Help me mod this high pass filter (HPF pre)

[HelpingFriendly]

Hello
I just got a fdeck HPF pre high pass filter.  It is a 12+12 filter curve that is two filters in one. The first filter has a fixed corner frequency of 35 Hz, and 12 dB/octave slope. The second filter is "adjustable" from 35 to 140 Hz, and also has a 12 dB/oct slope. Thus it always has 24 dB/oct slope below 35 Hz.
I would like to mod it so that the "adjustable" frequency is greater than the default 12dB. Any help is appreciated. Here's the schematics :

[hymenoptera]

If you want to steepen the slope (18db/octave, 24db/octave, etc) it's going to require more poles. Two of these circuits in series comes to mind.

[HelpingFriendly]

Yea I thought of that. I just didn't know if I could save the space on my pedal board instead of having 2 pedals

[hymenoptera]

Right, sorry. I was thinking this was a DIY project. You want to mod an existing product.

Well, if you know what frequency you will be setting it at, you could lower the value of C1 (33nF), that's one of the fixed HPF poles right there. Not seeing the second pole of the fixed HPF...

[Transmogrifox]

You can pretty easily stick another board in there with the 2nd 12 dB/octave stage on a little "floating" perf board.  If you were willing to do SMT it would fit on your fingernail.  Even with TH parts it would probably fit in the box.

[HelpingFriendly]

Thanks 4 the reply.  So another quick question.  If I swapped out the 100k pot that controls the frequency for a higher value, would that give me the ability to choose a frequency higher than the cut off default value of 140hz?

[midwayfair]

Thanks 4 the reply.  So another quick question.  If I swapped out the 100k pot that controls the frequency for a higher value, would that give me the ability to choose a frequency higher than the cut off default value of 140hz?

You need LESS resistance from R10B+R12 OR a smaller C10&C11 (& C2 when taking signal from the drain!) to get a higher cutoff.

Simplify the lower part of it to this:



Smaller resistances will load the Q1 drain output more, and it might become more noticeable if you go below about 10K, but that would be a fairly high cutoff for a bass cut.

[ashcat_lt]

R12 sets the highest frequency of the pot.  Decreasing that will move the whole range of the pot upwards, and it's basically a linear relationship - half the value is double the frequency. 

But that's only one of the poles.  R11 would be the other.

Course, you may notice that this value is exactly 10 times that of R3 and R4, which speaks to that whole loading thing mentioned above.  This 10:1 relationship is a pretty standard "lowest safe ratio" because it makes less difference than part tolerance.  If you go smaller, you'll start to notice broadband attenuation along with the low-cut.

[HelpingFriendly]

Great info thanks. Im using this on a guitar rig that's why I would like a higher cut off. What values do u suggest to get the cutting off closer to 200hz-220hz?

[amz-fx]

Here is the original build document:

http://personalpages.tds.net/~fdeck/bass/hpfpd.pdf

The developer generously allows everyone to build for personal use, though he sells finished units at a fair price.

regards, Jack

[midwayfair]

What values do u suggest to get the cutting off closer to 200hz-220hz?

Frequency cutoff is 2*pi*R*C

... Or do what I do and use a calculator: http://sim.okawa-denshi.jp/en/Fkeisan.htm

[HelpingFriendly]

Thanks 4 the reply.  So another quick question.  If I swapped out the 100k pot that controls the frequency for a higher value, would that give me the ability to choose a frequency higher than the cut off default value of 140hz?

You need LESS resistance from R10B+R12 OR a smaller C10&C11 (& C2 when taking signal from the drain!) to get a higher cutoff.

Simplify the lower part of it to this:



Smaller resistances will load the Q1 drain output more, and it might become more noticeable if you go below about 10K, but that would be a fairly high cutoff for a bass cut.

Thanks. I was thinking just to replace R12 with a 33k variable resistor. Then I can adjust down in resistence and listen by ear until I reach a good value. This should work I assume? R11 is the one to change if I want to alter the fixed 35 Hz cut off right? If so I would like to leave that maybe and just raise the other filter.

[hymenoptera]

I believe the fixed 35Hz occurs around Q1. Notice C1 and C10. Both 33nF. Reducing them to lower values will shift the fixed HP frequency higher.

I believe that if 33nF gives you 35Hz, then 3.3nF would give you 70Hz? I'm not great with math, so someone will hopefully correct me. Double this, ten times that, I get confused easily ;P

[amz-fx]

The first stage around Q1 is ruler flat from 10Hz to 25kHz and beyond. There is no fixed 35Hz rolloff. There is about 2db loss in this buffer though.

The second stage around Q2/Q3 is a second order equal-value-resistor high pass filter (Butterworth) and the dual 100k pot adjusts the corner frequency from 35 Hz to 140 Hz. My calculations match that close enough. Low resistance on the pot corresponds to the highest corner freq (140Hz) and high pot resistance is the lowest corner (35Hz).

The high pass rolloff is 12db/oct. That appears to be the bulk of it.

regards, Jack

[Bill Mountain]

That schematic is for the HPFpre-1 which is only a single 12dB per octave variable HPF.  The HPFpre-3 is the new one with a fixed 12dB at 35Hz plus a variable 12dB HPF giving 12dB roll off above 35Hz and 24dB roll off below.

Fdeck has not posted a schematic for the HPFpre-3.

I'm being brief because I'm on my phone.  I can get more detailed tomorrow when I get near a computer but my guess is that the HPFpre-3 is 2 of these circuits in series using opamps with one filter fixed at 35Hz.

[hymenoptera]

The first stage around Q1 is ruler flat from 10Hz to 25kHz and beyond. There is no fixed 35Hz rolloff. There is about 2db loss in this buffer though.

The second stage around Q2/Q3 is a second order equal-value-resistor high pass filter (Butterworth) and the dual 100k pot adjusts the corner frequency from 35 Hz to 140 Hz. My calculations match that close enough. Low resistance on the pot corresponds to the highest corner freq (140Hz) and high pot resistance is the lowest corner (35Hz).

The high pass rolloff is 12db/oct. That appears to be the bulk of it.

regards, Jack

Thanks for clearing that up. That was my suspicion at first, too, as I can up with something like 4.1Hz or something in my calculations, I just figured I misplaced a decimal place or something.

Now it all makes sense.

I'd recommend like Transmogrofox says and add in your own filter caps, and resistors on a small PCB. Might even be able to just forgo the PCB and hang them right off the input and output jacks if there's room. Otherwise I think it's going to take a 4-gang pot or some other wizardry to make a 4-pole variable frequency filter using the existing circuit.

[ashcat_lt]

I believe the fixed 35Hz occurs around Q1. Notice C1 and C10. Both 33nF. Reducing them to lower values will shift the fixed HP frequency higher.

I believe that if 33nF gives you 35Hz, then 3.3nF would give you 70Hz? I'm not great with math, so someone will hopefully correct me. Double this, ten times that, I get confused easily ;P

It's linear.  If 33 hits 35, then 3.3 would hit 350.  You'd need 16.5 to hit 70.

[PRR]

> If I swapped out the 100k pot that controls the frequency

Plan shown has *two* pot-sections to set frequency.

If you want to go to 18dB/8ve or 24dB/8ve filtering, you will end up with 3 or 4 pot sections, all tracking.

2-section pots can be found. 4-section pots exist but are rarer.

While either R or C will change frequency, pot selection is slim and pot value was selected for general system impedance happiness. The 33K end-stoppers can't be changed much either. They should be at least 5X the size of R4 or loading is a problem (and this stage is already clipped on very strong guitar).

OTOH you can change the C values without changing anything except Frequency. Want it to work an octave up? Half all the caps.

> Double this, ten times that, I get confused easily ;P

So we see.

[Transmogrifox]

If you want to go to 18dB/8ve or 24dB/8ve filtering, you will end up with 3 or 4 pot sections, all tracking.

Although I wonder if tracking is necessary.  The main reason for tracking on a single 12db/8ve is to maintain the butterworth response on that filter section.

Say, you cascade another one of those 2-pole sallen-key butterworths with just another 100k dual-ganged pot, drill another hole in your enclosure and then you have a significantly improved control over where you transition from 24 dB/8ve to 12 dB/8ve to flat.

For example, you could set one down at 80 Hz and the other up a couple octaves at 320 Hz and then you have a steep slope below 80 Hz, and not so steep slope up to 320 Hz.

With un-coupled pots you would have the flexibility to do this, and I think it would be more useful IMHO.

For guitar to shift the whole thing up to useful frequencies:
Change R11, R12 to 10k
Change C10, C11 to 22n

A simulation shows this will give you a range of 62 Hz up to ~720 Hz.  Those are simple parts-change mods you can do without having to add a little board.  That's probably the first thing to do and then decide if you actually need that extra steep rejection.

Then another simple mod to add a fixed 6 dB/8ve HP is to change the capacitor at the input to a smaller value.

[ashcat_lt]

For guitar to shift the whole thing up to useful frequencies:
Change R11, R12 to 10k
Change C10, C11 to 22n

You'll get increasing broadband attenuation almost to -6db when the pot is at minimum resistance...