v2.0 : Adjustable anti-aliasing filter (3rd order)

[gwEm303]

Using the unused U2b we may implement this 18dB/octave adjustable anti-aliasing filter:

https://www.dropbox.com/s/3snxq4brou7bscp/OAS%203rd%20order%20adjustable%20AA%20filter.jpeg?dl=0

This has been tested in circuit and produced audibly better results than the simple filter especially for medium and low sample rates. The musician may dial in the filter by ear between 15.1KHz and 1.8KHz to reduce or maximise aliasing as they see fit.

In order to implement this with a v1.0 PCB is possible but challenging. Would recommend sticking with the simple anti-aliasing filter there unless you are an intermediate or advanced solderer.

Construction notes Although the circuit may seem difficult to implement on a v1.0 PCB, it is not as bad as it seems. The two 5.6K resistors and the 4n7 capacitor can be mounted directly to the dual gang pot. The 1nF can be connected on the solder side of the PCB, between pins 4 and 5 of U2b. This just leaves three flying leads to the pot - from U2a pin 1, U2b pin 5 and U2b pin 7.

The 82ohm and 100nF capacitor follow the same layout as the simple anti-aliasing filter and can be mounted to the solder side of the PCB board.

Don't forget a diagonal cut above the ADC (U1) pin 1 in order to cut the existing trace to U2a pin 1.

BOM 1 of 82ohm resistor 2 of 5K6 resistor

1 of 1nF capacitor 1 of 4n7 capacitor 1 of 100nF capacitor

1 of dual gang 47K or 50K log pot (sometimes called a stereo pot)

[jnlundberg]

photo of soldered pcb would be great

[gwEm303]

sorry, I can't take one of those at the moment

[jnlundberg]

"The 1nF can be connected on the solder side of the PCB, between pins 4 and 5 of U2b."

-should this instead be pins 5 and 6 of U2b, as in the schematic?

[butterrsoft]

This mod replaces the previous single-gang pot, is that correct? When implemented, does the new pot still dial in variable gain, and the filter follows? Or does it just adjust the filter, and the gain is fixed?

And no, jnlundberg, the cap is not shown between pins 5 and 6. It goes between 4 and 5 (4 being ground).

[gwEm303]

This mod replaces the previous single-gang pot, is that correct? When implemented, does the new pot still dial in variable gain, and the filter follows? Or does it just adjust the filter, and the gain is fixed?

And no, jnlundberg, the cap is not shown between pins 5 and 6. It goes between 4 and 5 (4 being ground).

that’s right!

[butterrsoft]

Err, sorry, that answer is a little confusing! So the new pot replaces the old one, and still adjusts the sample gain, yes?

[gwEm303]

:-D

no, you end up with two pots : one for the gain and one for the anti-aliasing filter

[butterrsoft]

So the old pot connects in after U2A pin 1 but before the leftmost 5k6 resistor, yes? (Leftmost in your circuit diagram linked above)

[gwEm303]

the pot to adjust the filter is a dual gang pot and only does the filter. one gang of it connects after the first 5K6 and the other gang of it connects after the second 5K6

to control the gain there is a 100K log pot which is in the feedback loop of U2A (see “improved input stage”)

[butterrsoft]

Righto, thanks for the pointers :)

[jnlundberg]

Hey guys, I never got this to work. Please let me know if succeed. Would be nice with a picture. And an updated schematic with all the mods. Not sure how to trouble shoot it. Pcb should be ok, I ordered with a midi interface pcb and it worked out fine.

[butterrsoft]

I'm marking up a new schematic to make an smd pcb from. Not sure if the 3rd order filter will fit, but I can share the schematic when done.

[jnlundberg]

I got parts to do a few more of this through hole so it'd be great if I van use the schematic with it

[butterrsoft]

OAS schematics: with 1st order filter - https://fastupload.io/6788f0d7c0f750b3 with 3rd order filter - https://fastupload.io/49db4cdfce478cad

Both use the v2.0 conventional input stage from gwEm303 :)

[jnlundberg]

Amazing, thnx a lot!

[echolevel]

Good stuff, everyone! @butterrsoft - would you mind reporting back on whether you have success with either/both circuits? If you can confirm success, we can add the PDFs to the repo (if you give us permission, of course). Thanks!