This is the parts list and descriptive text for my layout(s) of the
Univibe.  I have made two versions, one with the bulb-and-photocell
setup of the "original" and one with LED's driving photocells. The
bulb-and-photocell is "vibe-a" and the LED version is "vibe-b".

As I noted in my original post, there are some changes between the
circuitry of this layout the original Univibe circuit.  I started with
the Univibe circuit from Unicord's service/owner's manual for the
Univibe.

The signal path is completely unchanged.

Let me repeat that.

The signal path is completely unchanged.

The only changes were in the photoresistors, driver, and power supply.
The orignal power supply was half-wave rectified.  This led to the
need for lots of power filter capacitance to hold hum down.  I changed
this to full wave rectified and took out a capacitor.  I thought that
three more diodes was a good trade for one fewer big capacitors.  You
can leave the diodes out and hack-wire in another capacitor if you
like.  The original used series R-C filters to make the signal path
hum-free.  I put in a three-terminal regulator to keep the signal path
power clean.  This is much cleaner power than the original.  Don't try
to tell me that additional 60-cycle hum was an important part of the
original sound.  I stand there snickering if you do.

I checked though a lot of catalogs and directories and never found a
photocell with a number corresponding to the note on the original
schematic.  An interested friend who had access to a Univibe was kind
enough to measure the resistance swing of the photoresistors at 50K to
250K.  I wound up making a version with LED's and phototransistors
first, then retrofitting to available incandescent bulbs and
photocells per the "original". The 50K-to250K range did not give me
a good swing by ear, so I increased the range to 10K to 400K in my
prototype.

Any method you use to put in a variable resistor that makes that kind
of swing will sound the same.  The electrons don't know or care
whether they're going through cadmium-sulphide, carbon, silicon or
whatever.  All they know is resistance, capacitance and inductance.
If you can't or won't believe that, you're running on superstition.

I laid out two different version of the circuit to give you several
choices on photoresistors.  The vibe-a circuit layout accepts a
12V/40ma grain of wheat bulb and four .250 inch photocells in a layout
which fits a cut-down 35mm film canister as a light-tight shield.  The
vibe-b circuit layout will accept the CLM6000, the Vactec V5C2 or
V5C4, and the H11F1 LED to photo-FET optoisolator.  You should be able
to make any one of these work.  The vibe-b will also accept discrete
photocells and separate LED's.

If you use the vibe-b layout for one of the LED driven methods, you
will need to experiment to find out how much LED current gives you how
much change in the photocell resistance, then tinker with the driver
circuit to get the resistance swing you need.  I think that a
five-to-one resistance swing is OK.  Wider will make the effect more
extreme, narrower will be less extreme.  The actual resistance that is
in the middle of the swing determines the bass-to-treble center
position of the effect.  The two resistors specified as "39K" in the
collector circuit of the driver help determine the static current of
the driver transistor and hence the "rest" position of the sweep..
Change these two resistors and trim with the trimpot in the emitter of
the driver until you get the resistance center and swing you like.

If you use the vibe-a layout, you'll still need to adjust the emitter
resistor pot on the driver transistor to get a good light-to-dark
swing on the light bulb.

From an engineer's point of view, there are some things that look
like poor practice in the 'vibe. The half wave rectification is poor
economy. Maybe caps were cheaper than diodes back when it was designed.
The bulb driver is a poorly stabilized circuit. This thing probably drifts
and has wide variations in the originals. On the other hand, the LFO is
a neat design, elegant and practical.

On the LFO: In the original circuit, the LFO produces a rising amount
of signal as the speed goes up. I think this helps to compensate for
the decreasing sensitivity of the bulb to changing current as the
speed goes up. The back-to-back diodes in the LFO phase shift
oscillator circuit seem to be the biggest contributors to this. If you
pull these out, the LFO produces a much larger signal and it remains
(relatively) constant with speed. There is a 47K resistor specified in
series with the base of the driver transistor. If you don't get enough
swing on the driver, reduce this to 4.7K, or even to 0.

In the vibe-b LED/photocell circuit, pull the diodes out. and increase
the 100K resistor from the collector of the driver transistor to its
base to between 470K and 750K, again to get the necessary current
swing. I provided a space for you to use two transistors as a
darlington in the place of the original single transistor if you need
more current gain in the driver; there is an intentional short between
the base and emitter of this transistor marked with an arrow on the
copper pattern. If you use one transistor, leave the short in. If you
use two in darlington, cut the short at the arrow.

The 250K dual pot should be reverse-log taper to give linear apparent
control of speed. 250K dual reverse lot pots are at least as scarce as
original "Univibe"'s. I did some math, and put in two 330K resistors
that connect from the ends to the wipers of the dual pot. The math
says that this will give a good approximation of the reverse log taper
if you use a 1M dual linear taper pot, which ought to be a whole lot
easier to find.

There was an ongoing battle in the audio-engineer world back when phase
shifters (for that is what the 'vibe is) were new. The phase shifting
capacitors in the shifting stages are staggered in value in the Univibe.
An opposing opinion was that shifters with all-the-same cap values would
make a more effective phasing sound, and that is the opinion that was
finally accepted in the pro-audio world. You can make your own choice by
tinkering with the values of the four film phase shifting caps, if you
like. Oooops.... but that wouldn't be original, would it????

Resistors
All 1/4W unless otherwise noted
150     1
1K 1/2W 1
1K2     1
3K3     2
4K7     17
6K8     1
22K     2
39K     2 (vibe-b only)
47K     8
68K     1
100K    11
220K    2
330K    2 (use with the dual 1M linear taper speed pot)
470k-750k 1 (vibe-b only)
1M2     2
2M2     1
1K pcb mount trimpot (bias adjust for the phasing opto's) - 1
250K dual panel mount pot (LFO speed) -1
(note: you can use a dual 1M linear taper if you use the 330K's)
50K panel mount pot (LFO drive level) -1
100K panel mount pot (output level) - 1

Capacitors
1uF/25V Alum elec. radial pcb mount -16
100uF/25V Alum elec. radial pcb mount -2
220uF/25V Alum elec. radial pcb mount -1
1000uF/25V Alum elec. radial pcb mount -1
330pF film - 1
470pF film - 1
4nF film   - 1
15nF film  - 1
220nF film - 1

Semiconductors (in order of preference; watch pinout for your choices)
input transistor (1)  - 2SC539(orignial), 2N5089/2N5210
remainder of signal path (9) - 2SC838c (original) or 2N5210/2N4124/2N3904
 /2N2222A/2N4401
I have built one with all 2N3904's except the input transistor, which
was a 2N5089. For each darlington pair you can use an integrated
darlington pair like the MPS-A13.

LFO transistors (2) - 2SC838c (original) or 2N3904/2N4124/2N2222A/2N4401
photocell lamp driver (1) - 2N3904
LFO diodes (2) - 1N914/any silicon diode
rectifiers (4) - 1N4001 through 1N4007
7815 three terminal regulator

Photoresistors
For vibe-b, four of:
- CLM6000 LED/photoresistor ($3.50 each, Hosfelt Electronics, see Effects
  FAQ)
- Vactec 5C2 or 5C4
- H11F1/H11F2/H11F3 LED/photo-fet analog isolator (Digi-Key, $1.95 each)
- Photocell with 10K to 250K light to dark range (Mouser Electronics,
  $1.50 each, possibly Radio Shack) and four high brightness green
  LED's

For vibe-a
- Four photocells and a 12V/0.04A incandescent bulb
  (a Radio Shack 8V grain-of-wheat bulb works well)


Misc other
SPST stomp switch (original) or DPDT stomp switch (true bypass)
2- 1/4" phone jacks for input/output
knobs for amplitude and speed
jack to match wall wart transformer
wall wart transformer, 14-18VAC or 20-25VDC, 50ma or more
Light shield if you make vibe-a
Box to put it all in


To make the light shield for the bulb version:
Get a Kodak plastic film canister, black with a grey snap on cap.
Remove the cap.  On a flat surface place a 1/2" (12mm) spacer and hold
the point of a thin, sharp knife blade flat on the spacer so the point
sticks a little off the edge of the spacer.  Place the closed bottom
end of the canister on the flat surface, and slide it over so it
touches the point of the blade.  Press the canister side into the
blade and rotate the canister so the blade cuts into the canister.
Do not hold the canister so if you slip or if the canister cuts
through suddenly your finger or other personal part will contact the
blade.  Use light pressure and make several rotations until the
canister cuts through.  Clean up the edge if needed.

Take the now-shortened canister and press the edge of the open end
over a relatively sharp corner, as of a table or cabinet, and again
rotate the canister, pressing the edge into the edge of the corner.
Rotate and press evenly so the corner edge forms the open edge of the
canister into a small lip turned outwards.  Try to work evenly so the
plastic forms into an even lip.  The lip is needed so the cap will
stay attached; you cut the original plastic lip away when you
shortened the canister.  Try the cap, and then keep forming the
plastic until the cap stays snapped in place.

Once the canister is formed, remove the cap. Cut a very small hole
exactly in the center of the bottom, smaller than the space between
the two holes on the circuit board for the libht bulb. This will be
used to align the shield.

Take your circuit board and mark a spot exactly in the center of the
space between the two holes for the bulb on the component (not copper)
side of the board.

Cover the bottom of the canister with double sided foam tape, leaving
the covering paper on the outside.  Cut through the foam tape so you
can see through the alignment hole you made in it earlier.  Place the
canister bottom down on the with the center alignment hole exactly
over the alignment mark on the curcuit board between the holes in the
circuit board for the bulb. Practice this a couple of times before you
remove the covering paper to stick down the canister. When you can do
it every time, remove the paper covering the sticky surface of the
foam tape on the bottom of the canister and stick it down exactly
centered.

Turn the board over and use the holes in the circuit board as a guide
to drill through the canister bottom so the leads for the bulb and
four photocells can pass through the bottom.

Now, form the photocell leads so they come about 0.1" out from the
back of the photocell and bend at at right angle to be parallel to the
body. Stick the leads through from the inside of the canister through
to the copper side of the circuit board and solder them in place. When
you are done, there will be a little square formed inside the canister
by the photocells, with the holes for the bulb in the center. You may
have to slightly bend or resolder the leads the get them all nice and
neat and the same height, etc. Solder in the bulb in the center of the
photocells. When you snap the cap on, the bulb and photocells are held
inside the light shield.

I recommend doing all this is two phases: Make the shortened canister
and place it on the board and drill the holes before you place any
other components on the board. Solder the photocells and bulb in after
all other components are on the board.

