PAIA2720-3B VC Bandpass Filter_Augmentations

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

I tried out the PAIA 2720-3B VC Bandpass Filter on the breadboard. I wasn't initially that impressed with its range. However I swapped the diode for a red led and drove it from an exponential current source, the result was a great increase in the range of the filter. I also changed the way in which the Q control was implemented, resulting in, in my opinion a more natural control contour.

I also made a second version with a simpler current source using just one NPN transistor. The results were still very impressive.

The most interesting characteristic of these filters is the way they behave when overdriven. With high Q settings, the LED becomes modulated by the very signal that it is controlling and things freak out somewhat, the LED flickers too in response to the large signal swing!

The original is included for comparison.

PAIA 2720-3B_LED_ExpoCurrentSource_Version.jpg

Description:

Filesize:

288.89 KB

Viewed:

1086 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

PAIA2720_3B_LED_Simple_CurrentSource_Version.jpg

Description:

Filesize:

212.14 KB

Viewed:

1073 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

paia_2720_3b_vc_bpf_Disrete_Design.gif

Description:

Filesize:

11.33 KB

Viewed:

1000 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

I tried out the PAiA 2720-3L Lowpass filter. Again it worked better with LEDs instead of diodes and a current source to drive the LEDs. I chained two Lowpass filter sections together to make a 4 pole filter. It was quite good but would really benefit from having a Q / resonance control. I'm going to have to figure out if this can be done.

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

This version of the 2720-3B bandpass filter has a linear current source that doesn't have any tempco problems.

1495219802500.jpg

Description:

Filesize:

93.87 KB

Viewed:

994 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

I was just trying out the circuit from my last post again and it would seem that 68K is a better choice for R12 as it allows the filter to be completely closed.
I suspect that there might need to be a bit of experimentation with some of the resistor values to accommodate different LED's. Specifically, R9 & R12.

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

I built 2 of the PAiA 2720-3B bandpass VCFs and designed a little two transistor mixer and output stage with 10 V peak to peak maximum output.

The sinewave generators represent the 2 filters, the 10K resistor on the right represents the output load.

2Ch Discrete transistor mixer for dual VCF.png

Description:

Filesize:

48.18 KB

Viewed:

902 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

n.d · Joined: Dec 15, 2011 Posts: 52 Location: Talos IV

Nice work!

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

Fernando · Posted: Mon Aug 27, 2018 11:26 am Post subject:

very nice gasboss!

I always think about trying the most simple PAiA circuits.

Well for the twin T filter (2720-3b) I'd try with a photocell instead of the diode and a circuit to control the LED. Or a real vactrol instead of a DIY one. But DIY you can use a more adequate range for the photocell and have less work fine-tuning it's right resistance range.

There are many schematics for vactrol CV around from other circuits (Buchla LPG, etc.) but I'd try something much simpler first, the simplest possible.
Try this: http://thesquarewaveparade.com/images/VtoR.jpg (second option, with offset)
And read here: http://www.electro-music.com/forum/topic-37628.html

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

Fernando · Posted: Mon Sep 17, 2018 10:04 am Post subject:

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

Yeah, Ge diodes sound like a good idea, never tried them at the time.

AlanP · Posted: Tue Sep 18, 2018 10:21 am Post subject:

My experience of clipping diodes in pedals is that Ge diodes have a "softer" sound than Si diodes. You need a couple in series to match the Fv of silicon, though.

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

Fernando · Posted: Mon Oct 01, 2018 10:38 am Post subject:

but in the same position as in the PAiA band pass?
I mean, as the variable resistance to ground at the bottom of the twin T?

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

Yes the diode was in the same position as with the PAiA circuit.

AlanP · Posted: Wed Oct 03, 2018 8:40 am Post subject:

LEDs have a fairly high Fv compared to silicon or Ge, they're pretty harsh in pedals, too Smile

translates to high levels before clipping.

cslammy

gasboss775 · Joined: Jan 02, 2016 Posts: 217 Location: Scotland

vladosh · Posted: Fri Dec 07, 2018 11:33 am Post subject:

Here's some PCB layout for this ,if that's ok . I didn't made the add-on mixer ,just the last linear response version ,wanna try this version , have built this filter once earlier as the original on 18V ,PCB is 70x100mm

Paia BPF Adaptation - Parts View.jpg

Description:

Filesize:

496.74 KB

Viewed:

372 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Untitled.png

Description:

Filesize:

19.19 KB

Viewed:

368 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

danielfiction · Joined: Jul 15, 2009 Posts: 7 Location: nyc

i'm really interested in this filter. i'm going to build it.

but what about modifying it for "notch", aka, band stop output? how would you go about that?

daniel

vladosh · Posted: Wed Dec 25, 2019 2:13 am Post subject:

really like how this filter sounds ,however i replaced that 47K resistor with 62K and it wouldn't close all the way , it's powered with 15V so thinking to try 82K and see how that goes i even wonder if connecting the freq pot to -V via 47K resistor is better ,nice resonance

vladosh · Posted: Wed Dec 25, 2019 5:12 am Post subject:

i that's how i messed it up ..

MapacheRaper · Joined: Feb 15, 2018 Posts: 166 Location: Spain

So a lot of chat on how good it sounds but not a miserable audio demo? C.mon guys...

I would love to build one of these but Im a disaster breadboarding

cslammy · Posted: Mon Feb 03, 2020 5:18 pm Post subject:

I am terrible at breadboarding/stripboarding as well.

I have created BRD and SCM in eagle, posting it here in zip along w/ gerber
I'll send it off on my next fab and let you know what happens.
Relatively Big board, not a lot of vias, room for jumpers, kludge wires etc.
No ground pour, no fancy stuff.

Warning: doing this while at work--checked it once, think i got it right, but may be mistakes. Passes DRC and ERD checks.

PAIA4720BOARD.PNG

Description:

Filesize:

48.53 KB

Viewed:

115 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

4720FILTER-EAGLE.zip

Description:

Download

Filename:

4720FILTER-EAGLE.zip

Filesize:

170.43 KB

Downloaded:

148 Time(s)

MapacheRaper · Joined: Feb 15, 2018 Posts: 166 Location: Spain

Great!! That´s exactly the kind of project for your blog. Waiting to read the post Smile

It´s a pity that China is halted, tho. Who knows when they will produce it. I have some PCBs paid, and in the schedule to fab in JLCPCB. Not looking good at the moment...

Whatever, cool project