Tim Escobedo's 9v MS20 / Sallen-Key LP filter: how to turn it to HP?!

[HeaD]

Hi everybody. Could someone help me to turn this low pass VCF to high pass?



Thanks!

[frequencycentral]

Highpass can be derived from lowpass by mixing the unfiltered signal with the filtered signal, providing that one is 180^o out of phase with the other.

[HeaD]

Thank you Rick,

so this means that I should split the signal and then mix the two resulting signals (clean/filtered) again?

I was wondering if there is a way to modify the schematics making it HP rather than deriving an hi-pass adding some other stage to this circuit. Usually simple Sallen-Key filters are reversible (from LP to HP) changing positions and values of resistors and caps. But it's hard for me to figure it out in this case.

Is it possible?

[oskar]

Maybe this link could be of interrest.
http://www.modular.fonik.de/files/EFMforum/orgEFMfiles/vcf8e.pdf

It's electrons for musics take on the ms20 filter.

[HeaD]

Hi oskar, I had already seen this article using the search tool... is very interesting but unfortunately I need a 9v powered filter :/ Thank you

[PRR]

> unfortunately I need a 9v powered filter :/

Yes, but the DC set-up and the signal path are different (though interconnected) things.

Look at the EFM 8E filter. It does both low-pass and high-pass. In low-pass, ignoring the DC set-up, the added control voltages, and some part values, this is the "SAME" as Escobedo's 9V filter. Input, transconductance amp, cap to ground and buffer, transconductance amp, cap to ground and buffer, variable "resonance" feedback (transconductance amp or simple pot), output.

Now throw the EFM 8E low/high switch to high-pass. What changes?

Signal ground to transconductance amp, cap to ground and buffer, transconductance amp, cap to Signal Input and buffer, variable "resonance" feedback (transconductance amp or simple pot), output.

The EFM 6E's signal path is at DC ground. To run on single 9V grounded battery we must lift the signal path up around 4.5V, and block any DC path to power ground. So keep the 0.1u cap at Escobedo's INput, but connect the left end to ground. The second 680pFd cap, tied to signal ground in Escobedo's low-pass, gets connect to the box INput (because it is a cap there is no concern about DC upset).

Rick's "Highpass can be derived from lowpass by..." is universal and direct. EFM's re-wiring is subtle, clever, and very elegant, but needs profound insight into how this filter works. I used to do this stuff but I forgot this trick.

[Taylor]

You don't need to copy the whole circuit, just look at what the HP/LP switch does. In lowpass configuration, as in Tim Escobedo's schem above, the input goes to the first OTA's inverting input (via a resistor) and the cap right after the second OTA is grounded (in Tim's it goes to VR).

Then in HP mode, it just switches these: the input goes to that cap and the first OTA's inverting input is grounded (via the resistor). In Tim's circuit, you'd want to send the first OTA's input to VR rather than ground.

That's my guess anyway. Anyone feel free to correct me.

Edit: Well, Paul beat me.

[HeaD]

Thanks to both of you. So, is it something like that?



Thanks!

[frequencycentral]

If it were me I'd try something like this, maybe it needs a few tweaks:



The output of the Escobedo VCF is non-inverting with respect to the input, so the input is inverted U2a. The (inverted) unprocessed signal and LP output are mixed into U2b. The extra pot at the input of U2b should be the unprocessed signal fully anti-clockwise, lowpass fully clockwise, highpass somewhere in between where the unprocessed and lowpass are mixed at equal volume.

Alternatively, you could have two outputs, one LP the other HP, then you'd get supercool stereo filter swoops. I'm able to do this with the Oberheim SEM multimode filter in my DIY modular. A must-have sound once you've discovered it. You'd want to make the extra mix pot into a trimmer and set in just right.

[HeaD]

Thanks Rick. So... according to the mods you've proposed, I could use a 2pdt to switch between lp and hp sections, assuming that the trimpot will be set at the right value (where the two signals are mixed at the same level). Is that right?

[frequencycentral]

Thanks Rick. So... according to the mods you've proposed, I could use a 2pdt to switch between lp and hp sections, assuming that the trimpot will be set at the right value (where the two signals are mixed at the same level). Is that right?

Yeah, that sounds like a plan!

[Taylor]

Hi again, could you please take a look to this configuration? Thanks!

That's what I would try, except that Paul is saying you should keep that .1uf cap on the OTA's inverting input instead of losing it when you switch to HP.

[PRR]

[mwvm]

hello...i did a little google and found this page....

i built the lpf roughly 3 year ago... the only thing i needed to adjust from the original drawing was to add some resistance in series with the freq pot to sort some dead ends. i think tinkered with the 680pf caps too. i decreased the value to get more range as if i remember rightly - it starts to roll off 1khz, passing very little at 5khz.

i used a lm13700.

since then i bought a frostwave resonator, but had to sell it as i needed to cash. I MISS THE HPF!!!!





if i built the above hpf circuit would this work if the lpf followed the hpf in series?

[mwvm]

well - parts are ordered.

i assume lpf following the hpf will work, there's no reason why it shouldn't....just probably need to tinker with cap values so it passes a full range of audio when freq pots are fully clockwise.


i'm also having a think about using an lfo for the freq.