Capacitors: 68nF ok instead of 100nF?

lbenson

Senior Member
I've run out of 100nF capacitors, but by chance have a stash of small values of which the highest is 68nF. Are they suitable to use where where a 100nF capacitor would have been, adjacent to the picaxe power pins?
 

hippy

Technical Support
Staff member
68nF is probably good enough, and better than none, though not using any can often be fine in many cases.

You can use two in parallel to get a combined 136nF. Paralleling 68nF plus 33nF is pretty much 100nF. The cases where a specific value is essential are quite limited.
 

lbenson

Senior Member
I'll swap you some for a couple of toilet rolls.
Shipping across the Atlantic might be more expensive for the bulk of TP rolls than for little 100nFs. I probably have some somewhere in one of my 4 workspaces, meanwhile 68nF it is.
 

Eng460

Well-known member
The mention here by Hippy of using two capacitors in parallel when the right one is not available raises an interesting point that I have been pondering for some time.

The recommended power supply with the 7805 regulator has two capacitors in parallel on both the input and output sides. One is only 1/1000 of the other, so adding them in the normal way for parallel capacitors makes only 1/10 of 1% difference to components that are manufactured with a 10% tolerance. So what is the reason for the two? It is unlikely that the required value is that precisely known.

I suspect the reason lies in the impedance to alternating voltages which is affected by frequency, so for a sharp spike which is similar to a wide range of frequencies, the smaller one would behave differently to the larger one, it I have difficulty in getting my head around how this works. I can draw a vector diagram for the impedance on one capacitor, but can’t quite make the jump to two in parallel without just combining them first.

Is someone able to explain how they work a bit more clearly?

Eng460
 

datasmith

Well-known member
I think it has to do with frequency response. The small capacity value responds to/absorbs high frequency voltage spikes, while the larger capacity responds to/levels out lower frequency ripples. Thus they represent high and low frequency noise filters on a DC voltage. Well, that's somebody told me.😀
 

inglewoodpete

Senior Member
I think it has to do with frequency response. The small capacity value responds to/absorbs high frequency voltage spikes, while the larger capacity responds to/levels out lower frequency ripples. Thus they represent high and low frequency noise filters on a DC voltage. Well, that's somebody told me.😀
Correct. The small capacitor absorbs the high energy, low duration spikes, often caused by digital switching (both inside and outside the nearby chip). The larger one acts more like a storage "tank" of energy to handle surges in demand.
 

premelec

Senior Member
FWIW - every component has some resistance, capacitance, inductance - and different sizes, types and composition have different characteristics at various currents. and frequencies... What datasmith says is basically true. Empirically we use what works with a particular IC - I usually use a ceramic low value paralleled with an electrolytic much higher value... For some circuits the capacitor ESR is of much importance...
 

AllyCat

Senior Member
Hi,

A capacitor basically consists of two sheets of conducting material separated by an insulator. To make a typical capacitor physically small, the insulator may be a thin sheet of plastic (film) and the conductors aluminium foil or even metallised onto the plastic. The capacitor still will have quite a large area so it is then "rolled up" into a cyclindrical shape. The problem with coiling up a conductor is that's also the way you make an inductor! So a relatively large capacitance will have a significant "parasitic" inductance. There are other construction methods such as "ceramic" (typically for low capacitances) and "electrolytic" for high capacitances, but these have various disadvantages such as voltage limitations (including polarity) and poor accuracy, etc.. Even if a capacitor has an ultra low inductance, you may still need to consider the inductance of its leads or of the PCB tracks which connect it.

The problem with the inductance is that it forms a series resonant circuit with the capacitance, which is a high impedance resonance and means the capacitor effectively "vanishes" at that frequency. If the capacitance is large then so may be the inductance and the resonant frequency can be at a frequency which causes "issues" for the circuit, perhaps in the low MHz range. Thus a much smaller capacitor (typically ceramic) is put in parallel with the larger (usually electrolytic) capacitor to take over at the resonant (and higher) frequencies.

Particularly with switching circuits, there's then the potential problem that the smaller capacitor can become parallel resonant with the inductance of the larger. Fortunately, electrolytic capacitors are relatively "lossy" which can help to damp the resonance in the same way as a resistance in a tuned circuit lowers its "Q" (the Quality or sharpness of the resonance). Thus the types of capacitor can be important, not just their values.

Cheers, Alan.
 

Eng460

Well-known member
Thank you all for the most helpful replies. And particularly to Allycat for the extra detail which I very much appreciate. I had noticed some types described as non-inductive in my local suppliers catalogue, but did not appreciate what that meant before. I now have a much clearer picture of what is going on.

So from now on it a large electrolytic plus a small ceramic of the values specified.

I had been using tantalum for the smaller ones, after all more expensive should be better, right? I had noticed the polarity and believe that I connected them the correct way around, but recently one failed in a most spectacular manner. Glowed red hot, let the smoke out and other nearby components seemed to suffer as well, so all removed and thrown out. I was sure Inhad the polarity correct, but perhaps the result means I had it wrong. Anyway, twice shy on that one and it’s ceramics from now on, thank you.

I also like the neat look of the polyester type, and they seem to have plenty of voltage rating. Are they also ok?

Eng460
 

premelec

Senior Member
Tantalums don't take reverse polarity well and sometimes aren't marked clearly... sometimes I go for the looks - color, size etc... polyester fine for caps and clothes... ;-0 what ever you like... BTW ceramic capacitors can blow up artistically with purple smoke... at higher voltages and frequency in my experience and I have seen very pretty melted polyester types - sort of spill their melted guts... [all this at high voltage/frequency]. Many foil types are 'self-healing' meaning that if you get a high voltage foil puncture it vaporizes the foil at the fault...
 

Buzby

Senior Member
... The recommended power supply with the 7805 regulator has two capacitors in parallel on both the input and output sides. One is only 1/1000 of the other, so adding them in the normal way for parallel capacitors makes only 1/10 of 1% difference ...
I've mentioned this before, but it won't harm to say it again ...

Years ago I built a gadget using CMOS counter chips, and a 7805 ( with no capacitors ) provided the supply. The circuit didn't work.

Lots of probing with a multimeter, checking connections, etc. did not reveal why my counters wern't counting.

Eventually I got my scope on it, and was astonished to see that the 5v supply had a huge 2MHz oscillation on it, almost rail-to-rail.

Breaking tracks to locate the source showed that the 7805 was the culprit.

Put 100nF caps either side of the 7805, and the circuit sprang into life.

Lesson learned !.

Now I always put a 100nF on the output of any power regulator chip.

Cheers,

Buzby
 

AllyCat

Senior Member
Hi,
I also like the neat look of the polyester type, and they seem to have plenty of voltage rating. Are they also ok?
Yes, Polyester is one of several plastic dielectrics used in the "typical" style of capacitor I described above. Other names/types are Mylar, Polycarbonate, PTFE and Polyethelyne, etc.. IIRC Polycarbonate was used in some of the "better" (industrial) capacitors and Polyethylene has very low dielectric losses so is typically used at GHz frequencies. However, as I described above, they may have a rather higher inductance, particularly if the connecting wires are simply slipped into the winding whilst the capacitor is being "rolled up".

I think self-healing is probably limited to metallised foil types, which may have a rather higher series resistance because the aluminium layer is so thin. For high current and low inductance types, two strips of aluminium foil may be slightly offset (sideways) before being wound between dielectric layers. After winding, one conductor protrudes at one end of the "cylinder" and the other conductor at the other end. Then, the conductors can be bunched together (with the wire lead) at each end to give very low inductance and low series resistance. That's why many "mains" (line) suppressor and "Power Factor (correction)" capacitors are still cylindrical with the wires protruding at the centre of each end. Also, for mains capacitors in many parts of the World (230+ volts ac) don't forget they must resist the peak ac of around 350 volts, plus some additional spikes, so are rated to at least 400 and sometimes 630 volts.

Cheers, Alan.
 
Last edited:
Top