Last update 10/8/07

2005 TAPE OPTO COMPRESSOR TECH SUPPORT

There will not be any more boards (sorry), these are not available now.

Below is the "Tape Opto" compressor designed for the 2005 conference, these are not being offered outside of the conference.   This page has been updated with all the information you should need to finish it.  Please read this thoroughly, and view all the pictures before emailing.  There were also almost 180 people that took the class, so if everyone starts emailing, it will be a bit of an overload.   I tried to include ever question I have been asked over the last couple weeks, so please wade through this before emailing., but if you are stumped, then I am here to help.  SH

GENERAL DISCLAIMER

When given the challenge of designing a compressor for $20, there were a lot of hard choices to make.  Craig Schumacher and I talked about a number of important things, and we agreed that the proper power supply and power entry jack were important to avoid anyone blowing it up with the wrong power/polarity.  We also felt that things like boxes, knobs, jacks... were easy things to get, and added to the unique outcome of each kit (large oven knobs are my favorite, set to broil).  This left about $14 for the printed circuit board and parts.

My biggest goal was to make building it not only fun, but worth the parts you paid for.  To me it wasn't $20, it was $20 x 180 people!  So the audio path was made discrete class A, and even with inexpensive parts, it has excellent bandwidth and phase response.  The control circuit that activates the opto pad, was designed for +4dBm signals, and works adequately for as primitive as it is.

IN DEPTH GARBLE FOR SUPER GEEKS

The control circuit could have been better, given infinite time and two more dollars.  I had a working version that I thought sounded good so I took it to Larry Crane's studio to test it out.  He thought it sounded good too, and dialed in the settings for attack and release.  At that point, I was afraid to start messing with it, so the design stayed. 

The control circuit needs more gain for -10dBm signals.  The compression circuit was designed to have enough gain so that a 0VU (+4dBm) signal would send it into full compression at 3 o clock (a little gain to spare).  With -10dBm signals, the diode in the detector is not forward biased, so little compression occurs.  However I did use it with a cassette 4 track that had -10dBm level, and I did get full compression, but the tracks were pretty hot, and I used the line outs on the back.  I played with the non linearities of the vactrol, log amp, detector, and attack release circuits to try to make it behave as linear as possible for +4dBm signals.

CIRCUIT DESCRIPTION: schematic and parts list.

BOARD LAYOUT / PCB ART

Below is the layout for the TapeOpto compressor pcb.  The overall dimensions should be 1.82" x 2.7".  This is the only format I can export in with the program I use for PCB layout.  Please do not ask for the board file, or another version of it.  This is a component view, meaning you are looking at the top of the board.  To make a pcb you will have to mirror this image.  I authorize the use of this for PERSONAL USE ONLY.  Any use for deliverable products sold for profit will be in violation of the copyright.

The circuit is broken up into two parts, the audio path drawn on the top, and the control circuit on the bottom.  R1 and U1 (the vactrol) form a voltage divider which is what does the actual compression.  The circuit drawn between R2 and R10 is a make up gain stage with approximately 34dB of gain.  To increase the overall gain, increase R6, to decrease it, make R6 smaller.

The signal goes through the opto pad (R1,U1), and is fed into the input of the volume pot R2.  This feeds the first gain stage (R3,C2,U2,R4,C1) which amplifies it by a factor of 17 (24.6dB).  Because this is a discrete class A single gain stage, it is inverting, meaning the signal coming out of C2 is 17 times larger than it is going into the gain stage, and the phase is reversed.  The signal then goes into R5/R6, a -15dB pad, where it is fed into the second gain stage (R7,U3,C3,R8) which amplifies the signal again by a factor of 17, and flips the phase again, making it in phase with the input.  U4, C4, R9, and R10 are an Emitter follower, which buffers the signal to the outside world.   The voltage at the base of U4 and the voltage coming out the emitter  are almost the same amplitude, but the output has greater current drive (like a mini power amp).  Because all the gain is after the pad, the design has essentially unlimited headroom on the input, the output is limited to about +20dBm, or +16vu, which is actually quite good considering it runs on a 20vdc rail.

The control circuit uses another JFET gain stage at the input, and then D1/C5 to convert the AC audio signal into a 1/2 wave rectified DC voltage.  This then passes through R16,R17,C6 where the dynamics of it are altered giving the attack/release characteristics of the compressor.  The last part of the circuit (U6,R18,R19,R20,U7) is a voltage to current converter with gain.  This takes the signal from the attack release circuit, amplifies it, and converts it to current to drive the vactrol.  R21 sets the idel current of the vactrol, so its internal diode is turned, this gets the photo resistor into a more linear operating regoion.  U8,R11,D3 is the display circuit which drives the LED.  If you want more compression at lower input levels, try decreasing R19, this will increase the voltage to current converter gain.

HOW IT SHOULD WORK.

A +4dBm (0vu on a professional machine) should be able to drive the unit into full compression with the compression control (R12) at about 3 o'clock (the LED will light up with max intensity).  Driving the circuit with -10dBm signals will not drive the control circuit hard enough, therefore it is recommended you use this with +4dBm signals.  The intensity of the LED is the indicator of how hard you are compressing.  If you have no test equipment, the best way to test it is to run audio through it.  The LED will light up briefly at power up, and then only come on as you begin to compress.  If your LED is stuck on, you have a problem in the control circuit.  With the compression knob down (off), the circuit should have considerable gain which is a function of R2 (volume control), as the compression is turned up, the level will drop as the opto pad begins to compress.  Turn up the volume to make up for the gain loss.  The ratio at full compression is 3.5:1.

TESTING

Check the following DC voltage, if yours are off, then look for problems in that part of the circuit.  These are a quick check to make sure the power supply and gain stages are working right, which covers much of the circuit.  All measurements are relative to ground.

V+, output of voltage regulator.  Measure at the positive end (no band) of D2, should be 20vdc, +-0.5 volts, and very little AC.

Measure DC volts at the Drain all all three 2n5457s, they should all be around 10vdc +-1.5vdc.

MINE DOESN'T WORK, WHAT SHOULD I DO

9 times out of ten the problem can be seen.  First check all the parts (except resistors and the two green square caps) for polarity.  The pictures below should help determine the orientation of the parts.  My boards have the point-point wires for power  inputs, outputs, and attack/release.  The LED has a flat side around the base, this indicates ground and should face the outside of the board.  I would also mount the LED up high for test, so you can remove it and mount it in the front panel when you box it up.

Check for solder bridges between pads that should not be connected. 

Check all solder connections, are they all soldered?

Clean the board with alcohol, don't get it in the pots R2 and R12, scrub the solder side clean with Q-tips.

Check the polarity of all transistors and the 8 pin chip, they should match the drawing.

Check diode polarity, the silver band on the two 1n4007 diodes should match with the band on the layout.

INTEGRATING INTO A BOX

Below are pictures of the pair I built up.  Start (if you want) by drawing it to scale on the computer and printing out templates of the mechanical layout.  Glue stick them to the front panel and drill out to the appropriate size.  Its best to step up slowing in bit sizes for cleaner holes.  If you are drilling steel, use low RPM and lubricant or you will dull bits very quickly.

  R2 and R12 have a small tab that needs to be removed in order for the control to mount flush to a panel.  The two controls are 1" spaced. 

Here's the template I used for my pair I brought to the conference.  The dimensions at the top are of the outer gray box.  The two brown circles are R2 (top) and R12 (below R2).  The green rectangle is the pcb viewed from the edge.  The lower two controls are the 2M pots (for adding adjustable attack and release) listed on your parts list draw to size.

ADDING ATTACK AND RELEASE

Follow the notes on the schematic for details, the pictures below show the trace you need to cut for the release.  There is no trace that needs to be cut for the attack.

This has nothing to do with the compressor, my wife just put these cool red knobs on our tube mixer and I thought it looked rad!

WIRE ROUTING AND GROUNDING

The picture below shows one of the units I built opened up.  The power comes in via the purple and black wire (bottom right).  Notice this only goes to the board to the pads provided, and runs at 90 degree angles to the audio (input top, output on bottom).  I used TRS jacks and grounded the ring so I could use these with transformer balanced gear.  There is green felt under the board to prevent it from touching the metal panel behind it.  Never allow the back side of the circuit board to touch metal!

POINT-POINT

Audio input (pointer).  There are two ground pads below it, one for audio, one spare.

Audio Output (pointer), ground pad on right.

Jumper "FB" to pad right above it.  "FF" pin was used for test only.

R22, as indicated on the schematic, is a jumper, it must be in place for the circuit to work.  Also used as a ground test point.

POWER FROM WALLWART

The pointer shows the positive input to the board labeled +24vdc (ground pad is on the left), the middle picture shows the positive terminal of the power entry jack (red) and ground (black), the last picture shows the no load voltage of the wallwart.  If you mix this up and power the circuit with the polarity reversed, irreversible damage will occur.

Great to see you all in NOLA!  If you get this far and still have questions, email me at Scott@hamptone.com.