Electrical Faux Pas

[z_p_e]

wattsup,

There are two main capacitor types: polar, and non-polar.

Non-polar caps are ceramic, film, mica, mono etc. These can be used for AC or DC with no concern about polarity.

Polar caps are electrolytics, tantalums, etc. They will have a + or - sign on their case. These can be used for AC or DC as well, but care must be exercised in both cases.

When used for DC, make sure you observe the polarity of the capacitor in relation to the power supply. Of course make sure the voltage rating is ok too.

When used for AC, the "safest" way to use them is by connecting two of them in series, and back-to-back. This is now a non-polar electrolytic, and these can actually be purchased, although they are not too common. Remember that when you do this, your final capacitance value will be one-half of that screened on the case. Back-to-back means connecting both +'s or both -'s together. It should not matter which one you choose.

You can "cheat" with an electrolytic when a high value capacitance is required and you don't want to use two in series, but only with "line-level" type signals....nothing of significant power or voltage swing. "Cheap" or poorly designed audio electronics often does this. It will work, but don't expect hi-fi results, the audio will actually become distorted.

The final way of using an electrolytic, or polar capacitor with AC is with a circuit that has a large DC offset. A perfect example would be to decouple the output of an audio amplifier that is using only a single DC supply. The amplifier will normally be biased half way between the supply voltage, and since you do not want this voltage constantly powering your speaker, you insert a capacitor in series between the amp output and the speaker to decouple this DC. In this case, you may use a single electrolytic cap, just make sure the +'ve lead is at the amp output, and the neg lead is connected to the speaker terminal. The other terminal of the speaker must be connected to ground in this case.

Cheers,
Darren

[sparks]

    Sorry for messing up this thread

[wattsup]

@Sparks

You should post this in the TPU Discussion thread instead of this thread cause we are using this thread to discuss general electronics. Good post though.

@z_p_e

Again regarding the caps, and thanks for your previous post which I am starting to understand.

On DC circuits, I rarely see a capacitor that is in series (except on the large TPU that has those two black caps in series). They are usually in parallel to other components or between the + and - lines. I have tried putting caps in series but this does not work as I cannot get any voltage out of the other cap terminal.

Let's say you ran a small dc motor off of a battery. You can put the cap on the + and - before the motor, but you cannot put it only in series on the positive line. Even with transformers, I have not been able to. Are there special caps for this or is this simply impossible.

[z_p_e]

wattsup.

Capacitors are used as follows:

For DC, the capacitor is always in parallel to the load or battery. So the capacitor is "across" the battery or load, not in series with them. Capacitors do not "pass" DC voltage.

For AC, the capacitor is usually used in series for AC coupling (or decoupling of DC), except when used in filter circuits. In filter circuits (such as a simple RC), the capacitors can be in a "shunt" (parallel) configuration.

A couple other tidbits:

A transformer gives a similar situation to a series capacitor; unless the input is changing (i.e. AC), there will be no output. It's a little more complicated than that, but in general and to keep it simple, this is true.

A pure DC voltage is one that does not vary with time. But don't call a DC voltage that changes (drifts) from 10V down to 9.5V over a two day period AC either....it's just a drifting DC voltage

So any time you are "switching" your battery on and off in relatively quick succession (either manually, or with an oscillator and MOSFET switch), this should be considered an AC source.

[z_p_e]

Vortex,

I think in terms of ?transit time? and ?switching time?, SM may have been making a generalization and perhaps meant both in combination? If precision timing and instantaneous correction is required and involved, then I believe the bottom line is ?how fast can the device cleanly transfer its input to its output??

Inter-electrode capacitance in BJT's will cause group delay and slower switching for pulses, and I suspect this will have a far worse effect than the transit time through semiconductor material. In the end, it messes with the intended output, delaying higher frequencies more than lower ones. As a result our output pulse is "sloppy".

Tubes typically have much lower inter-electrode capacitance compared to BJT's and MOSFET's, but the transit time (strictly speaking) may not be faster. As you say, there is a physical distance the electrons must travel, and that takes time. In semiconductors, junction distance is very short compared to tubes, but the velocity of propagation through the pn material may be significantly lower than c, and as such the race might even out.

According to this web page: http://www.john-a-harper.com/tubes201/  tube transit time is about 1ns.

Let?s assume a Plate to Cathode distance of 1cm (0.01m). If the electrons traveled at a constant speed of light (they don?t), this transit time would be considerably less?.around 33ps. This is a factor of about 30 times less than c. Evidently, the average electron velocity of propagation in tubes is much less than c?unless of course the 1ns figure is incorrect.

Just for comparison, here is a recent comparator device from National:

?Comparators feature sub-nanosecond propagation delay.

May 22, 2007 - With 700 ps propagation delay, dual 21 mA Model LMH7322 features rise and fall times of 160 ps and dispersion of 5 ps at greater than 100 mV overdrive.?

In other areas, transistor switching times are getting down to a few pico-seconds:
http://www.cs.clemson.edu/~mark/464/transistors.html

So ultimately, I don?t know what to say of the assertions that tube transit times are much less than SS, other than in regards to the TPU, it may not matter because in the end both work.