5V: the surprising misbehaviour will still occur, and there won't be any of the distortion you imagine!

Indeed, but that is not the situation presented in the simulator's schematic, & in itself, introduces a greater complication to the situation than a couple of resistors.

[ArdWar]

A lot of my problems with breadboard usually boils down to cheap one having ridiculous amount of burrs, with the bad contact either from the burr itself or the eventual weak contact pressure after I yank the previous component getting caught in it.

That, and the use of soft non spring steel on those $3 boards. Not even nickel plated.

[tggzzz]

Yet another example of why solderless breadboards are a good way of misleading yourself[1] and discouraging beginners

Fortunately there are better prototyping techniques, technologies, and components: https://entertaininghacks.wordpress.com/2020/07/22/prototyping-circuits-easy-cheap-fast-reliable-techniques/ It looks like you know some of them, but there might be a few surprises there.

[1] another example: https://entertaininghacks.wordpress.com/2024/03/16/practical-traps-with-a-one-transistor-audio-amplifier-solderless-breadboards-and-oscilloscopes/

In the latter example, how come the transistor has no provision for forward bias of the base-emitter junction?
Looking at the simulation's input & output signal, it seems that the former very conveniently has its most negative excursion at around +2.0 volts, neatly disposing of the bias problem, as it seems did the "real world" generator they used.

In general it is good practice to demonstrate effects using the simplest possible circuit/code/mechanics/etc. That enables people to concentrate on the effect, and not be distracted by extraneous components.

What is regarded as 0V is purely arbitrary. Your point evaporates if a +-7.5V symmetrical supply is used.

The TAoE x-chapters example noted didn't even have a generator: they simply applied a constant voltage from a PSU to the base.

In most people's "real life" the waveform will either be symmetrical around zero volts, or the most negative excursion would be to zero volts.
In both cases, the substantial distortion of the output waveform would be far more obvious than anything as subtle as oscillation.

You are missing the point.

Change the +15V supply to +-7.5V: the surprising misbehaviour will still occur, and there won't be any of the distortion you imagine!

Indeed, but that is not the situation presented in the simulator's schematic, & in itself, introduces a greater complication to the situation than a couple of resistors.

It isn't presented because it is neither relevant nor important.

This input symmetrical-about-zero should satisfy you; I can't be bothered to duplicate the simpler simulation.

[tggzzz]

A lot of my problems with breadboard usually boils down to cheap one having ridiculous amount of burrs, with the bad contact either from the burr itself or the eventual weak contact pressure after I yank the previous component getting caught in it.

That, and the use of soft non spring steel on those $3 boards. Not even nickel plated.

... and then there are the oxide layers on the component leads, the parasitic capactances and inductances.

Too much hassle when there are easy cheap alternatives.

[shapirus]

Too much hassle when there are easy cheap alternatives.

Cheap yes, easy no.

Anything that involves soldering is not even remotely comparable to solderless breadboards in terms of easiness for R&D. Nothing beats solderless breadboards for trying different circuits and different part values in hardware quickly and with zero waste.

It's only when you are settled on a specific circuit and component values that proceeding to soldering to build a working prototype makes sense, or, another scenario, when parasitics of the solderless breadboard make it impossible or too unreasonable to use in a specific case.

The world of electronics isn't limited by HF and fast digital circuits. And even for those solderless breadboards can be quite usable, as long as you know what you're doing.

[tggzzz]

Too much hassle when there are easy cheap alternatives.

Cheap yes, easy no.

Anything that involves soldering is not even remotely comparable to solderless breadboards in terms of easiness for R&D. Nothing beats solderless breadboards for trying different circuits and different part values in hardware quickly and with zero waste.

I disagree, for the reasons I have stated - especially wasted time when (predictably) solderless breadboards don't perform as expected.

Beginners are going to have to learn soldering sometime. "Practicing" while making prototypes is a great way of enhancing their skills before they need to do it on a "real" PCB.

The "zero waste" concept is a red herring. Components soldered on breadboards can be reused: desolder them and put them back in the drawer. I have such components dating from the 1970s! Seriously But thanks; I've updated the page to trap out the "waste" contention.

It's only when you are settled on a specific circuit and component values that proceeding to soldering to build a working prototype makes sense, or, another scenario, when parasitics of the solderless breadboard make it impossible or too unreasonable to use in a specific case.

Many acknowledged experts jump straight to soldering, well documented examples being Jim Williams, Bob Pease. They do that for a reason. Even experts such as Horowitz and Hill can be caught out.

It requires a lot of experience to predict when parasitics cannot be a problem.

The world of electronics isn't limited by HF and fast digital circuits. And even for those solderless breadboards can be quite usable, as long as you know what you're doing.

Beginners - by definition - don't know what they are doing.

"LF" prototypes sometimes exhibit unexpected "VHF" behaviour, with baffling consequences. I showed one very simple example, another traditional example involves LDOs.

Overall with solderless breadboards it is too easy to get a circuit that appears to be working, but on closer examination exhibits weird or intermittent behaviour. I hate intermittent behaviour, and like predictable behaviour. Don't you?

[shapirus]

Overall with solderless breadboards it is too easy to get a circuit that appears to be working, but on closer examination exhibits weird or intermittent behaviour. I hate intermittent behaviour, and like predictable behaviour. Don't you?

I am fine with it when I understand what may be causing it so that I can safely ignore it and concentrate on what I am testing.

I use solderless breadboards to test isolated subcircuits to verify the results I get in simulation and adjust the values of passive components. Also to test things that are hard to simulate, such as power-on transients and state assertions.

So far this has worked in almost 100% of cases. Rare exceptions are things like high(ish) frequency switching converters: for example, a SEPIC converter based on MT3608 (1.2 MHz switching frequency) failed to work almost completely -- that's one of the few cases where I had to etch an actual PCB to test things (maybe a perfboard would have worked too, I'm not sure).

Otherwise, I've had very little issues with everything from DC to elaborate high-speed discrete logic (74LVC) circuits. But I'm not into RF or microwave. That may be a different story, but it's known that some people were able to buld oscillators in the multi-GHz range on solderless breadboards: https://www.eevblog.com/forum/projects/challenge-thread-the-fastest-breadboard-oscillator-on-the-mudball/
People even build entire computers on solderless breadboards: https://hackaday.com/2021/12/18/a-well-documented-breadboard-computer-shows-dedication/

There's no reason to blame the tool. Solderless breadboards work just fine to verify concepts for me and many others. Their parasitics and intermittent contacts have rarely been a problem for me and I think they are usually greatly exaggerated. YMMV. Every tool has its pros and cons and knowing them allows to use the tool most efficiently. Speaking of beginners, being a beginner doesn't change much. You do something, you encounter a problem, ask yourself why it happens, find an answer, learn and make your tools more useful for you.

[joeqsmith]

Same old song and dance

A reminder from four years ago
https://www.eevblog.com/forum/projects/challenge-thread-the-fastest-breadboard-oscillator-on-the-mudball/

More recently I was showing a member how to measure an inductor's self resonance and discuss some of the problems
https://www.eevblog.com/forum/beginners/measuring-the-self-resonant-frequency-of-an-inductor/msg5122983/#msg5122983

[Peabody]

As a hobbyist I don't have any problems with breadboards so long as I use quality boards.  But the ever-growing problem is parts that only come in hobbyist-unfriendly packages.  There are little adaptor boards available for things like SOT-to-SIP which can be used with breadboards, but there's not much you can do with a QFN or BGA.  And it seems most of TI's cool new stuff is really too small to solder to an adaptor board.  And even if you lay out a board for a project and order it from the Far East, you still need a stencil, unexpired paste, and hotplate.  Or get the board house to assemble it for you.  But that's all time and money.

[zapta]

When it comes to solderless breadboards, I trust Ben Eater.

https://shop.eater.net/products/bb830-breadboard