4 Pole Butterworth high pass filter using TL072

[jimladladlooklike]

Hey folks. So an update on this project for those that are interested:

My buddy wanted a steeper slope, so I said I would make him a 6 pole filter. I played about with the filter tool and Q calculator resources some more (whilst not really understanding it fully) and came up with this:



As you can see the response is not totally flat, it has a high Q factor. I made the prototype circuit just to see how it sounded anyway - this contains mainly sockets in order to swap components out. (Also worth mentioning I took Rob Strand's advice and stuck my filter in between the parts of the circuit shown to me by him, thanks again Rob.) It sounds good, but I wanted to see if I could rid that slight hump in the low mids to avoid it potentially not being up to scratch.

Anyway, have been doing some more reading of the article shared by ElectricDruid and soon realised I could use the specified DC gain values of the 6 pole filter in the table and apply them to the equation (2) to figure out what Q factors would give the desired, flat Butterworth response.

So, I am going to do some more work with the filter design tools and get back to you all with what I get.

Thank you all again for the amazing help.

EDIT - Yes I am going to be using a switch, but for true bypass.

Also the Q values I have found (if anyone is curious to check my work) are: 1.934, 0.707, 0.518.

cheers

[PRR]

6-pole may not be more musically useful than 4-pole, because all musical sounds have harmonics, and the ear will recreate a missing fundamental from the harmonics.

Go ahead, build it, take his money, but don't assume it will nail the sound in his head.

[jimladladlooklike]

Hey all,

So I've built this thing finally. Messed about with the circuit whilst my friend played through it and came up with the following layout. (Cuts out so much bass, but he was thrilled so great!)

We noticed a popping when the effect is engaged/disengaged, then realised I hadn't put a pull down resistor on the output. After adding said resistor it seems as though the popping persists. Does this mean I need to add JFET switching?

Thanks.

[ElectricDruid]

Which stage is which? The last schematic you showed has 82n for the last stage, but you've clearly tweaked it some more since then (and there's nothing wrong with that!). I can't work out which bit does what, so I'm finding it very hard to say anything useful!

[jimladladlooklike]

Apologies! let me get the new schematic for you. Bear with...

[jimladladlooklike]

Here it is. First stage is an input buffer, the rest is the 6 pole filter, cut off point 200hz

[antonis]

I presume bias configuration is omitted for simplicity..

[ElectricDruid]

Thanks. Does that 2M2 go to actual Ground, or does it go to 4.5V like the 330K/100K/82K's?

It *should* go to ground, not 4.5V. It won't do any anti-pop duty tied to 4.5V.

[jimladladlooklike]

Yes bias config is omitted for that reason.

The 2m2 does go to actual ground. cheers

[antonis]

In case of pop issues derivation doubt, implement 3PDT switch wiring of IN & OUT grounded..



In case of the above mod isn't anti-pop effective, also implement LED anti-pop configuration..

[jimladladlooklike]

Thanks antonis

[PRR]

Here it is. First stage is an input buffer....

First stage lacks bias. The JFET opamp may "work" by clipping the start of notes and holding that charge a long time. But it's not proper. Add 1Meg to Gnd.

[Rob Strand]

Wait a sec. I used the values in Reply #20 but the ckt you posted later is different
It looks like you have fixed things somewhat in Reply #25


Response to reply #20

Also the Q values I have found (if anyone is curious to check my work) are: 1.934, 0.707, 0.518.

Those Q's are correct for a 6th order Butterworth filter.

However, your actual circuit seems to have Q's of 1.23, 0.605, 1.17.  Also the each stage of the Butterworth filters have the same f0 but on that circuit the first two stages have f0 of 196Hz and the last stage has an f0 of 251Hz.   So as far as implementing a 6th order Butterworth filter correctly it could be off.

As you can see the response is not totally flat, it has a high Q factor. I made the prototype circuit just to see how it sounded anyway - this contains mainly sockets in order to swap components out. (Also worth mentioning I took Rob Strand's advice and stuck my filter in between the parts of the circuit shown to me by him, thanks again Rob.) It sounds good, but I wanted to see if I could rid that slight hump in the low mids to avoid it potentially not being up to scratch.

Having said that it's possible to have different filters with different Q's and un-equal f0's and still produce a Butterworth -like response, especially for high order filters.  So if it sounds OK maybe you should leave it.  In fact if it sounds OK do you care if there is a small peak.   The simple way to knock out the peak is to just decrease one of the grounded resistor values by one step.   That has a side effect of nudging the cut off up a bit.  Try each resistor and see which one has the nicest response (I suspect the second stage).

When designing filters with real discrete part values it's very common to have small peaks in the response.  If they are objectionable then you have to manually tweak some of the parts.   (You can do it with computer design but the level of difficulty skyrockets by a large factor compared to hand calculations and you need a way of telling the computer you don't want any peaking.)

[jimladladlooklike]

Thanks guys!

So the first stage is actually biased by a 1M resistor from the non inverting input of the first stage to 4.5VDC, is that sufficient?

Cheers

[jimladladlooklike]

Can I use that LED configuration in combination with the 3PDT switch wiring from the first diagram? (Input and output grounding) As in can I wire up the footswitch the first way and just add the 4.7k, 300r resistors and 10uF cap to the LED? That's how I have it wired up currently (with added 100k from LED cathode to footswitch to reduce brightness of the blue LED I'm using.) Unfortunately this doesn't seem to be working either

[antonis]

Both 3PDT switch wirings are well verified..

You can use 1st diagram (LED cathode to lug 3) and split CLR (whatever value you use) into 2 X CLR/2 connecting 10μF cap to their joint..
Cap's negative leg should be connected to any convenient permanent GND..
(not to lug 3 / LED cathode..)

[ElectricDruid]

So the first stage is actually biased by a 1M resistor from the non inverting input of the first stage to 4.5VDC, is that sufficient?

Yes, that should be fine. That's exactly what PRR suggested a bit earlier on.

[jimladladlooklike]

Okay so I have carried out all suggested mods and still have that popping sound.

[jimladladlooklike]

So I've been reading this article:

https://www.mrblackpedals.com/blogs/straight-jive/6629778-what-really-causes-switch-pop

I think there may be something to be said about the values of the current limiting resistors on the input and output?

Perhaps if I up the value of these resistors to 100k?

Also might I have got the 100nf cap and 1k resistor on the input of the first stage the wrong way round? Should the signal go input > CLR > capacitor > opamp?

[antonis]

Mr. Black's aspects should be considered only after serious lookup...

e.g. all those about input current limiting resistors are true & correct ONLY when cap right leg, hence circuit input is grounded - and I haven't met many of that kind..
Usually, input cap is (or should be) connected to circuit's input impedance, the resistive part of which is significant so another series resistance adds nothing but signal degredation 'cause it's effectivelly  added on signal source (or previous effect) output impedance..

P.S.
Of course, like almost all aspects, it's a point of view matter..
(but that's my point of view..)