Howdy,
So, doing some EQ things. It's been long time since I've remembered filter math, but I've slowly noodled my way through it (helped by building myself some very useful spreadsheets).
I've been looking at the MXR-ish 6 Band, which became the basis for the PedalPCB 6 band and a few others I've seen. While I've got the bandpass section sorted, there's something odd on the schematic I don't get.
At the first op-amp stage, it's a basic NI gain stage with a gain of 2 - except for R5/C2 in parallel with R4. My thoughts are this would increase gain over a specific frequency, and with some basic modelling I was correct.
Ignoring the gyrator stages for now, hanging off the output stage is a basic voltage divider (halving the output signal), then a cap to ground - a basic RC filter.
Out of interest I set these filters together up in Falstad, and lo and behold - they almost cancel each other out. There is some slight attenuation, but it's like 3db at 25khz with both these sections in vs reasonably flat with them out.
With all that being said, is there other function these sections are doing that I'm missing? Is there any other reason they are left in, or is it just a case of everybody copied what MXR did?
(https://i.ibb.co/HP6B7dt/Screenshot-20231025-223307-2.png)
Quote from: drdn0 on October 25, 2023, 08:54:29 AMIt's been long time since I've remembered filter math,
Time for a refresment.. :icon_wink:
(https://i.imgur.com/nPPpZ6h.png)
followed by
(https://i.imgur.com/II2IcoK.png)
P.S.
C3/R8/R9C4 form a band-pass filter of 8Hz and 3180Hz..
In my very limited experience, pre-emphasis/de-emphasis circuits are meant to reduce noise.
Quote from: Ben N on October 25, 2023, 09:40:12 AMIn my very limited experience, pre-emphasis/de-emphasis circuits are meant to reduce noise.
I'd definitely agree with most other things I've messed with, but they are boosting treble (significantly) before the filter stage, then cutting it out at the end.
Surely this would have the opposite effect?
Quote from: drdn0 on October 25, 2023, 09:51:33 AMSurely this would have the opposite effect?
No, because the high frequency content of the noise ("hiss") is more objectionable. Dolby NR works around the same principle.
+1 agree with what the others have said.
You boost the treble going in (pre-emphasis), so that when you cut it later (de-emphasis) the most objectionable part of the background hiss is also reduced.
Quote from: ElectricDruid on October 25, 2023, 01:11:32 PM+1 agree with what the others have said.
You boost the treble going in (pre-emphasis), so that when you cut it later (de-emphasis) the most objectionable part of the background hiss is also reduced.
Well consider me learned! Thanks guys.
Thought I'd be able to cut some more parts out :icon_razz:
There is nothing new about pre-emphasizing the highs before processing and applying the inverse de-emphasis after processing.
This pre-emphasis / de-emphasis processing is built into FM radio. In Europe, the pre-emphasis is 50 µsec and in North America, it is 75 µsec. If a stereo decoder is used, the de-emphasis is placed at the output of both channels. Without it, there would be a lot of hiss on FM. Some tuners have a switch or jumpers to select the time constant whereas others have different models for different countries.
FM can be useful without EQ. But most speech/music sources tolerate a lot of EQ, great improvement for 30 cents.
Magnetic tape is all about high-boost into the tape and high-cut coming off the tape. This may be harder to see because the wound head does a lot of EQ which is not accounted in the amplifier description.
You can also do it pre/post a distortion, to get a not-so-nasty result. But the recipes are empirical and rarely documented well.