Quick question
I am deisigning bandpass filters using this TI document.

http://focus.ti.com/lit/an/sloa093/sloa093.pdf

When I choose a capacitor to start with do I use pF, nF, uf, or maybe even just F?

Thanks for your help!

 

If you always include your units, then it does not matter.

Example: From "here" to "there" is 1234 millimeters, or 123.4 centimeters, or 12.24 decimeters, or 1.234 meters, or 0.001234 kilometers, etc. It all means the same distance. It's the same thing with Farads...

You will, however, find that 1234 Farad capacitors are hard to find. You have to use a value that you actually can find. A sure sign of a complete amateur is a design using X Farad capacitors and single digit resistors. Use microF or nanoF values and K-Ohm (usually) in your designs. Just be sure to include the units in all your calculations.

 

well, my problem is when i do the equations with say a 0.01uF cap and 30hz....

my resistor values come out as follows
0.0053
0.1007
0.000279

what do I multiply those by to get the correct resitor value.
do i divide it by 0.00000001 (0.01uf)?
because that even gives me resistors in the Megs decade.
way too much.

 

Easier to enter values as 0.01*10e-6, it also keeps me from having to count the number of zeros. I work with MathCAD (not Matlab) where I can enter the values as 0.01μF (or whatever).

I would suspect the calculations that gave you fractional ohm resistors is right. Since you can't be expected to actually use these values in a circuit, you have to choose a different value capacitor. That's the point of doing the calculation.

If you want a bigger resistor, you need a smaller capacitor.

Try that.

 

So the values of the resistors in that TI document will be actual?
I just have to work with the cap values until I get a realistic resistor value?

i guess i just have to start outside of thier suggested window of 0.01uF and 100pF.

EDIT:
The document should give me a decade to start with.
should i do the calculation is uf or pf or what.

this is really getting to me.
i'm not stupid so i know the design guide is missing some info.

 

I don't really know what you're trying to do, but if you just use all values as:

Frequency in Hertz, not KHz o MHz.
Time in Seconds, not milliseconds, hours, days, etc.
Capacitors in Farads, not μF, or pF.
Resistors in Ohms, not kΩ or MΩ.
And if you get around to using inductors, Henrys, not mH or μH.

If your values are very big or small, supply the proper number of zeros, before or after the decimal point, to make it the units like I said above. And do so with powers of ten (or I won't read it!)

Do this, and your answer will be correct. If it is not the answer you want, you will need to change your inputs.

 

OK.
I got it.
I'm using a basic scientific calculator for this so i have to convert the cap values to just farads.

works just fine with typing lots of zeros.

I was having a brain fart.

thanks for your time and patients.

 

It's good to keep the R and C values within sensible bounds too. As a general nice-to-have rule, make the caps > 100 pF and the resistors between 1k and 100k or so. This is mostly so the parasitic stray capacitances of the op-amp don't influence the filter response too much, but also to to give good interference immunity.

For a 30 Hz active filter, you might want to start with a capacitor seed value of 1 uF or so.