Jump to content

Talk:Roll-off

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Two suggestions

[edit]

Good job on creating a solid article! I have only a couple of suggestions:

  • It's probably worth making clear in the article that this concept isn't exclusive to electrical networks; I'm sure it applies to other frequency-dependent systems such as mechanical resonators. It also has meaning in the digital domain.
  • I'm not sure the distinction between passive and buffered higher-order systems is relevant, as in both cases the pertinent point is that the asymptotic roll-off is the same. And at the moment, the article sort of implies that Butterworth, etc. filters only exist in passive topologies.

Oli Filth(talk|contribs) 11:05, 5 July 2009 (UTC)[reply]

Thank you for the kind words. I have reworded the first sentences in the lede to make it more general. Do you think that is ok now? On your second point, did you write that before I made this edit which was meant to address that very point, I suspect we may have "crossed in the post", or do you think it still needs some work. On the digital issue, a method of synthesising higher roll-offs in the digital domain is mentioned, but you are right the article is skewed to passive analogue, that's my background and I tend to write about what I know. SpinningSpark 11:25, 5 July 2009 (UTC)[reply]
Your rewording largely covers it, although perhaps "LTI system" would be better than "transmission function".
As for higher-order filters, my thought was more that we could merge the two sections into one quite safely, as I believe that the details on buffering and coupling are largely unrelated to asymptotic roll-off. I think all that needs to be something to the effect of "higher-order systems have multiple poles; however in general the roll-off still converges to a constant gradient at high frequencies". Oli Filth(talk|contribs) 12:56, 5 July 2009 (UTC)[reply]
Is this an improvement? SpinningSpark 15:58, 5 July 2009 (UTC)[reply]

Reference missing

[edit]

The "Mayer et al." reference listed under the Notes section isn't mentioned in References or anywhere else. Which article/book is that? --Yerpo (talk) 13:11, 13 July 2009 (UTC)[reply]

It is there, right at the end: ISBN 089004385X. SpinningSpark 16:28, 16 July 2009 (UTC)[reply]

Circuit error

[edit]

Concerning following image: Multiple order RC filter buffered between stages The output of the op-amp must feed back to its minus input which is wrongly connected to ground! —Preceding unsigned comment added by 213.102.118.193 (talk) 01:12, 19 December 2009 (UTC)[reply]

It's not an error, and it's not an op amp. The error is that you are assuming that all amplifier symbols are op amps. SpinningSpark 03:46, 19 December 2009 (UTC)[reply]

Attenuation at Corner Frequency for Various Orders

[edit]

It is worth drawing attention to the definition of corner frequency for higher order filters. For a 12 dB/octave filter, should the attenuation be -6dB at the corner frequency? Yes is the most sensible answer, since an inclination to define the corner frequency as the -3dB point will become rather meaningless for a very high order filter with all of its poles at the same frequency (which will be a long way off the -3dB frequency). — Preceding unsigned comment added by David.J.Greaves (talkcontribs) 14:06, 7 May 2019 (UTC)[reply]