THE ENGINEER'S THUMB... At last, a better compressor!

[Cliff Schecht]

> why he chose the exponential current source over the simpler linear

Thre's no exponential current source here.

Ah yes you are correct, the current source itself just produces an output current based on the input voltage. I guess the overall circuit could be considered exponential just in the sense that the charging and discharging of the caps happens exponentially, but the current source itself is linear. If his circuit acted in a truly linear nature the compressor might sound weird and probably unnatural to our ears. Or maybe it would sound cool? You could probably figure out a way to finagle one of the floating darlington pairs into a current source that forces the caps to charge linearly.

[Jdansti]

Has anyone built the vero version yet?  I'm having a little trouble. I built it according to the vero on page 1 with the exception of a jumper to move the input (row 2) down to an unused row at the bottom. This places all of the leads going to the switch close together.

The signal switches between the board and bypass when I activate the DPDT switch, but the indicator LED stayed on in both conditions. While the switch was in bypass, I happened to turn the volume pot and the indicator dimmed of and on as I rotated the pot. While I was doing this, the volume pot started smoking.  I disconnected power, the switch and pots.

I reconnected just the power and the LED came on. I connected the "FS" and "Out" leads to simulate the switch in bypass and the LED stays lit. The other thing I noticed was there was 9v on the lead that goes to the center lug of the volume pot.

Q1: Is there supposed to be 9v on the center lug of the volume pot in bypass?  
Q2: Possibly a bad transistor?

Edit:  I'm using a tiny indicator LED (see photo below) instead of a standard 3mm or 5mm LED.
Q3: Could this LED's leakage be out of spec resulting in  the bypass circuit not operating properly?


Edit 2:  As my cousin Rosanne Rosanna Danna used to say, "Never mind...".    FS was soldered to the wrong track. I corrected it and now it turns off and on. Everyone else is in bed at this hour, so I'll test drive it tomorrow.


Here are some pics:

Vero layout:



My board:

[Jdansti]

Good news and bad news.

Good: Millenium bypass works fine. I get signal with effect on and in bypass.

Bad:  When the volume pot is at max and the ratio pot is at min (I assume 100% dry), the output volume is about 1/2 of the unbypassed volume. When the ratio pot is 100% wet, the output volume equals the bypassed volume, but there is no compression (actually, no compression on any settings).

Any ideas?

[rousejeremy]

Your Q1 looks backwards

[Jdansti]

Your Q1 looks backwards

Thanks for taking a look. I think the data sheet for the BC327 shows that Q1 is oriented the same as the vero layout.  Could you please look at the configuration below and let me know if you agree?

[merlinb]

I can't see any glaring errors, so I guess it's time to post your voltages. Also check the orientation of your diodes, since we can't see that in the photo.

[rousejeremy]

My mistake. I must have been looking at the pinout for a 2N pnp

[Jdansti]

Here are the voltages. I noticed that out of pins 1-8 on IC2, only pin 6 is connected to anything else.

PS 9.5

Q1
e 4.7
b 4.2
c 1.1

Q2
e 7.47
b 0 (LED turns off when I touch b with probe.)
c 9.5

IC1 
1.   4.7
2.   4.7
3.   3.2
4.   9.5
5.   2.4
6.   4.7
7.   4.7
8.   4.7
9.   4.7
10. 2.4
11.   0
12.  2.4
13.  4.7
14.  4.3

IC2
1.   0
2.   0
3.   0
4.   0
5.   0
6.   0
7.   0
8.   0
9.   0.01
10. 0
11.  9.5
12.  4.7
13.  4.7
14.  4.7
15.   0
16.  1.1

Here are the diodes:

[merlinb]

Hmm, voltages check out.

Next step, check the wiring between the board and the off-board stuff (jacks / pots etc.)

EDIT: Aha! Double check R7- should be 10k but looks like 100R in your photo?

[Jdansti]

Hmm, voltages check out.

Next step, check the wiring between the board and the off-board stuff (jacks / pots etc.)

EDIT: Aha! Double check R7- should be 10k but looks like 100R in your photo?

It sure looks like 100R!   I'll take a look at the board as soon as I can get to it.  Thanks!

[Jdansti]

Thanks for the eagle eye, Merlin!  I checked all of the resistors and found another wrong value. It works great now that I have the correct resistors in place!   

[The_Armadillo]

So i build this little guy up, per the PDF with the PnP in it. I'm getting this nasty, gated distortion with the ratio anywhere above 9:00. Below there it cleans up significantly. Possibly a bad OTA? I got it on the cheap from eBay, so possibly a fake?
Only part subs were that I haven't put in the power filter cap, the output electro for a 1u poly.
I'll post voltages if the problem doesn't stick out.

[midwayfair]

Hey, folks, I posted a video demo of this pedal on my YouTube. Hopefully it will convince some of the fence sitters to build this excellent pedal!

http://www.youtube.com/watch?v=n7VqdIlHkI0

[mth5044]

Has the VU'd, expanded edition come to fruition?

[StephenGiles]

Hey, folks, I posted a video demo of this pedal on my YouTube. Hopefully it will convince some of the fence sitters to build this excellent pedal!

http://www.youtube.com/watch?v=n7VqdIlHkI0

Excellent demo, many thanks.

[Cliff Schecht]

To be honest I have my doubts that any compressor will ever replace my Orange Squeezer. It's been one of my mainstays for so many years and is on 90% of the time during gigs. For what I want a comp to do, it's perfect.

I still need to build Merlin's circuit correctly. I've screwed it up twice now (bad perf build and etched a board backwards, neither worked ) and really want to play with it still. I usually don't screw circuits up this badly when building but I've been pressed for building time between finishing up grad school and playing gigs.


Also a comment on the LM13700: IIRC it doesn't have fantastic noise performance. You have to cut the signal down to the mV level at the inputs and then amplify the crap out of it again which in itself tends towards being noisy. Typically when I use a 13700 I try to keep the signal swings as large as possible before clipping to maximize noise performance. Merlins clever trick was to put the 13700 in the op amp feedback loop where the noise performance of the 13700 has a less significant contribution to the overall system noise (compared to if you were relying on the 13700 as a signal-path device).

[midwayfair]

Also a comment on the LM13700: IIRC it doesn't have fantastic noise performance. You have to cut the signal down to the mV level at the inputs and then amplify the crap out of it again which in itself tends towards being noisy. Typically when I use a 13700 I try to keep the signal swings as large as possible before clipping to maximize noise performance. Merlins clever trick was to put the 13700 in the op amp feedback loop where the noise performance of the 13700 has a less significant contribution to the overall system noise (compared to if you were relying on the 13700 as a signal-path device).

Yes but compared to the CA3080 the noise performance is spectacular. I agree that it's not quiet like a TL074 or anything like that.

p.s. Re: Orange Squeezers, I'll see your 90% of the time and raise you 10%. Always on during shows. I can't live without it.

[PRR]

> put the 13700 in the op amp feedback loop where the noise performance of the 13700 has a less significant contribution

Right: Dang near NO noise contribution when not limiting.

This can't be compared to "normal" use of OTA; in fact it is mostly the reverse of OTA flaws. The hiss only happens in heavy limiting, and at that point who cares? As the note decays and gain rises, hiss gets _less_. I'm not going to work-out the idle noise because numbers don't mean much, and it is (should be) quick/easy to just try it. (No, I'm not giggling at your expense-- I do little PCB these days and you should have seen my 2-day screw-up of a simple 2-hour screen door last week.)

[merlinb]

Hey, folks, I posted a video demo of this pedal on my YouTube. Hopefully it will convince some of the fence sitters to build this excellent pedal!                            http://www.youtube.com/watch?v=n7VqdIlHkI0

Great demo- you've done me proud! Also what's kinda spooky is that you bear more than a passing resemblance to me... (or perhaps I should say that I look rather like you )

[midwayfair]

I recently did a "fill this box" thread on Madbean's forum, offering a free PCB for anyone who told me what to fill a black box drilled with two pots and a toggle with. One poster suggested the Engineer's Thumb with treble boost not knowing I had already built one & how much I liked it ...

The trouble was, my previous layout was huge and long and skinny, and it wouldn't fit in the box I had. So that meant I had to come up with a new layout. I thought I'd share it with community for the perf lovers out there (or just guys who don't etch).



VERIFIED 8/22/12 (I built it last night and tested it tonight.)

There are several jumpers -- it's just unavoidable with this circuit. So it's not the absolute cleanest circuit layout ever, but it's very compact for this build. There are no fully standing resistors or diodes, though a couple 1M and the 220R are squeezed onto 3 holes instead of 4. The board will fit with a battery in the case ... or even maybe even a charge pump daughter board for 18V if this circuit will allow it! I believe there are enough pads to accommodate adding the attack and release controls, though you may need series resistors to avoid weirdness.

PDF for those who want it: http://jonpattonmusic.files.wordpress.com/2012/08/engineers-thumb.pdf