The need for decoupling

neiltechspec

Senior Member
So, yet another problem has been resolved today by the addition of a decoupling capacitor. (Not me !).

It's been seen so many times now.

So why don't Rev Ed put the need for Decoupling Capacitors in big Red Letters at the start of all the manuals ??????.

Seems an obvious omission to me.

I always include them on anything I build, be it analogue or digital - even long before I discovered PICAXE, when I started using 74 series & CMOS4000 series chips I fitted them.

Neil
 

nick12ab

Senior Member
Totally agree, the importance of decoupling capacitors cannot really be understated!

Additionally they should be mentioned in EVERY piece of troubleshooting text that Rev-Ed has written, including the one that is shown when a PICAXE download fails.

Here is a picture of a non-Rev-Ed PICAXE-08M project board that has no decoupling capacitor. Surprise surprise, some of the large quantity of these boards were not working for no apparent reason until a decoupling capacitor was added. Rev-Ed's minimum circuits show decoupling capacitors but some people seem to think that they can get away without.
 

Attachments

AllyCat

Senior Member
Hi,

Yes I agree too, except with:

Rev-Ed's minimum circuits show decoupling capacitors...
In trying to "justify" adding decoupling capacitors in this recent thread, I struggled to find an example in my User Manuals.

Admitted, the decoupling is primarily required and determined by the output components, but surely decoupling should be shown on the minimum circuit diagrams on pages 27, 28, 31 , 35, etc. of Manual 1 ?

Cheers, Alan.
 

Dippy

Moderator
The requirement for decoupling/bypass is very much down to the power supply quality/noise the device 'sees' and the amount of noise it might 'send' into the supply line or I/O lines.

Microchip PIC datasheets (not all esp not old ones) make mention of the neccessity wrt intosc and threshold behaviour.
I agree that it really should be considered a minimum spec though maybe when Rev-Ed designed the early stuff they found that many simple educational circuits would run without them.
I daresay many Rev-Ed boards were designed a long time ago before Microchip stressed it in their own datasheets.
 

jims

Senior Member
Picaxe Manual 1 eludes to this on page 26. I agree that it should have more emphasis and spell-out when/where to use them along with capacitor values and the type of capacitors to use (ceramic, electrolytic), etc. Folks seem to get caught in this situation too often. Jims
 

oracacle

Senior Member
Totally agree, the importance of decoupling capacitors cannot really be understated!

Additionally they should be mentioned in EVERY piece of troubleshooting text that Rev-Ed has written, including the one that is shown when a PICAXE download fails.

Here is a picture of a non-Rev-Ed PICAXE-08M project board that has no decoupling capacitor. Surprise surprise, some of the large quantity of these boards were not working for no apparent reason until a decoupling capacitor was added. Rev-Ed's minimum circuits show decoupling capacitors but some people seem to think that they can get away without.
the reason people omit it, is, they do not understand why it is there. alot of the boards are developed by people who see a way of making money, with less electronic knowledge than I have, and i dont have very much of it.
My rail had capacitors just for decoupling, it had lage elctrolitics on both sides of the switching regulator, 100uF acros the picaxe power rails, 0.1 uf ceramic (iirc) across the power pins of all the other ICs plus the ones one the motor driver board.
 

John West

Senior Member
My rail had capacitors just for decoupling, it had lage elctrolitics on both sides of the switching regulator, 100uF acros the picaxe power rails, 0.1 uf ceramic (iirc) across the power pins of all the other ICs plus the ones one the motor driver board.
"When in doubt, spread capacitors about."
- JW
 

oracacle

Senior Member
it doesnt seem to do much harm, but i had presumed on it being a noisy circuit with the swuitching regulator and micro stepped bi-polar stepper motor.
the nice thing is, i got a little curious and put the PCB scope on it, it showed and absolutly flat 5.1v with not ripple with with the motor under load, i was chuffed and impressed at the same time
 

nick12ab

Senior Member
the nice thing is, i got a little curious and put the PCB scope on it, it showed and absolutly flat 5.1v with not ripple with with the motor under load, i was chuffed and impressed at the same time
According to EEVBlog Dave, oscilloscopes like that are "toys".

There could be high frequency interference on the power rails in a decoupling capacitor-less circuit that is beyond the bandwidth limitations of the PCB scope.
 

oracacle

Senior Member
yeh that guys pi$$es me off no end, his voice is the sound of a condesending twit.

the scope will come in handy later when i come to RPM from HT lead project i am thinking about, dont worry i will be sure to decouple when i build the circuit
 

jims

Senior Member
I just noticed that the Picaxe manual #1 (online version) has been updated since I printed my "hard copy" about 1 1/2 years ago. It now shows 100nF caps on the minimum circuit diagrams on pages 27, 28, 31 , 35, etc. Not really sure when the update occurred. Good job Rev Ed. Jims
 

womai

Senior Member
According to EEVBlog Dave, oscilloscopes like that are "toys".

There could be high frequency interference on the power rails in a decoupling capacitor-less circuit that is beyond the bandwidth limitations of the PCB scope.
Always depends on the application. A motor typically produces major spikes somewhere in the lower Hz to the lower kHz region. The little PCB scope will be able to show these (bandwidth ~300 kHz). Of course there are things it can't do (e.g. show radio frequency interference); but also consider the fact that the "non-toy" scopes (according to Dave) cost approximately 100 times more - yes for two orders of magnitude more $$$ I'd would expect them to do more :=) (and even these scope have plenty of areas where they are not sufficient - in my previous job we needed scopes to go into the 20+ GHz region, and such instruments again cost ~100 times as much. A 100 MHz bandwidth scope is no more than a toy in those areas!).
 

neiltechspec

Senior Member
The online Manual 1, the only one I checked, does indeed show a capacitor. Shame it's all the way down starting at page 27.

Not everybody uses the online manuals (me included).

Wouldn't have been an big issue to have added a note nearer the beginning I would have thought.

After all, how many actually read that far down the likes of manual 1 - I'm guilty of that myself. I only went as far as the download cct & pinouts.

Like I originally said, I always fit them anyway.

Neil.
 

AllyCat

Senior Member
Hi,

"When in doubt, spread capacitors about."
Hmm, whilst there are occasions when that's appropriate (e.g. a PCB packed with TTL or memory chips), IMHO it's not really relevant for a PICaxe, maybe "scaring" some novices and even producing "bad" layouts. Firstly, this thread was concerning not using ANY decoupling capacitors; fitting almost ANY SINGLE capacitor should "tame" a PICaxe. I'd suggest a 100 nF ceramic for small designs, 10 uF normally, and perhaps 100+ uF if the chip is directly driving signifcant loads (e.g. LEDs) and the PSU is only "basic" (e.g. primary cells).

However, if the load is "nasty" (a motor, solenoid, etc.) then certainly choose a suitable additional capacitor (or more) and mount it close to the load (not the PICaxe). Then use appropriate power supply arrangements: perhaps individual supplies, or at least "star" connections and/or isolation/filter components (a resistor or inductor) in the supply rail to the PICaxe.

Probably not relevant to PICaxe, but in more "advanced" circuits, the connection of several capacitors in parallel can cause major problems. Some capacitors have significant series inductance (and ALL connections, including PCB tracks, have some inductance) so multiple capacitors can create series or parallel resonant circuits capable of radiating interference, perhaps in the 100 MHz - GHz region (i.e. radio communication bands). In these cases a single tantalum capacitor may be preferable, or we often used an electrolytic in parallel with a ceramic (or foil) type, because (aluminium) electrolytics are sufficiently "lossy" at high frequencies (i.e. low Q) to damp out most HF resonances.

Cheers, Alan.
 

john2051

New Member
Regarding decoupling. When I was studying at college, it was hammered into us that each 74 series ttl or 4000 series cmos ics needed a 0.1uF as close as poss to the
power pins. In addition, a 10uF tantalum for I think for every four ics as well as a bulk capacitor on the main power rail(s). The number of commercial products that as people have mentioned decoupling seem optional. On some boards you can actually see the holes where they should be, but in an effort to save a few tenths of a penny, they are left out!
I find you can sometimes tell how much decoupling or filtering a module has by its level of interference to a radio in close proximity. Although a scope comes in handy here.
Sometime ago you could buy dil sockets with caps built in. Great so long as the power pins are in the right place.
Maybe it would help newcomers to picaxe to stress the importance of these caps..
Regards john
 
Top