PT2399 hi/lo pass filtering

[John Lyons]

I have a few examples of how it's been done but
I can't find a simple primer on theory with examples.
Yes I know there are some but in the examples I can't
find much on hi pass filtering.
Here is one but in the post it seems there there is an issue
with it as far as the resisitance to ground affecting the opamp.
http://i1294.photobucket.com/albums/b605/cdouch/echobasefilter_zps7e556fd5.jpg
https://www.diystompboxes.com/smfforum/index.php?topic=60662.msg895129;topicseen#msg895129

[Ripthorn]

Are you talking about high pass filtering the entire input into the PT2399, or just the repeats, or something else? The typical filtering for all input to the PT2399 is done via a multifeedback low pass arrangement. For just the repeats, sometimes it's done with simple RC filters (like what you have shown in your example), sometimes a tone pot (like the EQD Disaster Transport Jr.). Something like the Deep Blue Delay will take the output from pin 12 and run it through another multifeedback low pass filter using the pin 13/14 opamp. The arrangement shown in your example looks like it's intending to (after the selection of the high or low pass RC circuit), but it's not quite right. The electrosmash article on PT2399 circuits is great (https://www.electrosmash.com/pt2399-analysis). You can easily do a multifeedback high or low pass, it just depends on the arrangement of the caps and resistors. Here is the calculator of multifeedback high pass: http://sim.okawa-denshi.jp/en/OPtazyuHikeisan.htm. You will likely want some low pass in there as well to get rid of any potential high frequency hash left over from the PT2399 output.

[John Lyons]

Thanks
I was looking at repeats filtering.
In the examples I've seen (ther than the one linked)
I did not see a series cap and resistor to ground as
a typical high pass filter but maybe I am looking at it wrong.
I'll check out theose inks.

[anotherjim]

We usually don't need to add anything for high pass filtering in these and many other circuits. They are there already in the series AC coupling caps and the PT2399 opamps input resistors. Generally, the standard datasheet examples choose R and C values that give full audio bass and all you need to do is make the cap smaller to reduce the bass.

[John Lyons]

Thanks Jim
So in this example which I like and uses some
high pass filtering seemingly...Where is the hi pass
filtering coming into play? Seems like it use standard
cap values.
http://thermionic-studios.com/wiki/images/3/32/1776Effects_MultiplexEchoSchematic.png

[Rob Strand]

The *low* pass filters are MFB types like in section 7 of this document,
https://www.ti.com/lit/an/sloa049b/sloa049b.pdf

The (low frequency) input impedance is R1.   At low frequencies it looks like an inverting amplifier with R1 = input resistor and R2=feedback resistor.

With this in mind the input impedance setting resistors in the your example circuit are.
R14, R32, R12, R16, R23.

R16 and R12 act like an inverting summer.

The input impedance setting resistors which have series caps and form  *high-pass* filters are:
C10 + R14
C9 + R12
C12 + R16

The C9 and C10 cases have large caps (1u) so there isn't lot of high-pass filtering.

The key cap is C12 as it is a small value.   it's also the one for feedback!
In simple terms the high-pass cut-off frequency is set by f3 = 1/(2*pi*C12*R16).

However, it's not that simple! 

When the feedback pot is full.  R19, R21 and  R16 all add together to form the input resistance.
So the cut-off is really 1/(2*pi*C12*(R19+R21+R16)).

When the feedback pot is backed off the part of the pot between wiper and C12 add resistance to R16.
At the mid setting of the feedback pot the pot resistance is 25k.  When that adds to R16 the pot
is contributing most of the resistance.

So the high-pass cut-off increases as you advance the feedback pot.  There isn't a single high-pass frequency.

(I've actually simplified some some stuff in my description because it's already complicated enough.  For example when
the feedback pot has say 2k between the wiper and C12 then you need to consider R19, R21, R16 and the 2k from the pot.
The pot to ground changes things a bit as well.  Yada, yada, yada  of details.)

[John Lyons]

Thanks Rob
I guess I feel a little less dense
since the answer is more dense.
I'll work towards digesting this.

[Rob Strand]

Thanks Rob
I guess I feel a little less dense
since the answer is more dense.
I'll work towards digesting this.

The quirks of that circuit make the high-pass more complicated than usual.
If it was just one input resistor and one high-pass capacitor it's not too hard to work out
and you get a single cut-off frequency.