How can I adjust the Afterlife's compression parameters?

[solderfumes]

Hi all,

I'm pretty new to building pedals and have been lurking around the forum for a while taking stuff in.  I got started through the DIY club on tonefiend.com and have built all five of the entry level projects (Electra distortion, Bazz Fuss, FET clean boost, Rangemaster, Fuzz Face), and have just finished building a five-knob Orange Squeezer, with mods lifted from Mark Hammer's Tangerine Peeler and GGBB's Orange Smoothie, and with some advice/explanation from Mark Hammer via personal message.

Next up on the docket, and currently on the breadboard, is an Afterlife/Flatline; actually, it's a Ghost Note, as I copied Jon Patton's threshold control mod.  Because I'm a sucker for punishment, and because I'm trying to learn all I can from each circuit, I'm attempting to make a five-knob version, trying to figure out what I can do to adjust attack and release.  The Afterlife already includes a "sustain" control but it seems to me that that's really a ratio control.  Jon kindly answered some of my questions via personal message and suggested the following:

• replace R6 with a small-valued pot to adjust the attack
• add a 1M pot (plus series resistance) across C3 to shorten release

I replaced R6 with a 1K trimmer, and it seems quite effective.  I haven't tried #2 yet, but I did try adding a trimmer in series with vact_1A (the LED) to lengthen the decay, and it seems to work.  Jon mentioned that the release in the Afterlife is already quite long, but I've noticed that the LED pretty much instantly turns on and off (I don't have a Vactrol so I rolled my own with shrink tubing, an LED, and an LDR, which allowed me to see how bright different types of LED got before they went into the shrink tubing) unless I turn the trimmer up, and the release feels pretty quick to me unless I do that.  In fact, I can't really tell any visual difference in the behaviour of the LED if I just straight-up remove C3 from the circuit.  The LED still just lights up when I play and turns off when I don't.  Thus, my questions:

1) If I'm not misunderstanding the circuit, R6 controls how hard the signal gets "pumped" to C3, much as the 1k5 resistor before the diode in the Orange Squeezer controls the attack in that circuit.  But, that's also what adjusting the gain of IC1_B does (i.e. Jon's threshold control), albeit from "the other side" of the signal.  So, are the effects of these adjustments completely conflated, or do they actually have any distinct effects?  It does seem to me that adjusting R6 changes the attack while changing R7 adjusts the threshold, anyway, but I would need to fiddle around with it more to try and identify if they're doing the same thing.

2) Should I be worried that there is by default no resistance protecting the LED?  I'm not sure of the kinds of voltages we're feeding the LED, but from my rudimentary knowledge of diodes and Kirchhoff's laws, I don't quite follow why we don't see an unbounded current across the LED when the voltage of C3 exceeds Fv.  Moreover it seems like C3 is either instantly dissipating its charge across the LED if I don't put in the series resistance, or just not doing anything.

As a side note, I found that I'm getting really good results (i.e. strong compression) using what I think is an MEC MLG5547 LDR (http://leeselectronic.com/index.php?id_product=7363&controller=product&search_query=photo+resistor&results=11).  I started off with an old LDR (marked NSL-32 CANADA, if that means anything to anyone) that I got from a friend, and thought "really subtle effect, can't quite tell if it's on but it seems nice" but then when I tried the MLG5547 it clamped down really hard.  I can't compare it to a Vactrol VTL5C3 but the datasheet tells me that its turn on time and turn off time are pretty snappy (20ms and 30ms respectively), and its light and dark resistances are in the same realm as the VTL5C3.  So don't lose hope if you're rolling your own!  Actually, that brings me to another question:

3) The datasheet says I should be using a green LED, but my green LEDs aren't even close to as bright as my orange ones, so I'm getting better results with an orange one.  I looked it up and saw that green LEDs have a very wide range for Fv; does that mean that their turn-on voltage also varies?  Should I shop around for a brighter green LED?

Apologies if these questions have been addressed somewhere that I haven't already looked!

[solderfumes]

After thinking about it a little bit longer, I think I may know the answer to question 1: the signal coming out of the inverting op-amp doesn't actually build any charge in the capacitor, per se; it simply creates a voltage difference relative to the signal coming in through R6.  Thus R6 really does control the current flowing into C3, and thus controls how fast the charge accumulates in C3, while R7 increases the amplitude of the voltage appearing across C3.  Am I way off base here?

[solderfumes]

Actually, let me rephrase that -- I should really stop using the word "signal", haha.

The voltages produced by the outputs of both op amps create a voltage difference across C3, and R6 and C3 are connected in series between the voltage difference.  Thus the magnitude of the voltage difference (dictated by R7) determines the voltage achieved by C3, while the current (dictated by R6) determines how fast it gets there.  Does that sound right now?

[solderfumes]

I thought of another question: I notice that there are two power-filtering capacitors, C4 and C6, of values 100uF and 100nF respectively.  I've seen this in other circuits as well.  Shouldn't their capacitances simply add up?  Is 100.1uF really that different from 100uF?  Something else must be going on; what is it?

[solderfumes]

OK I think I may have figured out #2 as well, and also why removing C3 doesn't seem to make any difference.  I had thought that C3 was responsible for powering the LED, but I'm now guessing that C3's only role is as a smoothing capacitor in the rectifier stage; the voltage presented to the LED is actually a heavily rippled DC voltage, and C3 smooths out the ripple.  R6 protects the LED because it controls the current flowing through the diode network.  Removing C3 just feeds the LED the raw, rippled DC voltage, with current still being determined by R6.  Maybe the ripple is just fast enough that I didn't notice it when I took out C3.

Please do correct me if I'm completely wrong though!  I promise I didn't set out intending to talk this through with myself in a public forum.

[midwayfair]

I thought of another question: I notice that there are two power-filtering capacitors, C4 and C6, of values 100uF and 100nF respectively.  I've seen this in other circuits as well.  Shouldn't their capacitances simply add up?  Is 100.1uF really that different from 100uF?  Something else must be going on; what is it?

The 100nF is a film cap. Film caps are better at filtering high frequencies than electrolytic caps are. It's the two different dielectrics that matter, not the values.

OK I think I may have figured out #2 as well, and also why removing C3 doesn't seem to make any difference.  I had thought that C3 was responsible for powering the LED, but I'm now guessing that C3's only role is as a smoothing capacitor in the rectifier stage; the voltage presented to the LED is actually a heavily rippled DC voltage, and C3 smooths out the ripple.  R6 protects the LED because it controls the current flowing through the diode network.  Removing C3 just feeds the LED the raw, rippled DC voltage, with current still being determined by R6.  Maybe the ripple is just fast enough that I didn't notice it when I took out C3.

Please do correct me if I'm completely wrong though!  I promise I didn't set out intending to talk this through with myself in a public forum.

This is correct. The guitar is not a constant current or voltage source, so you have to use a capacitor's charge to smooth out the variations in the signal or you'll get pumping. The vactrol itself does some of this as well.

There are other places you'll see things like that 100uF cap -- they're in AC circuits. Pull up an older amp schematic, and I think some older EHX units with their own power chords. If you really want to get deeper into what's going on with the rectifier circuit in this, look at bridge rectifier circuitry. Also, have you read John Hollis's original article? It's sparse, but it might still be some help. The Philosopher's Tone might have some ideas for you as well -- that uses a 2200uF cap!

If you find a brighter green LED, let me know. The ones I have suck.

[solderfumes]

Ah okay I didn't know that about the different types of capacitor.  I don't have film caps so I substituted a ceramic, but I'll have to do some further reading on capacitor types.  Does anyone have any recommendations on the literature?

So what is happening when I add the series resistance to the LED?  It definitely changes the decay of the LED's brightness, but I don't understand why anymore.  And what will happen when I add the 1M resistance in parallel with the LED?  I had thought both of those mods would work by affecting the discharge rate of the capacitor, but now I'm not sure.

[midwayfair]

So what is happening when I add the series resistance to the LED?

Voltage is stored on the 100uF cap. The voltage is used by the LED.

To light the LED, current must pass through it. For current to pass through it, voltage must be drawn from the + side of the 100uF cap through the LED's anode to its cathode.

The LED is the only drain for the capacitor besides the capacitor's ability to hold a charge itself. Obviously, the LED conducting is a far, far quicker method of draining the capacitor than letting the cap drain all on its own.

When you put resistance in series with the LED, less current is drawn. You probably didn't notice using your eyes, but the LED lights up less brightly. So while it increases the decay, you're also reducing the amount of the effect, so there's probably a tipping point where the series resistance stops being useful.

If you put resistance in series with the cap, it would charge the capacitor less readily, as well as limit the current that could be drawn from the cap. Part of why I suggested not doing that.

If you put resistance in parallel with the cap, charge will leak through the resistor more readily than through the capacitor on its own. But you're right ... there might not be much of a difference if the LED itself is draining the cap so quickly.

This is one of those cases where the circuit wasn't necessarily designed to do much more than it already does.

[solderfumes]

Voltage is stored on the 100uF cap. The voltage is used by the LED.

To light the LED, current must pass through it. For current to pass through it, voltage must be drawn from the + side of the 100uF cap through the LED's anode to its cathode.

The LED is the only drain for the capacitor besides the capacitor's ability to hold a charge itself. Obviously, the LED conducting is a far, far quicker method of draining the capacitor than letting the cap drain all on its own.

Right, yes, okay.  I was caught up in thinking that the voltage was entirely coming from the rippled DC voltage, forgetting that the capacitor also acts as the voltage about half of the time.  I get caught up not considering the circuit as a whole sometimes.

When you put resistance in series with the LED, less current is drawn. You probably didn't notice using your eyes, but the LED lights up less brightly. So while it increases the decay, you're also reducing the amount of the effect, so there's probably a tipping point where the series resistance stops being useful.

Oh yes definitely it's noticeably less bright when the release trimmer is adjusted.  When I was playing around with it, I'd use the threshold or ratio controls to make the LED light more brightly.

I guess the threshold control and my attack trimmer are highly interactive, although not completely overlapping in effect -- the threshold control is needed to adjust the voltage high enough to exceed Fv of the diodes, while the attack trimmer has a more direct effect on the current.  I will probably still try the resistance in parallel with the cap, though as you said I suspect it won't do much good.

I imagine that, as you said, the circuit isn't designed to allow these kinds of adjustments.  I'm probably just still riding high on the success of my Orange Squeezer, haha.  Still, it's fun to try and there does seem to be at least small amounts of adjustability of the usual envelope parameters.  I'm just glad to learn the ins and outs of different kinds of compressor circuits.

[solderfumes]

There are other places you'll see things like that 100uF cap -- they're in AC circuits. Pull up an older amp schematic, and I think some older EHX units with their own power chords. If you really want to get deeper into what's going on with the rectifier circuit in this, look at bridge rectifier circuitry. Also, have you read John Hollis's original article? It's sparse, but it might still be some help. The Philosopher's Tone might have some ideas for you as well -- that uses a 2200uF cap!

I Googled but didn't find the original article about the Flatline.  Does anyone have the link?

[kothoma]

I Googled but didn't find the original article about the Flatline.  Does anyone have the link?

You've probably found this:
http://www.hollis.co.uk/john/circuits.html  (http://www.hollis.co.uk/john/flatline.jpg)

I don't know it there's more from John Hollis himself. But here's some additional information:
http://www.aronnelson.com/gallery/main.php?g2_view=keyalbum.KeywordAlbum&g2_keyword=john+hollis+flatline+compressor+vero+layout+ulysses&g2_itemId=6537
http://www.geofex.com/PCB_layouts/Layouts/flatline.pdf

[solderfumes]

Thanks for tracking those down.  I had seen most of it, as you guessed, but hadn't read the description by Ulysses.

Having played around with the circuit a bit more tonight I'm actually reasonably pleased with my release control (a 10K variable resistor in series with the LED).  When the release is lengthened (i.e. pot is turned up), it does reduce the overall brightness of the LED and therefore the overall amount of compression, but it's still quite usable and gives me the option of a little bit of duck-and-swell, which I like from a guitar compressor.  Moreover, I'm finding that the attack, release, and threshold controls aren't quite as conflated as I thought they were, with all of them giving decent adjustability of their respective parameters regardless of the others' positions.

I also tried Mark Hammer's Punchline mod to turn it into an expander, and might wire that into my pedal too.  Seems like it might function as an interesting booster in certain capacities.  Gonna be a crowded enclosure!

[thesameage]

For anyone wondering, here's how you do the Ghost Note mod:

Connect lugs 2 and 3 of the pot. Solder one side of the resistor to lugs 2/3, and the other side goes to the PCB. A wire from lug 1 of the pot goes to the other resistor hole on the PCB. Order doesn't matter. You could use a 100K pot without the limiting resistor (or even larger) instead if you like, but I found the 50K + 24K gave plenty of range.

I used a linear so that noon was the original value, but feel free to deviate!