Bearhug Compressor - schematic, demo, layouts, & forthcoming 1776 PCB!

[midwayfair]

I was tinkering with the Bearhug and wanted to see if I couldn't maybe get a wee bit more out of it. I found that making it compress more revealed a bit of a limitation in the attack, and further limitations when placed after an overdrive.

Anyway, these values will make it more a traditional compressor and less of just a 'boost with compression.' Most of the same sounds should still be in there, but you won't be getting above unity volume while compressing at max compression. (Of course, you can just turn down the compression.)

If you're happy with your build, there's probably no reason to change it, but if you weren't happy, you might try these mods. This won't make it a different pedal, just make the comp control have a much wider range.

Changes:
R14: 100R-100k. This is the only change I think is essential, and leaving it at 1M was really a design flaw on my part, since it seems to increases the attack time enough that the low setting was doing too little on all but the highest threshold settings. There's not much difference between these values, just whatever you have on hand. This change also makes the low Fv diode choice a little less critical.

Remove R12. This gives you back the mild/boost settings. I'd also recommend a  reverse log for the comp control if you have one. This will keep it from bunching up.

Optional changes:
R4: increase to 4.7k. This boosts the output of q1, and I've suggested it on enough tech help threads that I think I should just make it standard. You can go higher, but there shouldn't be a need unless you use really low output pickups.

R3: I've been making this 4.7k on some builds, but I don't think it's really necessary with the R14 change, again unless you have low output pickups and you need a 'kill' setting.

All part numbers are on the 1776 schematic.

[midwayfair]

Summary of changes:
-Change R4 to 4.7K (should bias to ~5V)
-Remove R12
-Change R14 to 1K
-Use generic silicon diodes for D2 and D3.
-Use a 1KC for the comp pot to keep it from bunching up (because R12 was removed)
-Optional (for low-output pickups): Change R3 to 4.7K.

Why these changes are good:
-More reliable compression with a wider range of pickups
-Smoother sound overall.
-Slightly more output/boost
-Er ... a few cents cheaper to make!

Okay, I promised Josh a slightly more thorough round of testing and explanation before he adds this to the build doc.

I built a couple boards yesterday to test against my old stock version and did an a/b/c as soon as I got back from work.

Version A - old stock version.

Really only worked when it was just barely turning on. With my single coils, this was pretty much at the top of the compression range. It felt stiff. Also, the total output was less than the other versions here, even with the same perceived amount of compression. I felt like this had only one setting that worked with any given guitar, and that setting was never one that offered anything other than just a touch of compression.

Version B
This is on my development board, which has sockets for All The Things -- I set that up as described yesterday with all four mods (and R14 = 100R) to get the maximum squishiness.

This was the opposite. I felt like it was simply TOO easy to trigger. This meant that my minibuckers were sounding best at 0 or 9:00 on the Comp knob. It just felt too rubbery and extreme. Adjusting R3 back down to 2.2K or even lower really didn't back things off quite enough to use it as an almost pure boost. Max compression settings were essentially useless: it would compress even softly played notes down to unity volume, but because the decay is short it sounded like breathing. Ripple also became noticeable with higher output pickups. I can't imagine any pickups being weak enough to need these settings.

Since I had sockets, I tried a few different things and ended up with version C ...

Version C - new version
R14: 1K, removed R12, R4 = 4.7K (biases to ~5V), 1N4148 diodes (these could be 1N4001 or whatever).

This works out as a nice compromise. It has a sweet spot with my single coils between 3:00 and max, where it sounds very natural most of the time, and it's able to get above unity volume at any setting. The output is much higher than A overall with the comp backed off but it's still compressing even on the lowest settings with single coils. With humbuckers, it turned on reliably closer to noon but never felt out of control.

The nice thing about doing it this way is that if you WANT the squashed settings, there's room to go upward with minimal changes: increase R3 or even put the Schottky diodes back in.

-------------

Okay, so what's the deal with R14 and why was it 1M in the first version?

I've mentioned that the linearized FET set-up in Bearhug (everything between Q2 and Q3) was lifted wholesale from the Rothwell Lovesqueeze. Apparently, in the Lovesqueeze, the equivalent resistor to R14 is needed because it forms a voltage divider with the equivalent of R8 to split between the audio and control voltage.

In the Bearhug, that's unnecessary. The rectifier is fully decoupled from the audio path and no noise bleeds in. All it's doing is blocking the voltage needed to drive the FET reliably.

---------------
While I'm here ... some mods

-Hey! Anyone want a third "even longer" decay setting?
Use a SPDT on-off-on. Wire a 220K to lugs 1 and 3. Replace R10 with a larger resistor, 220K (for a 1s decay). Wire the switch as normal. In the center, R9 and our new R10 are in the circuit all the time = long decay. Flipped so that lugs 1 and 2 are connected, R10 is bypassed (normal short setting). Flipped so that lugs 2 and 3 are connected, a 220K resistor is put in parallel with R10, giving 110K total, for a decay of ~500ms, which is almost the same as our old "long" setting. Is this useful? I dunno. Might be for some people and it's very easy to do!

-I think how I did the emitter on Q2 was unnecessarily complicated. I set it up as if it were another MOSFET driving that stage, but ended up using a BJT because they're cheaper and the transistor characteristics didn't matter much. You could jumper C10 and leave off R13 if you wanted to save a couple parts.

[Mbas974]

Hi, thanks for your posts/design. I'll built it soon :-) and let you know.
I like trx design too!

Putting a resistor (or a pot) after C8, just before the diodes, should set the attack (charging time of C9).

For C9 have you tried the tantalum ones ?

[midwayfair]

Hi, thanks for your posts/design. I'll built it soon :-) and let you know.
I like trx design too!

Putting a resistor (or a pot) after C8, just before the diodes, should set the attack (charging time of C9).

For C9 have you tried the tantalum ones ?

I don't think the composition of c9 is important. Any internal resistance differences will be negated by the resistor that's in parallel.

And yes, you could add an attack mod, but I prefer a near- instantaneous attack here, especially since the decay in one mode is so short and the decay isn't very long. If you use a 1k or so, it'll feel more like an Orange Squeezer I think.

[Kipper4]

could i use a 2n5088 in the place of the mpsa18 Jon? please

[midwayfair]

could i use a 2n5088 in the place of the mpsa18 Jon? please

Yes, any medium-high gain transistor is fine there, and I suppose there was no real reason for me to specify MPSA18 when a 3904 would have worked with a slightly higher R3 and made the project cheaper.

2N5089 would be slightly closer in the gain range of the MPSA18, but it shouldn't make too big of a difference. You can always bump up R3 a little if you need to.

[Kipper4]

i'm certain i have a few 89's in.
Thanks for the quick response
Rich

[Kipper4]

Right now i'm confused with comparing the layout to the schematic
R4 appears as 2k2 in the BOM and schematic but as 4k7 in the layout correct?
Also i cant figure out how the top left hand corner is wired
ie junction of c2,c4 ,d2 goes to BS170 gate?
ie junction of C4 C6 C5 go to the drain

i cant for the life of me figure out how its wired underneath on the perflayout which btw i just used to build this
thanks for any help Jon
Right now its buzzing and beyond low volume.
time to get the dmm out and put it on buzzer.

[midwayfair]

Right now i'm confused with comparing the layout to the schematic
R4 appears as 2k2 in the BOM and schematic but as 4k7 in the layout correct?
Also i cant figure out how the top left hand corner is wired
ie junction of c2,c4 ,d2 goes to BS170 gate?
ie junction of C4 C6 C5 go to the drain

i cant for the life of me figure out how its wired underneath on the perflayout which btw i just used to build this
thanks for any help Jon
Right now its buzzing and beyond low volume.
time to get the dmm out and put it on buzzer.

Which BOM are you using? The BOM in Josh's document is correct, but the schematic still has the old values. (I'd update the schematic, but I don't have the eagle file it was created from.)

The connections on the schematic are all still correct, but a couple values are different. The values have all been updated on the layout. R4 = 4.7K now (it's the only 4.7K in the schematic).

Moving from left to right:
Gate connections: The 1uF input cap (C1), the 10M resistor connected to Vb (R2), the Zener's cathode (D2), and one leg of the 100pF cap (C4) are connected to the gate.

Drain connections: 100pF cap (C4), 4.7K resistor connected to Va (R3), 10uF output cap (C6), and 10nF cap that goes to the rectifier (C5).

The other connections in the top left:
10uF cap from Vb to ground.
100K from Vb to ground.

(Vb is the bias voltage created by the 100K/100K voltage divider)

Take some voltages, and post pics of both sides of the board. If you used a socket, you might want to pull Q2 and focus on getting the MOSFET booster half of it working first, since it's the really simple part of the circuit.

[Kipper4]

its almost certainly the mosfet booster thats the problem.
I dont know much about transistor voltages but i know this feels wrong

wallwart voltage when connected to circuit 8.7

Q1
D8.15
S 0
G 0

Q2
C 7.41
B 0.61
E 0.36

Q3
D 0
S 0
G 0.63

i'll take some pics in a minute

[Kipper4]

Vb is 5.46v btw

[Kipper4]

i think i have it wired right now but after about 15 seconds it loses all the volume and only works when i unplug it and reconnect it.
and it hums bad too.
i'll be back tommorow evening now its getting past my bed time.

[Kipper4]

new voltages
Q1
D 3018
G 0.92
S 3.1

Q2
C 7.74
B 0.63
E 0.39

Q3
G 0.63
S 0
D 0


WALLWART CONNECTED 9.2V
vb 5.8v

[midwayfair]

new voltages
Q1
D 3018
G 0.92
S 3.1

Source is too high. double check around there.

My internet is really spotty this afternoon, I might not be around to help. There's a bunch of general troubleshooting tips in the Madbean thread for this circuit. If I have a better connection tonight I'll come back and help more.

[Kipper4]

I neglected to mention I used 1n60 diodes too

[Kipper4]

Can I also just say don't worry tonight I should be asleep by now so ill debug it tomorrow night. I really appreciate your help john

[dan83lz]

Quick question.  I'm relatively new to circuit design and still don't have too solid a grasp of how compression circuits work, so forgive me if this sounds stupid, but have you tried replacing the switch, R9, and R10 with a 50-100k potentiometer to give more control over the decay?  Would that work or would the C9 value need to be variable as well?  I'm planning to give it a try anyway once the zener diodes I ordered get here, but was just curious of your thought on that. 

Or if my idea sounds silly, is there a better way to allow variable control of the decay?

[Kipper4]

Can I just say I fixed mine after finding a missing connection thanks to Jon for my best compressor yet.

[midwayfair]

Quick question.  I'm relatively new to circuit design and still don't have too solid a grasp of how compression circuits work, so forgive me if this sounds stupid, but have you tried replacing the switch, R9, and R10 with a 50-100k potentiometer to give more control over the decay?  Would that work or would the C9 value need to be variable as well?  I'm planning to give it a try anyway once the zener diodes I ordered get here, but was just curious of your thought on that. 

Or if my idea sounds silly, is there a better way to allow variable control of the decay?

The decay works thusly:
Your guitar signal is rectified to a negative voltage from D3 and D4. (Read up on envelopes for further understanding, especially Mark Hammer's "technology of envelope effects" on Geofex.) This is DC, so remember it has peaks and valleys.

Negative charge is stored up in C9. This is necessary because otherwise your DC-converted signal would make the compression pulsate. The cap smooths things out. The same thing happens at the DC input of a circuit ... that's what power filtering caps are for.

To make C9's behavior more predictable, we can set a time constant by putting some resistance in parallel with the capacitor to drain off a certain amount of voltage at a certain speed. The rough calculation is uF * KOhms = mS. So 4.7uF * 22K = 103mS decay. That's pretty short. 4.7uF * 122K = 573mS. That's kinda medium. For comparison, it's a tiny bit longer than an Orange Squeezer and a second and a half shorter than a Ross Compressor.

The reason I used the switch is that the maximum threshold (comp knob) setting on the long decay is already pretty much more compression than I can imagine anyone wanting. And lowering the threshold will make the decay seem a little shorter by just not pumping as much voltage into C9. This is why I described the long switch as sort of taking over from where the short setting leaves off. A lot of settings on compressors are interactive.

This isn't to discourage you from just wiring the switch up as a 100K-500K pot. But remember too that a switch is usually cheaper than a pot + knob.

Kipper: Thanks again for building it. Glad you like it.

[midwayfair]

Posting for clarity on the most recent version:

The links on my page for the schematic and layout correspond to each other. My schematic doesn't really correspond to anything else at the moment, since it was a minor change but it can be built on the PCB as follows:

http://1776effects.com/wp-content/uploads/2014/02/Bear-Hug-Fab-Updated-2-15-14.pdf

Omit R13
Jumper C10
You can jumper or leave R11, it's just a limit on the comp control. (I actually forgot to leave this in on my layout and schematic, so I'll go ahead and put them in)

Follow the bill of materials otherwise.

If Josh does another run of boards, they will correspond to those changes.

So, anyone building this on my perf layout, the perf layout is exactly how I currently build them. The changes above will work for the PCB. I don't deal with vero, but it shouldn't be too hard to backtrack the current component values on IVIark's layout, since the topology is the same except for a couple omitted parts.