I don't understand how to use AP

Now that the AP is working properly on the JP6 filter, I'm trying to understand how to use it.

Most places say that you should mix it with the original signal, but when I do that I get exactly what I get from my notch:

Red: Notch. Blue: 50% AP + 50% Dry

JP6 gain through cell 2

In the simulation, cell 2 outputs a +/- 2.6V signal when output from Cell 1 is 4V. This in turn arrives at the output as 3.1V and is overdriven to 4V with a nice rounded shape. 

Now, the output gain feedback resistor is 100k. Let's do some calculations.

LP and HP are tapped through 33k + 47k = 80k. That should give a gain of 1.25, which should give an output of 2.6 * 1.25 = 3.25.

The output of the BP requires a gain of 1.33 in cell1 in the simulation, and 1.5 on the breadboard, to be equal to the LP and HP outputs

That means a total gain of 1.25 * 1.33 = 1.66 (or 1.88 on the breadboard)

Thus, we need 100k / x = 1.66, x= 60k (or 53k for the breadboard).

For the direct output from cell 1, which is at 4V in the simulation, to get it to 3.1 requires a gain of 3.25/4 = 0.81, or a 125k input resistor. 

Now, wait a minute. Gain < 1 is generally not something we do, we usually do it by using a resistor divider. But let's not, let's rise the output level from 3.25 to 4 instead, meaning we only need unity gain for cell 1.

That gives us the input resistor for LP/HP as

gain = 4/2.6 = 1.54, R = 65k

and for BP

gain  = 1.33 * 1.54 = 2, R = 50k (and 1.5 * 1.54 = 2.3, r = 43.5)

Let's see what this does in the simulation:

Well, LP still looks good, at just a tad below 4V. Then, turning up overload to max fivces a slightly more distorted wave, but still cool, and with an amplitude of 5.1.

Now for another check, what happens with a 20Vpp signal. Does it distort?

It definitely does, but that's not because of the feedback, it's just how the circuit works.

We could get around this by using a 200k input resistor instead of a 100k, and then double the gain at the end of the circuit instead. We'll just have to try this to see how it affects both a 5V and 10V signal when doing distortion.

On breadboard

I tried sending a 5V signal through cell 2, and I get 3.23V on LP, 3.32V on HP and 2V on BP, meaning we need a gain around 1.6 for BP. 

To get all up to 5V at the output would require gains of

HP/LP: 5 / 3.23 = 1.54

BP: 5 / 2 = 2.5

and unity.

100k / 1.54 = 65k

100k / 2.5 = 40k

So a combination of 47k and 18k, and 22k and 18k is a good starting point.

Testing: 

LP: Near perfect 5V out

HP: Output is 5.2V, so a tiny bit too high

BP: Damn close to correct.

This means I will try the following for cell 2 in "production":

HP and LP: 47k +18k

BP: 22k + 18k

Direct: 82k + 18k

For cell 1 I will use 56k for everything, but BP is boosted by 1.5 using a 150k feedback and 100k input resistor

Next things to try on breadboard:

- overdrive

- no resistor pre-mux for cell 1

- swap cells to see if they behave similarly

- output vca.

- polaritites for output VCA and overdrive, see if we should modify cell circuit

See "Jupiter 6 filter - no VCA - 12V - JOVE trials -overdrive"

JP6 All-pass - inverting BP

In my simulations and breadboarding I added an inverting op amp for the input to get AP. But what if we instead inverted BP? It is already 180 degrees off compared to HP and LP. 

I simulated this, and it looks like it works well. Instead of going from -360 to 0 degrees it goes from -180 to 180, which is just to say that the output is inverted compared to the input - just like HP and LP.

As I'm already planning on having an invert button/switch on the SVF this doesn't matter :-D 

It also saves me one opamp.

As for the opamp gain, to keep the gain at 1.48 (or 1.5, it probably doesnt matter too much) we can go with 100k input and 150k feedback, or 68k/100k.

I will try this now.

JP6 All-pass tested

I've breadboarded and tested the AP-filter from my all-pass simulation, version AP-C: https://atosynth.blogspot.com/2023/04/jp6-filter-allpass.html

It only required minor tweaks.

During testing, my BP outputs +/-3.5V and the HP/LP outputs +/-5.2V. I assume that BP is affected the same way HP/LP is by the resonance feedback, so adjusting feedback so that HP/LP are normally +/-5V probably also affects BP amplitude.

Anyway, to get AP we need to mix 2x BP with the inverted input. But the BP needs to be unity gain.

If we assume that unity gain is +/-5V when properly adjusted, it is 5.2 when HP/LP is 5.2. That means that to bring it up from 3.5V to 5.2V we need a gain of 1.48. The simulated circuit has a gain of 1.33 when using a 100k/33k combination. Replacing the 100k with 68k gives us the gain of 1.48 that we want. 

Here is a video of turning the cutoff CV knob, see how the phase changes.

The resulting circuit is like this:

PS: We need to sum everything BEFORE it reaches the filter as we need input both for AP and normal input. But this affects the polarity of the input and probably also the feedback circuit.

PPS: I am not sure how we can use AP/Phasing. Do we need to be able to pan wet/dry to different channels? Should this be an option anyway for filter outputs?

Update: Here is an alternative circuit, the AP output phase is 180 degrees different from the one above but it saves one op amp that can be used to sum stuff before the filter

New measurements with JOVE CV generation and working Cell 1 and 2

NB: R_abc is 10k instead of the 18k found in the JOVE circuit.

Resonance

I think the resonance looks quite good

Here are some quick measurements of the resonance voltages using the JOVE resonance circuit. This seems to give approxmiately unity gain for HP/LP and self resonance (could perhaps be a bit better).

Output of trimmer 1: -2.83V

Output of trimmer 2: -10.83V

Reso CV: 0V

Output CV mixer: 2.9V

Base voltage: -52mV

Reso CV: 5V

Output CV mixer: 7.94V

Base voltage: -10mV

Cutoff

The cutoff range seems too limited, but with this, a 50Hz wave has an amplitude of approx 130mVpp. At max, the HP filter lets a 1.5kHz wave through with 180mVpp. An 8k wave has approx 5Vpp amplitude.

Cutoff trimmer 1: -8.4V

Cutoff CV 0:

Output CV mixer: 998mV

Base voltage: 100mV

Cutoff CV 5:

Output CV mixer: -1.25V

Base voltage: -100mV

Polarities

LP output, Cell 1: inverted

BP output, Cell 1: normal

HP output, Cell 1: inverted

Amplitudes 

HLP Max: +/-5.25V

BP Max: +/-3.5V

HP Max: +/-5.25V

High frequency oscillations

The high frequency oscillations seen previously are completely gone (though I have not tested with higher CV), perhaps they were caused by the erroneous 3p filter cap

JP6 filter oscillations and latch-up

I've had serious troubles with the JP6 filter since I started working on it again about a week ago.

- it latched up whenever the cutoff CV was > 1.5V

- it had severe high-frequency oscillations

Today, I finally got it working again (with the Jove CV generation circuits). Here are the three things I fixed:

- Cutoff and Resonance Iabc has to go to two cells, not only one (or they will be too large)

- I had messed up and used a 3pF instead of a 330pF cap

- And even after fixing that, I realised that the caps were probably not in contact with the connectors. I redid the wiring there and everything started working.

Back to the real testing!