Bypass caps 'Hands on demo' help ...

Go To Last Post
41 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So, I've been an on-off-on electronics hobbyist 35'ish years now ... Started using breadboards pretty early ... Never got into 'proper' hardware or PCB design, and after getting into microcontrollers at eng. school I've always been more interested in the programming bit than the hardware ... But I've still got a decent grasp of electronics ...

 

A couple of comments after posting a picture of my breadboard got me really interested in bypass caps for IC:s ... 

 

I don't doubt for a second that they're there for a reason ! I've seen them on every commercial product I've ever opened up ... Read the AppNotes and documents and everything makes perfect sense ... 

But I've never used on my breadboard experiments ... And well, so far so good you might say?
I'ts easy to see using a scope that adding caps 'cleans thing up', but I've yet to see/hear/smell a circuit spiral completely out of control due to the lack of bypass caps ... 

So, if You were an Electronics teacher and I was the annoying kid always asking 'why', how would you demonstrate the need for bypass caps in an AVR based circuit ? 

 

Picture of my current 'bypass capless' setup :
ISP (AvrISP MKII) connected to a tiny85 via a 74hc4053 used as an "ISP isolator". Tiny85 as an i2c master using the USI controlling a PCF8574 with the Led Bar and a Maxim 519 D/A converter.

Running the i2c bus at 490kHz ! Pushing/testing the limits ... 

The other setup is a Mega168 reading a ADXL335 via a LM258 opamp. The mega then sends data to the serial port and over MIDI.

Everything's fed (5V to the onboard chips, and 3.3V to the ADXL335) from the USB port of my laptop through the UM232 module in the lower left corner there.


 

Last Edited: Mon. Jan 18, 2021 - 03:06 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

if You were an Electronics teacher and I was the annoying kid always asking 'why'

I would accidentally bump you when you were in a high place......like a 2nd storey window..... cheeky

John Samperi

Ampertronics Pty. Ltd.

https://www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, circuits do "spiral out of control" without bypass caps. You may not have experienced that, but there are plenty of posting on this list with problems that resolve to that solution.

 

My philosophy is "why piddle with the little stuff that will only help you"? My (remaining) life is too short to waste on problems that would have been solved by spending $0.25 (U.S. Coin of the Realm) a couple of times.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jussishow wrote:
I'ts easy to see using a scope that adding caps 'cleans thing up'

What further proof do you need than that!!!  smiley

 

Jim

 

(Possum Lodge oath) Quando omni flunkus, moritati.

"I thought growing old would take longer"

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk wrote:

jussishow wrote:
I'ts easy to see using a scope that adding caps 'cleans thing up'

What further proof do you need than that!!!  smiley

 

Jim

 

But 'cleaning things up' doesn't make the two AVR experiments run any better than they already do  ... 
I'm still waiting/looking for that EUREKA moment ! The 'Ok, That's why bypass caps are so important'-moment/discovery ! ;-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you can stand an Aussie shouting at you for half an hour: https://www.youtube.com/watch?v=...

 

His previous whiteboard session on the subject is worth a look too.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Try running the ADC and you will quickly see a behavior and performance difference.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

Last Edited: Thu. Jan 14, 2021 - 02:34 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Run your circuit in a noisy environment...be it EMF, or RF.  As West Coast Jim noted, try running the ADC and see what happens.

 

 

East Coast Jim

 

 

 

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"Step N is required before you can do step N+1!" - ka7ehk

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB, RSLogix user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've had trouble with systems that run just fine on my bench at 115,200 baud but out in the field are just shot through with errors until you get down to 9,600 or so, and even then error correction is a good thing.  Now, the 'field' was an airfield with a pretty honkin' big radar, but hey... 

 

Not much to do with decoupling - more about about cable shielding - but the point is that what works somewhere might not work everywhere. 

 

S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jgmdesign wrote:

Run your circuit in a noisy environment...be it EMF, or RF.  As West Coast Jim noted, try running the ADC and see what happens.

 

 

East Coast Jim

 

 

The Mega168 is reading two A/D channels. The signals from the accelerometer amplified by the LM258. Aref set using a simple voltage divider.

Works just fine ! Not 100% rock solid readings, and adding caps here would shurely smooth things out but then we’re talking about filtering/smoothing caps.

 

Wish I had access to a lab of some sort where I could introduce my gadgets to ‘electrical noise’ in a controlled manner ! :)

Last Edited: Thu. Jan 14, 2021 - 05:01 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jussishow wrote:

Wish I had access to a lab of some sort where I could introduce my gadgets to ‘electrical noise’ in a controlled manner ! :)

 

Stupid suggestion:  Build a spark gap.  Not hard - get a couple of transformers to pop the voltage up, and (gently!) adjust the gap for a nice continuous spark.  They make all sorts of fascinating broadband noise. And ozone.

 

In the USA, doing that will get you a nasty letter from the FCC, the government radio watchdog, telling you to stop it.  You might hear from your neighbors first.  Dunno about Norway.  S.

 

PS - Other fun ideas include lousy old fluorescent lights (get one that buzzes and flickers) or, as I found out, go loiter near a large radar installation.

edited to add:  And ozone.

Last Edited: Thu. Jan 14, 2021 - 05:21 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Lets say your circuit has no decoupling caps at all......

 

You start playing around with your circuits at a time when there is not much noise so everything works well.

You start doing a massive code update that you partially prepared earlier and are now going to do.

During that process you start the environmental noise ( laptop almost empty so connected to the mains..... you turn on a modern day dimmeble light that you bought of of ebay/alie/amazon delivered through a very cheap chineese supplier)

All of the sudden your circuit starts resetting, but as your start-up sequence has also gotten a major overhaul were will you think the problem is at first????

 

bet you will blame your code and not the lack of capacitors that keep your supply clean and in effect make sure the supply never drops below the minimum operating voltage.

when it then recovers you will get a reset as the reset line has been low ( according to the processor) or you had a brownout failure

 

As said we have seen more examples here of people spending days on debugging their code as 'everything used to work just fine, I only changed code' and it turned out they went from a quiet to a noisy environment, some even without knowing.... were in the morning every morning things worked just fine but in the afternoon things got strange........

 

Getting unexpected resets is just one of the things that can go wrong when not using proper decoupling caps.

ADC faulty readings is another, how about the processor skipping instructions, that is a good one to debug ;)

 

So decoupling caps are a minimum that is why all the freaks here hammer on them as they do solve tons of issues.

Make it standard to put at least a 100nF at every IC supply pin you have on the board, so not every IC but each supply pin..... and then one or two 10-100uf ( depends on drawn current) general buffering capacitors to handle real surges. make the 100nF ceramic ones and the other 2 are a do not care ( for now as you are not building a product that goes in mass production) .

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My demo would be to go to your car, lift the bonnet and desolder all the decoupling capacitors on the electronics that control your brakes, steering, and airbags.

#1 Hardware Problem? https://www.avrfreaks.net/forum/...

#2 Hardware Problem? Read AVR042.

#3 All grounds are not created equal

#4 Have you proved your chip is running at xxMHz?

#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Brian Fairchild wrote:
My demo would be to go to your car, lift the bonnet and desolder all the decoupling capacitors on the electronics that control your brakes, steering, and airbags.

Is that not how they invented the very first the self driving car??????

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Brian Fairchild wrote:
My demo would be to go to your car, lift the bonnet and desolder all the decoupling capacitors on the electronics that control your brakes, steering, and airbags.

You'd probably demonstrate/illustrate gravity by kickin' the students of a rooftop as well ! ;-)

 

 

Got a bunch of tubeamps, and some spare transformers ... I'll see if I can use them to generate enough some "RF/EMC noise" to cause some issues ! 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jussishow wrote:

You'd probably demonstrate/illustrate gravity by kickin' the students of a rooftop as well ! ;-)

 

Only those student who are gravity deniers.

#1 Hardware Problem? https://www.avrfreaks.net/forum/...

#2 Hardware Problem? Read AVR042.

#3 All grounds are not created equal

#4 Have you proved your chip is running at xxMHz?

#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Gravity IS a myth.  It's just that we're living on a really sucky planet.   cheeky  S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Brian Fairchild wrote:

jussishow wrote:

You'd probably demonstrate/illustrate gravity by kickin' the students of a rooftop as well ! ;-)

 

Only those student who are gravity deniers.

 

And anyone who asks you to explain gravity by the means of a demonstration/experiment is labelled a gravity denier and kicked of the rooftop ? ;-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You just throw yourself at the ground and miss.  S.
.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Scroungre wrote:

Gravity IS a myth.  It's just that we're living on a really sucky planet.   cheeky  S.

 

Alleged graffiti in the ISS:

 

Seated on this high-tech loo,

This bog that cost a billion bucks,

Gravity is a myth, my friends,

The ISS sucks...

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Scroungre wrote:

You just throw yourself at the ground and miss.  S.
.

Wasn't there a quote a bit like that in The Hitchikers Guide To The Galaxy or in a Discworld novel or ...? 
Learning to fly : Jump up and forget to fall down ? :-D

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hitchhiker's Guide, Douglas Adams.  S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jussishow wrote:
I'ts easy to see using a scope that adding caps 'cleans thing up', but I've yet to see/hear/smell a circuit spiral completely out of control due to the lack of bypass caps ... 
Well for one thing try doing anything with computer and audio. Any kind of noise/wobble on power lines becomes only too obvious to the very sensitive ear/brain combination built into most humans.

 

A more practical example is actually power caps in a satellite receiver/recorder we supplied about 2 million units of. Admittedly these were large electrolytics not just decoupling caps but what happens to electrolytics if someone underspecifies the part to buy and then they are used in a box that runs at 65C because of heat from PSU and hard drive is that the electrolyte eventually dries out and the caps then offer zero capacitance so it's as if they weren't there any more. The caps were actually smoothing the 5V rail to satellite front end tuner (actually two) and it had a HUGE sensitivity to noise (remember that the signal coming in from a satellite has about as much energy as a housefly farting on the moon!). Because the Vcc rail was no longer smoothed the noise from the PSU overwhelmed the actual received radio signal so that there was no possibility of a decoding lock. So the tuner simply stopped working. While the box could allow the view to watch on just one tuner. If they were operating the box as watch one, record another then the second tuner maybe would not get the signal or at least have tons of noise so that episode of Baywatch or whatever lost large chunks at all the important places. Slowly but surely 2 million boxes came back to us. As this usually involved a service engineer from the satellite broadcaster going out in a van at something like £50.£75 per call out this added up a bit ! All for the cost of a $0.10 capacitor.

So don't underestimate the need to pay attention to the details in a design. Of course a one off hobby project is perhaps a bit different to a commercial design with millions of units. Just like things such as over-clocking you can get away with a lot in a one-off home design that could be suicide in a commercial product range.

 

(a bit like driving a car without seat belts? ;-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:

jussishow wrote:
I'ts easy to see using a scope that adding caps 'cleans thing up', but I've yet to see/hear/smell a circuit spiral completely out of control due to the lack of bypass caps ... 
Well for one thing try doing anything with computer and audio. Any kind of noise/wobble on power lines becomes only too obvious to the very sensitive ear/brain combination built into most humans.

I'm a hobby musician, and I've got a bunch of homestudio gear and am "allergic" to noise/hiss when recording ... But that's another discussion altogether ...
And I can't really see what an 8 bit AVR would do in the signal path anyway ... Maybe in a LoFi Synth of some sort, where the noise would a part of the sound ... 

Also used to build my own guitar amps, so I've delt a lot with PS filtercaps, noisy transformers and such ...

 

And I've, quite recently, forgot to add filter caps when using 780x regulators, so I know from experience why those are essential ! ;-)
 

But back to the topic : These 100nF bypass caps that are essential at/as close as possible to every IC supply pin.
 
Just to clarify : I don't doubt the need for them, and I've read the AppNotes and I'm familiar with the theory behind them ... 

But ! I've now tried to make my breadboarded, 'bypass capless', AVR circuits crash/reset using some "improvised noise sources" that I found here ... A tablelamp with a dimmer (that emits an annoying Bzzzzzt at lower settings), cheap'ish bench supplies and I even dusted of one of my old amps with a Hammond 372BX power transformer ... But nope ... the AVR:s are chugging along as usual ...

Next plan is to setup some switching circuitry that, hopefully, will generate spikes/surges in the power ... I'm on a mission here ! :D
 

Measuring at pin 8 of the Tiny85  (the i2c master), and at pin 7 of the Mega168 (ADXL335 -> Serial port & MIDI CC "interface") shows quite a lot of noise ... but still no resets or crashes ... hmmm ...
The only caps on the board are the ones on the UM232R USB interface and a 2.2uF electrol. at the reset pin of the Mega168.
 


 

Last Edited: Thu. Jan 14, 2021 - 06:30 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

How many of these designs have been made into pcbs?

Let's not forget that breadboards have capacitance, because of the way they are built. I have run crystals with no caps with proto types on breadboards. Also the size of the busses in breadboards, the gauge of wire for all the jumpers, compared to the size of traces on pcb's. You are also getting your power from your laptop, so your supply is well requlated.

Try powering your circuit from one of those cheap $1 plug in 5v USB walwarts and see how it goes.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

On a PCB, crystal pin capacitance is typically a few pF. On a plug-in protoboard, it is typically more than 15pf and most crystal will run with that even if the capacitance is a bit high. 

 

That 15 or so pF has very little bypass effect on power supply pins where the desired C is around 65x larger than that.

 

I really do not understand this thread! For the cost of a few 1nf caps? Its hardly worth the effort to type into the thread let alone not do it on real-world prototypes.  I would FAR rather know what the real cause of problems might be instead of chasing a will-of-the-wisp, all because a bypass cap was forgotten,

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A 0.1uF is a very standard recommendation by most MCU companies to be placed close to the MCU and is a bare minimum in my opinion, it assumes a clean vcc supply input. I typically measure and use more than I need to ensure operating stability. A 10uF paired with a 0.1uF gives your MCU a stable supply rail, assuming your supply is reasonably clean to begin with. I always add in some extra bypass caps, Better to have a few extra bypass caps, than too few and regret it later.

Last Edited: Fri. Jan 15, 2021 - 05:20 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Noise source suggestion:  Hook up a stepper motor driver.*  A nice beefy one, like an LMD18201 (half-bridge), or built out of discrete TO-220 MOSFETs.  I had an AVR project driving around steppers (the motors were slightly smaller than beer-can size) and got spontaneous resets now and then.

 

That one did have lots of decoupling caps.  The motors ran off the same 5V as the AVR.  I'd brought up the power to the board to one connector, then made sure they were separate on the board, but that wasn't enough.  Cutting the connector in half and doubling up the wires to the P/S solved the problem.  It was radiated/reflected/coupled noise from the motors.  It was still the same power supply - and the same connector on the supply end - but using separate wires all the way back to the supply itself made all the difference.  S.

 

* Or, as you mentioned, something as simple as a big transistor switching something, preferably something low resistance and highly inductive (wire-wound resistor?  Throw at least an amp at it), off and on a lot.  S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I once had a stepper motor "light up an led" that was simply running through the same wire harness. Had to shunt the led leads with a resistor to prevent the motor current induction through the harness from lightning it up. Motor interface is no joke.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:

I really do not understand this thread! For the cost of a few 1nf caps? Its hardly worth the effort to type into the thread let alone not do it on real-world prototypes.  I would FAR rather know what the real cause of problems might be instead of chasing a will-of-the-wisp, all because a bypass cap was forgotten,

 

Jim


 

I’ts not about saving a penny or two, or ‘forgetting’ bypass caps ! (Or leaving them out intentionally for that matter...). I’ve got a box full of suitable caps here and I could happily add the to all the supply pins of all the IC:s and leave it at that ... 

But some of the comments sounded like ‘unless you add bypass caps your project will fail, period!, ... But ... they don’t !

 

And that’s where I am now ... Trying to make my circuits fail/restart and see the effect of the bypass caps with my own eyes ... Don’t see how that can be so hard to understand ?  ;-)
 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

12oclocker wrote:
A 0.1uF is a very standard recommendation by most MCU companies to be placed close to the MCU and is a bare minimum in my opinion, it assumes a clean vcc supply input. I typically measure and use more than I need to ensure operating stability. A 10uF paired with a 0.1uF gives your MCU a stable supply rail, assuming your supply is reasonably clean to begin with. I always add in some extra bypass caps, Better to have a few extra bypass caps, than too few and regret it later.

 

I did a lot of work with older AVRs in PLCC44 packages* (still have two on my desk right now, cheerily driving a lot of 7-segment displays!) and they always sprouted four 0.1uF caps, one roughly at each corner.  More get splattered about the rest of the board, depending upon what other chips and bits are loitering about.  S.

 

* Actually, most of the reason for this post is to whine at Microchip (and pretty much all chip vendors these days) about the death of the PLCC J-lead package.  You can surface-mount them!  You can socket them!  You can probe them without needing a microscope!  They have lots of pins!  They're nowhere near the colossal lump that a 40-pin 0.6" wide DIP is!  They're wonderful!  You can get them in 68 and 84 pins too!!  Now, if you want forty pins, either you have to get some giant slug of a chip - or solder it down and pray you don't brick it.  Grumph.  S.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'd bet that J-lead is more costly to manufacture.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I just priced the twelve caps necessary to decouple my most recent ARM M0+ design (10n/1u at each power pin): $0.04 - and that only because I only want a few. They're probably far more than are 'needed' but the same board has a $15 display bolted onto it - three times the cost of the rest of the hardware - and I don't want to have to throw them away when the board does weird stuff.

 

Every time a digital output changes state, there's a potential for a momentary short across the power rails. It's short, but if lots of outputs change state at the same time, it can glitch the supply sufficiently to trigger the reset, the brown-out detector, analogue input interrupts, digital interrupts... none of which are likely to have a wanted effect on the device operation. It's a no-brainer to fit the decoupling components.

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If I ever jump out of an airplane you can bet I'm going to be wearing a parachute.

 

Of course I've never seen any good, scientifically sound, well conducted study to show that their usage is beneficial.

 

But I think I'll take everyone's word for it, just to be safe. cheeky

 

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC wrote:

If I ever jump out of an airplane you can bet I'm going to be wearing a parachute.

 

Of course I've never seen any good, scientifically sound, well conducted study to show that their usage is beneficial.

 

But I think I'll take everyone's word for it, just to be safe. cheeky

 

JC

 

Ok ... Comparing jumping out of a plane with/without a parachute, or driving a car with/without a seatbelt which has also been mentioned here, to experimenting/toying around with IC:s on a breadboard (and the occasional strip-/veroboard) is just ... well .. childish ?

 

But for the sake of the discussion, I'll sink to that level for this reply ... :

If we limit the timeframe to my "AVR breadboarding days" which is roughly 15 years I've driven to the airport without wearing a seatbelt, got in the plane and jumped out without a parachute more times than I can count ... And I'm still here !

Remember the first couple of years were pretty intense, after that it's more like once/twice in a blue moon ... 
I'ts always been more about the software, so once I've had the hardware (AVR + some stuff I wanted to control/read from/measure...) up and running I've focused more on the code than the Hardware ...

 

The one I used the most is a Mega168 clocked by a 11.0592Mhz xtal. The Tiny85 and the Mega168 I use now run from the internal 8MHz oscillators.

I've hooked up RC Servos to (fed by a separate power ...), various i2c circuits, logic circuits, LCD displays (HD44780), tuned IR receivers to 'decode' remotes, used the A/D, PWM, UART e.t.c ...  

 

The times I've hurt myself in the landing or, landed without havin' a clue how I got there, there's always been a logical explanation to it ! Programming error, bad/wrong connection, using a PNP instead of NPN e.t.c ... 

 

But I've yet to end up with an experiment that just refuses to work, or starts self-oscillating and goes into an 'auto-destruct sequence' which then can be brought back to 'normal operation' by adding a bypass cap at the supply pins ! 

 

As I've said many times already, I don't doubt the need for them ! I'd just like to see it with my own eyes/tools here ! Just to satisfy my curiosity ! ;-)

Since they are so essential I was hoping there was a simple way to demonstrate it !  

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Just wondering what your thoughts are regarding the sphericality of the globe or the likelihood of NASA moon landings in the late 60's. Presumably this is a quasi-religious thing like covid existence, vaccine efficacy, etc, etc? Each to their own view I guess - ultimately it probably makes the world a richer place. :-)

jussishow wrote:
I've seen them on every commercial product I've ever opened up
that for me is the most telling comment on the entire thread. Would commercial entities, keen to save every last cent, perpetrate this myth if they didn't believe the scientific evidence?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:

Just wondering what your thoughts are regarding the sphericality of the globe or the likelihood of NASA moon landings in the late 60's. Presumably this is a quasi-religious thing like covid existence, vaccine efficacy, etc, etc? Each to their own view I guess - ultimately it probably makes the world a richer place. :-)

jussishow wrote:

I've seen them on every commercial product I've ever opened up

 

that for me is the most telling comment on the entire thread. Would commercial entities, keen to save every last cent, perpetrate this myth if they didn't believe the scientific evidence?

 

Have you read any of my comments ? 
Is my english really that bad ? 

If reply #30 doesn't explain it clear enough I really don't know what will ! 

https://www.avrfreaks.net/commen...

EDIT : And ... Grow up, will you ? ;-)

Last Edited: Sun. Jan 17, 2021 - 02:52 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You miss the point with your answer #30 when you say 'without these caps, projects will fail - they don't'.

 

This is not the case. Without these caps, projects *will* fail. Not all of them, and usually not all the time, but trust me - they do fail. Given enough projects and enough time, sooner or later they will fail and in often a way which is inconstant and particularly difficult to trace. I've had projects die because someone turned on an arc welder on the farm next door; I've had projects fail when the PSU design changed from linear to switched mode and vice versa.

 

You have just not yet been unlucky. In a similar vein, I have never lost data due to a disc crash - but it hasn't stopped me making backups.

 

You asked for an example: a 555 timer (the old non-cmos version, not the 7555) used to switch a high current inductive load - a relay or a motor - will cheerfully reset itself once it triggers rather than operating the timeout specified if it doesn't have enough capacitance across the rails and close to the chip. The reason is that both the transistors in the totem-pole output stage are momentarily on at the same time when the chip changes state; that short circuit can drop the power supply volts enough to reset the part. And the 555 is about as simple a chip as you can find... a microcontroller has dozens of totem-pole outputs.

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

barnacle wrote:

You miss the point with your answer #30 when you say 'without these caps, projects will fail - they don't'.

You misunderstood me there ... Should've been more clear ! With 'they don't' I ment that I've yet to experience it with one of my projects !  

The whole quote is : ‘unless you add bypass caps your project will fail, period!, ... But ... they don’t !

Quote:

This is not the case. Without these caps, projects *will* fail. Not all of them, and usually not all the time, but trust me - they do fail. Given enough projects and enough time, sooner or later they will fail and in often a way which is inconstant and particularly difficult to trace. I've had projects die because someone turned on an arc welder on the farm next door; I've had projects fail when the PSU design changed from linear to switched mode and vice versa.

I take it you didn't use these bypass caps on the ic:s on these projects then ?

 

 

Edit : Regarding disc-crashes : I'm not the only one in here who had one of the infamous IBM deskstar harddisks in the late 90's / early 2000's, am I ? Also known as Deathstars ? ;) And yes, I backup ! 
 

 

Last Edited: Sun. Jan 17, 2021 - 09:03 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Speaking of childish:

DocJC wrote:

If I ever jump out of an airplane you can bet I'm going to be wearing a parachute.

I HAVE jumped out of an airplane without a parachute.  They laughed at me, but I sure showed them!

(now, as it happens, the airplane was neatly parked at a gate, and I jumped out onto a very conveniently-located jetway, but that just shows excellent planning!  laugh)

 

More helpfully (perhaps?) to find a test that says "This circuit does not work without bypass caps - install them and it works, so there!" you might also try ramping up the clock frequency.  Dunno what you're running at, but it looks like an internal AVR oscillator.  Use an external clock and ramp it up as far as you can - then see if you can ramp it farther with decoupling caps.  S.

 

Edited because I re-read a previous post and found out what your clock speeds were and to fix 'without'.  S.

 

Last Edited: Mon. Jan 18, 2021 - 02:47 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

 

Ok, I think this has gone on far enough.  The OP is asking for someone to come up with a demonstration of the caps proving their need in a circuit, in spite of the datasheets recommending/requiring of the components.  The OP also does not seem to be interested in anyones comments on how the caps have helped them in the past.  The OP is looking for PROOF, or a demonstration thereof, of the caps actually saving the day.

 

THat being said, Unless one can provide a demonstration of the OP's requirements, there is no need to comment any further.  That includes parachute analogies et. al.

 

Cheerio!

Jim

Moderator.

 

 

EDIT:

Seems someone has become irritated so I am going to close the door on this subject

 

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"Step N is required before you can do step N+1!" - ka7ehk

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB, RSLogix user

Last Edited: Mon. Jan 18, 2021 - 03:05 PM
Topic locked