It's pretty simple, really. Non-polar caps are used in locations where they must pass AC. Film caps (of various constructions) are practical until the values exceed a few uF. Once you need more than a few uF of non-polar capacitance, it's time to look at NP electrolytics, as they are the cheapest route to a large-ish value without the polarity issue.
Film caps (in the 63-350V range) are available in values as large as 10uF, 22uF, 47,uF and even 100uF - or more. But those are usually high-priced specialty audio parts, for things like speaker crossovers and output coupling caps in some audio preamp circuits. A non-polar electrolytic is nothing more than a pair of normal (polar) electrolytics, with the negative (-) leads tied together. There is no difference between a (nominal) 10uF non-polar and a pair of 22uF polars tied neg to neg. Pos to pos works just as well, but it's just not the standard for a very good reason.. think about how a standard electrolytic cap case is built & sealed.. On Mon, Aug 1, 2016 at 2:30 AM, Jim Brain <br...@jbrain.com> wrote: > On 7/31/2016 12:37 AM, Ian McLaughlin wrote: > >> Jim, >> >> These are non-polarized (or bi-polar) electrolytics. An example is the >> following digikey part number: 493-12697-3-ND >> >> You can always whip one up out of your junk box - just put 2 normal >> (polarized) electrolytic in series with the polarities alternating (for >> example, connect the two positives together). Each capacitor has to be >> twice the value of the result - so for example, to replace a 1uF 50v >> non-polarized, you can put two 2uF in series. >> >> Hope this helps. >> >> Ian >> > Thank you (and Chuck) for the response. > > I guess it brings up more questions for me: > > * Why would Tandy spec such a cap in the design? > * As I look at the specifications of the design, it looks like the > rest of the design assume mac .250W of power on the audio line > (given the selection of resistor networks and other parts) . I > noticed the Digikey unit you referenced has 17mA of ripple current > capability. I realize audio is not truly a ripple current, but the > disparity between 17mA and 250mW seems like a problem... (Again, my > analog is severely rusty, so maybe I am looking at this wrong. I > spent all night trying to find a way to determine the power of a R2R > ladder given 5V logic and 8mA output drive of the 74ls374 IC, to no > avail, so maybe I am looking at this wrong) > o you'd think with the prevalence of R2R ladder logic, someone > would have whipped up a power calculator, if nothing else, to > determine the wattage of the resistors in the R2R, but I came up > empty... I'll have to either calculate it longhand by summing > all of the currents or sim it; not sure I can assume that since > the effective impedence is 10K, the amperage would be .5mA @ 5V > and thus the power would be 2.5mW > > Sometimes, I wish I'd cared more about analog in college... > > Jim > > > >
