Hi. In this circuit, why is C3 nonpolarized? I read that polarized is more compact and powerful and so preferred for DC. I don't get C3. And where is ground? Shouldn't there be a ground at COM?

 

At high frequencies (~ 100 kHz), the impedance of polar capacitors (aluminum) increases. To compensate for this harmful effect, a ceramic capacitor is included in parallel.

 

I am not sure you can find an electrolytic capacitor with such a low value.As Bordodynov pointed out ceramic capacitors are much better at higher frequencies, and in the case of the 7805, it might be size and cost that were factors in choosing that particular capacitor.

Some kind of a ground symbol connected to common would look nice, but it is possible that the designer intended that common to connect to something other than ground or be left floating.

 

Hi. In this circuit, why is C3 nonpolarized? I read that polarized is more compact and powerful and so preferred for DC. I don't get C3. And where is ground? Shouldn't there be a ground at COM?

Hi,

This has been this way as long as i can remember back into the early 1970's. Back then we were dealing a lot with standard 7400 series TTL (no LS yet for example) and those things put massive spikes on the 5v power supply lines so the first thing that went into every design was a somewhat large electrolytic in parallel with a small ceramic capacitor. Since even that stuff was quite fast, we had to specify that he layout must include SEVERAL ceramic caps usually one per chip or one for every two chips, spaced across the PC board. This was to keep the localized spikes low. At one time we even used 10uf in parallel with 0.1uf, one set every two chips or so.

Some of the systems i worked on maybe around mid 1970's contained several boards perhaps 6 inches wide by 12 inches long, each populated with many many 7400 series TTL chips. Proper supply bypassing was a must.
One of the projects i remember was a precision weigh scale used to measure animal organs for studies done to see the effect of certain drugs on the various internal organs. The organs had to be weighed with high accuracy so the scales had to be very fine tuned, but the main point here is that they were ALL using 7400 series TTL because that was the workhorse of the industry back then. Imagine how many chips it would take to create a multi digit weigh scale out of pure TTL chips like counters and gates. It was nuts, but it worked well and was fairly reliable despite all the chips. The main failure modes were connections because some of the stuff was wire wrapped (a PC board for a custom unit was not a thing back then).

Needless to say, we had to buy capacitors by the ton

 

Hi. In this circuit, why is C3 nonpolarized? I read that polarized is more compact and powerful and so preferred for DC. I don't get C3. And where is ground? Shouldn't there be a ground at COM?

A better starting point for your question is, since capacitors in parallel add, what is the point of putting a 100 nF (0.1 uF) capacitor in parallel with a 100 uF capacitor? Even if the tolerance on C2 were 1% (and it is more likely 20%), the uncertainty in the value of C2 would still be ten times the value of C3. So why is it there to begin with?

The answer is that no real capacitor is perfect and all real capacitors have resistance and inductance. The larger the capacitance, in general, the larger these parasitic values are. As the frequency gets higher, the parasitics begin to dominate and, at some point, that capacitor actually starts behaving like an inductor. To counteract that, you put smaller capacitors in parallel with it that behave like capacitors to higher frequencies. In very high speed, noise sensitive circuits you might have four to six capacitors all in parallel to give you a subcircuit that behaves closer to a single ideal capacitor over the frequency ranges of interest.

As to why it isn't polarized, the reason is primarily that it doesn't need to be. We don't use polarized capacitors because we want to, we use them because that's the only way to get sufficiently large capacitances in acceptably small packages. But if we only need 1/1000 of the capacitance, we don't have to live with the downsides of a polarized capacitor to get it.

As for "ground", that is an arbitrary decision on someone's part as to which node to assign a voltage of zero to. As long as you don't implicitly refer a signal to ground, you don't have to have one. However, the above circuit DOES refer a signal to ground, namely V1. It is 9 V relative to something. But what is that something? It's the ground node. So, yes, in this circuit there needs to be a common reference node defined. It is almost certainly the bottom node in that schematic, but that isn't absolutely guaranteed, so it is a poorly draw schematic. If it had shown a 9 V source connected between two nodes, then the ground symbol would not have been needed (but probably still would have been useful to show).

 

Some kind of a ground symbol connected to common would look nice, but it is possible that the designer intended that common to connect to something other than ground or be left floating.

But it leaves the question of what V1 is referred to. When someone hooks up their 9 V battery to power the circuit, they put the positive lead to the V1 terminal. But where to they put the negative lead? The schematic doesn't provide that information and forces the reader to make assumptions or guess. It's a poor schematic that does that.

 

But it leaves the question of what V1 is referred to. When someone hooks up their 9 V battery to power the circuit, they put the positive lead to the V1 terminal. But where to they put the negative lead? The schematic doesn't provide that information and forces the reader to make assumptions or guess. It's a poor schematic that does that.

 

You guys great! Thanks so much