[sdiy] Analysing 4-stage RC filters without the hot air

Tom Wiltshire tom at electricdruid.net
Tue Mar 22 20:59:45 CET 2016


On 22 Mar 2016, at 15:54, Timothy Daugard <daugard at cox.net> wrote:

> On 3/21/2016 5:14 PM, Tom Wiltshire wrote:
>> Yes, it was me. I was struggling with using impedances to analyse these things. It's ok for simple cases, but I couldn't get more complicated things to come out at all. This wasn't due to tedium - I was using the computer to do the sums. One thing I never got entirely clear from the discussion on the list was whether the impedance method actually *works* in more complex cases (in which case I had errors somewhere) or whether
> 
> The trick I found to doing this with my spread sheet was that you had to keep everything as a complex number and use imaginary math functions in the spread sheet. I started with 0 phase angle signal input to the array and then convert the output back to a real output graph.

Yeah, that sounds like what I did. I wrote specific functions that took complex arguments and did the typical operations - complex multiplication, division, additions, etc. Then I used those for a function for parallel impedances and one for a potential divider with impedances. Series impedances are obviously not so complicated.

The final step was just to write out the two-pole and three-pole RC networks as nested functions. Still couldn't get it to work. I think "Doh!" is the phrase I'm looking for...

I suppose I should go back and look it over. A few months away from something often makes spotting an awkward bug much simpler. When you're all tangled up in it, you can only see what you *think* should be there. Once you've forgotten what you thought you were doing, you can see what the code you wrote *really* does!!

Tom




More information about the Synth-diy mailing list