madbeanpedals::forum

Hey there-
First post, probably dumb question.

What I'd like to do is elegantly modify two Afterlifes to give myself dynamic control over Bass and Treble.
So- one AL just doing highs, another just doing lows, and running in parallel.

I figure there must be some sneaky places to insert a couple caps so they wont interact across the units-

It seems like Highpass would be easy, just change the value of C1 in one unit, or add a cap coming off of pin1 of IC1a, filtering it's lows without affecting the other compressor..
But for Lowpass- Can I get away just shunting the highs to ground on the positive side of C2? Or will I have to use the negative feedback loop of IC1a somehow?

Also, if that works- I should be able to make a mid control.. yes?

I have an AL, and I'll poke around with that this week, but any thoughts would be welcome-
Thanks!

Put a FET or BJT unity volume gain stage in front of both, with no source/emitter bypass cap. Run your Afterlifes directly into the source and collector. They've already got the decoupling caps. The emitter and collector will both send a unity volume signal into the Afterlife.

Solder a 2.2nF cap across R3 in the "bass" Afterlife. Make the input cap 10nF.

Change the output cap in the "treble" Afterlife to 22nF. You're changing the output cap on this one because it gives the compressor full signal and

Remove the volume pot from either one and simply connect the pad for lug 3 from that one to the lug 3 pad of the other. The sustain pot will now be your gain control for both, and the volume is a master volume.

These are theoretical changes, but the calculator says you should shelve the frequencies pretty high up. There should be enough crossover to keep things from sounding weird or scooped in the mids. I'd need a scope or sim to do more than make an educated guess.

There is one problem that MIGHT exist, and that's that the first stage puts the signals out of phase, and consequently I'm not sure if you need a buffer inverting the emitter follower back to "normal" to avoid cancellations.

Before you get too excited, though, a lot of people have tried parallel compression and often come to the conclusion that it's simply not worth the headache. Usually people just want more treble back in their compressed tone and the easiest ways to handle that are parallel compression (i.e., a blend circuit) or treble bypasses of various stripes -- the Engineer's Thumb has an excellent treble bypass mod that boosts the treble by leaving it untouched, and the Bearhug has a treble bypass that also functions as a treble cut (depending on whether the compressor is working very little or working very hard). The treble bypass in a Ross is less satisfying, since it just adds a little presence at the output, after the compression has already happened.

Hey Jon, thanks for the thoughtful response- Your demo videos are what got me on the madbean site in the first place-

You may have answered my question-

I should probably explain more what I'm going for, apologies if I sound redundant.
I'm a bass player, and I want a 2band dynamic eq. It would be my compressor, and my eq. So I'm not just looking to regain my highs, I want to compress and control each band separately. I need both Compression knobs, and both Volume knobs, and I would probably add your sensitivity control to each side as well.

There's quite a few units that do something like this, and I have a couple, but they're mostly designed more for general PA use than an individual instrument(I don't need a lot of range of controls, I'm never going to put a vocal through it). I know of two bass specific pedals, one with a million knobs and another with none. For a bass compressor(or eq), I just want a few knobs that are musical and to fiddle with a few component values until I'm happy with the way it works- which is why I like the AL.

From what you're saying, I need an extra stage in front of each compressor, perhaps there's a better compressor build for this to start with? I really like the opto, though- I'll have to research what you're talking about with emitters and followers before I attempt to ask a real question about that-

Thanks!
Ben

Quote from: btml on July 23, 2013, 02:38:02 AM
From what you're saying, I need an extra stage in front of each compressor, perhaps there's a better compressor build for this to start with? I really like the opto, though- I'll have to research what you're talking about with emitters and followers before I attempt to ask a real question about that-

OKay ... a source (FET) or emitter (BJT) follower means that you take the output from the emitter or source instead of the drain (FET) or collector (BJT). A simple buffer is an emitter follower:
http://www.muzique.com/lab/buffers.htm
(http://www.muzique.com/images/buff2.gif)

This will be unity gain as long as the transistor is biased correctly.

A basic transistor amplification stage will come from the collector or drain. Take a look at, say, the Big Muff schematic.

We can actually take the output from BOTH at the same time. They will not interfere with each other this way. It's the simplest possible method of getting a split output that I've found, since it just involves three resistors, a capacitor, and a transistor. Here are a bunch of things that use this trick in-circuit:

AMZ Mosfet booster with second output:
http://www.muzique.com/schem/mosfet5.gif

My Hamlet delay -- the delay line comes from the emitter while the dry path comes from the collector:
https://docs.google.com/document/d/1P7uOYFRH6ZW5eGoquhAiMQMyUU76bod_gNKkiSlPNJM/edit

My Blue Warbler -- imagine an Afterlife in place of the envelope section:
https://dl.dropboxusercontent.com/u/9878279/Jon%20Patton%27s%20layouts/Nonmini%20builds/Blue%20Warbler%20Vibe%20schematic%201.2.png

You can, of course, make a TL072 input buffer and connect both circuits to pin1. In my experience this loses a little bit of input gain, which is a critical consideration for reducing noise in a compressor design, and you might get some sort of parasitic crossover. It's probably worth testing -- it's really easy to set up on a breadboard, since it's just a chip and a voltage divider network.

I don't think the question is whether there's a better compressor to "start" with. I'm sure there are some DIY dual compressor designs, though I'm not aware of any off the top of my head that are known to be good ... perhaps you might be the guy to get them working? :) At any rate, since I'm a compressorholic and I can't think of one off the top of my head (and haven't built it), it's safe to say it's going to take some digging ...

I'll try to explain further why you might want to use only a master volume instead of individual volume controls. You can use the sensitivity (threshold) control and "compression" control to mimic the two stock controls on the Afterlife: The "compression" control is actually just the gain of IC1A, so if you turn down the sensitivity, you will get less compression at higher gain settings. Turn up the sensitivity and you'll get more compression at higher gain settings. This lets you use the gain control as a volume for each band.

The alternative is much more complicated. You'll need two more transistor or op amp stages, because you can't put the two volume controls in parallel with each other. So you have to run each volume control into a new buffer, then connect the outputs of the buffer. There is a benefit to doing this: You'll get super low output impedance, and the afterlife doesn't have the best possible output impedance at 100K.

If it were me designing this, I'd simply build two Afterlife boards with Comp and Sense controls (and otherwise as I described it above), then make the MOSFET booster as shown without C5 or R6. Put the Treble Afterlife on the Drain side of thing and it will be slightly boosted (~3 dB) -- that should help even things out compression-wise. You can also omit the decoupling capacitors (C4 and C6) because they'll be part of the Afterlife PCBs already.

That's why Jon rocks: explanations that are education.


Sent from my Nexus 7 using Tapatalk HD

Quote from: btml on July 23, 2013, 02:38:02 AM

I should probably explain more what I'm going for, apologies if I sound redundant.
I'm a bass player, and I want a 2band dynamic eq. It would be my compressor, and my eq. So I'm not just looking to regain my highs, I want to compress and control each band separately. I need both Compression knobs, and both Volume knobs, and I would probably add your sensitivity control to each side as well.
[...]
Thanks!
Ben

Hi Ben.

So; you want to split your signal in two bands - then run the two bands independently from eachother through some compressor(s) - and then mix them together again.

Now i havent read through Jons complete posts indepth (i will, just that its breakfast here), but i would start to play with a Linkwitz Riley phase compensated filter (google it). Because you have to take care of the phase shift artefacts in some way, and the best thing must be to compensate directly in the splitting filter stages. 

Take a look at my Blue Monster Project: http://www.madbeanpedals.com/forum/index.php?topic=6968.0 , where i do something similar, just that i run the separated bands through overdrive Circuits, not compressors. You may borrow the complete idea from there, i guess you wanna re-calculate Component choices for "your" split frequency though, as you aim for bass.

Cheers
/Bengt

Cool- I understand that-

I happen to have a karate shop laying around, which has a buffer in it- maybe not the right transistor for this application?
I was trying to use that effect before compression as a semi-parametric low/mid booster.. but of course that combo gets super noisy.
Now I'm wondering if I can (boxing them together) use the KS buffer in front of the AL's (in the way you described) and feed the compressor output into the KS post buffer...
KS transistor is a 2N5457, but I suppose I could change it out with a more appropriate part if necessary.

Seems like the buffer for the KS would be superfluous given the output of the AL, and the fact that I don't intend on switching parts of this contraption on and off. But I could be wrong?

You really should take a look at http://diy.thcustom.com/shop/pro-cessor-pcb/

And here's a rough picture of that idea- sort of- just wanted to make sure I understand.
Forgive me for just chopping up those schems, I don't have a drafting app.

Buffer from KS split into 2ALs, with Jons cap suggestions, sans side chain, one master volume, into the rest of the KS. I left out all the power sections-

Jon, your cap locations, I'll have to filter the sidechain audio separately if I don't want them to track full range, yes?

Bengt- I just read yours but will check it out as well as I'm heading to bed over here-

Cheers- Ben

Don't think this will work as expected. Even if you'd get that FET stage right (at least one resistor missing) then treble and bass will be out of phase? Mixing them at the VOL pot should create a nasty mid notch. Just my intuition, haven't tried it.

Here's a completely different approach:

Take an SVF (like in the Nautilus/Mutron or the Bloviator or the Boss Enhancer or Anderton Super Tone Control) with a fixed center frequency (and fixed low Q). Use vactrols to control the mix levels of the HP and LP outputs.

I confess I'm a big fan of SVFs.

Quote from: kothoma on July 23, 2013, 08:05:07 AM
Even if you'd get that FET stage right (at least one resistor missing) then treble and bass will be out of phase? Mixing them at the VOL pot should create a nasty mid notch.

That's what I was worried about ... this is tough on pre-existing PCBs.

btml, I think no matter how you go about it, this is going to involve some breadboarding.

Yeah, if it were an effect I would probably try it with a nasty notch, but I want to be able to leave it on- Then again, the notch filter on an acoustic amp is apparently how Albert Kind got that weird sound-

Found this on one of the above links-
The Mad Box http://diy.thcustom.com/shop/madbox-v3-1-pcb/ (http://diy.thcustom.com/shop/madbox-v3-1-pcb/)
It seems like it would be the simplest approach for me if it works. It's just a two channel effect send, and a two channel effect return with a blend knob. I would put the Afterlifes(with mods to bandpass them as above) one in each loop.

Does that sound feasible? And does Mad Bean have anything like this?

Quote from: btml on July 23, 2013, 08:57:10 PM
Does that sound feasible? And does Mad Bean have anything like this?

Sorry if i keep repeating myself, but you have to deal with / solve the phase shifting dilemma caused by filtering, and thats maybe not trivial. Or accept it as it is without phasecompensation, but then you will have artefacts at the crossover frequency, which maybe will be cool, but possibly sound like ****.

For the splitting and summing part of the circuit, THcustoms pcb (as in your link) looks clever though:)


Cheers

Vallhagen-
So I (sort of) understand now how the Linkwitz-Riley corrects the phase, with two lines of two stages which turn 90 degrees in opposite directions, and then 90 degrees again. I don't understand why there is a phase shift of this type in the first place, but there's probably some math involved?

That makes the Pro-cessor (also linked to above) a better place to start to see how this idea works-
I was starting to wonder if I will like the notch, but I can always put one in later.

But how about a different approach-
Could I use an Afterlife's Vactrol as the variable resistor in a tone control instead of for controlling the afterlife's gain?
Because I really like Ampegs use of the Baxendall in their B15, and I was going to be shooting for that voicing anyway-
If I don't need the resistor side of the vactrol hooked up as drawn in order for the afterlife to control the vactrol, would I be in business?

Except that it would then be an expander.. brain is scrambled.

Quote from: btml on July 25, 2013, 05:17:52 AM
Except that it would then be an expander.. brain is scrambled.

It would depend on how you hook up that vactrols. You can get expansion or compression in the resp. bands.

But for that Baxandall you would need 4 vactrols and the pairs would need to be complementary, i.e. (Rpot - Rldr) + Rldr, if that makes sense.

I was just looking at that- It would be a nightmare to tune two vactrols to act like two halves of a pot-

I could compress the full signal conventionally and get slightly more compression on highs by using the vactrol as a tone knob, essentially.

Too many ideas, need breadboard.

Quote from: btml on July 24, 2013, 05:59:19 PM
I don't understand why there is a phase shift of this type in the first place, but there's probably some math involved?

Yep, math is involved... but math is - in this case - "only" a way to explain physics/reality.

Verrrry simplified: Each time* you put a C or L (in this case, C) and some R in a signal chain, you will cause a time delay. And time delay=phase shift, just expressed in different units chosen by context. This time delay is related to the time it takes to charge/discharge a capacitor in a given circuit.

...so what happens here in your filtering is that the LP part of the circuit cause a time delay, and your HP-circuit cause another time delay. And the difference between those time delays will, at one specific (the crossover-) frequency, equal to a 180 degrees phase difference, causing a complete destructive interference when you mix them to eachother.  When You add a perfect positive peak to a perfect negative peak the result will be nothing...Just as 1+(-1)=0.

My reasoning here is based on the Linkwitz Riley 12dB/oct filter. Funny: Now as i see it and type it, the further developed (better!) Linkwitz Riley 24dB/oct filter must cause a 360 degrees phaseshift between the LP and HP sections. And 360 degrees means... we are in phase again, just that the LP is a cycle late (and we can live with that). No dip, no notching!!

So, yea, this takes me back to square one, THcustoms Pro-cessor is probably exactly your thing. I encourage you to try it:)

hehe, you see i changed my mind?

Cheers
***



**********************
* okokok, maybe not EACH time, but again, simplified example...

Quote from: Vallhagen on July 25, 2013, 08:24:06 AM
Yep, math is involved... but math is - in this case - "only" a way to explain physics/reality.

Math is always a way to explain reality :)

Quote from: alanp on July 25, 2013, 11:01:11 AM

Quote from: Vallhagen on July 25, 2013, 08:24:06 AM
Yep, math is involved... but math is - in this case - "only" a way to explain physics/reality.

Math is always a way to explain reality :)

At least it allows to make models of reality that can be used to make predictions.
Not sure about explaining anything.

Yeah- I've tried doing more advanced maths and surprised myself- But the only way I can do it is to find a visualization that jives, otherwise I'm lost.

I'm definitely going with the Pro-cessor so I can have conventional results, but I also want to see where this other idea goes cause it would invariably sound different than anything else out there. Two Builds-

So, the Vactrol controlling a tone knob thing-

In the AL, the LED idles at low (or off?), and when a loud enough signal is applied, the LED brightens lowering the resistance in IC1a's negative feedback loop, and this turns the loud notes down.

At least that's how I understand it-

Instead of resisting less in a feedback loop to turn down the signal, Id like to resist more in the signal path.

Is it possible to have the LED idleing high and send it less voltage when a loud signal happens? I don't see it in any schematics for compressors... Is that just because it would eat batteries, or is it not possible?

Cheers~

Quote from: btml on July 25, 2013, 08:49:50 PM
Instead of resisting less in a feedback loop to turn down the signal, Id like to resist more in the signal path.

You could use it as is resisting to ground, like many tremolos do.

Quote from: btml on July 25, 2013, 08:49:50 PM
Is it possible to have the LED idleing high and send it less voltage when a loud signal happens? I don't see it in any schematics for compressors... Is that just because it would eat batteries, or is it not possible?

You could flip the curve around with simple op-amp-mathematics, that's what they were invented for anyway. But there could be simpler methods. Sure it would need more current that way.

I can't use resistance to ground with this particular eq-

I guess it's the math for me.

I had the thought that if I were to go nuts and replace the pots with pairs of LDRs, A) they could do the compressing, but I could use pots to control the 'idle' voltages, and they would essentially set the default tone knob values. Which I think is spiffy.
And B) Id never make it work.

I still have trouble trying to understand what you want to achieve.
I know what multiband compression does to a mix.
But I can't quite imagine what compressing highs and lows separately on a single note would sound like, especially on bass. And mids stay uncompressed?
What would be different from mixing an eq'ed and compressed signal with the dry signal?

Kothoma- It's just multiband, but as it relates to the bass guitars spectrum and my playing style.

I don't like much compression anyway, but I do like the option of slightly more compression on my highs.
Say, if I wanted to slap or do something percussive- I like to have balanced tone, but that means when you do anything percussive, those things just poke out and clip more in the higher frequencies than I would like(especially if I've been DI-ed into a PA, and the vitamin root perm sound man assumes I want to sound like a machine gun).
As for the lows, I like to palm-mute a lot and go for a lower frequency percussiveness- as in mimicking the shorter attack and quicker decay of a note played on an upright. I'm pretty good at getting that bouncy swingin feel from my notes, but if I compressed that like I do the highs, it would get all smoothed out.

So- I use multiband compression to support the range of playing techniques I use. Does that make sense? I'm just looking for a version of that which is tailored for what I do rather than have 17 knobs.

The mids aren't a separate band to me- that would be going a bit nuts. On a conventional multiband I would keep mids with the lows- matter of taste I suppose. The baxendall idea is a whole different thing though, I'm not sure where that will take me.

Actually- I have tried eq + comp in parallel with dry- Read about Motown doing their vocals like that. I always felt like it was working against me though-

Quote from: btml on July 26, 2013, 09:23:46 AM
I do like the option of slightly more compression on my highs.
Say, if I wanted to slap or do something percussive- I like to have balanced tone, but that means when you do anything percussive, those things just poke out and clip more in the higher frequencies than I would like(especially if I've been DI-ed into a PA, and the vitamin root perm sound man assumes I want to sound like a machine gun).

So in essence you want a limiter for percussive playing. But you only want to limit the highs? But aren't the highs mostly in the attack anyways, so a limiter should do? Or is it really compression you want?

Quote from: btml on July 26, 2013, 09:23:46 AM
As for the lows, I like to palm-mute a lot and go for a lower frequency percussiveness- as in mimicking the shorter attack and quicker decay of a note played on an upright. I'm pretty good at getting that bouncy swingin feel from my notes, but if I compressed that like I do the highs, it would get all smoothed out.

You want a milder compression for that.

OK, I get basically you want to be able to play funk slap bass and palm muted jazz walking bass (or similar) with the same compressor setting? So you don't have to switch between two different compressor settings?

Exactly- except for limiting, I always felt that's the nuclear option.

Some use two compressors, and A/B between them-
One could compress the full signal, then compress the highs more-
You could do the motown parallel trick you described-
Some try sidechaining an eq-

Tried it all and I prefer the multiband approach- except that multibands are pretty much all 4 knobers, and those things never sound as musical to me as the simpler circuits like you see in ancient compressors like the Altecs, or in certain guitar pedals which we could geek on for days.

But yeah, it's all about having a setting that I can forget about and just play-

OK then, for two-band compression the easiest way should be to go with a Pro-cessor.
You only need to find the right corner frequencies.

This will yield 5 knobs, but you probably can eliminate the blend pot.
(I guess it would be more tricky to dial in your setting if you throw out the level pots.)
Still 4 knobs.

Tons of good information posted today about making multiband comps:

http://www.diystompboxes.com/smfforum/index.php?topic=103800.0;topicseen