The diagram at
http://focus.ti.com/lit/ds/symlink/cd4017b.pdf
shows 3 4017's, not two. It uses the first 4017
to control LEDs 1 through 9, the second to control
LEDs 10 through 17, and the third to control
LEDS 18, 19 and 20.
I can email a jpg of the schematic for that circuit.
I will "draw" an ascii schematic below that shows
the interconnections required between the 4017 ICs.
It does not show the LED wiring - that wiring will
be described. LED wiring is very close to what is
shown at the url you posted for free circuits. You
can use the 4011 clock circuit from that page.
CLOCK +-----------------------------------------------+
| | +-----------+ +-----------+ |
| | | | | | |
| ---15---- | ---15---- | ---15---- |
| | | | | | | | | |
+----14 | | +---14 | | +---14 | |
| | IC1 | | | | IC2 | | | | IC3 | |
| +-13 | | | +-13 | | | +-13 | |
| | | | | | | | | | | | | | |
| | --3--11-- | | | --3--11-- | | | ---10---- |
| | | | | | | | | | | | | |
| | +-- | ----+ | | +-- | ----+ | | +-------+
| | | | | | | | |
| +--------+ | +--------+ | +--+--[15K]----+
| | /e | /e | |
| +-[1K]-| 2N3904 +-[1K]-| Gnd [C1]
| \c Q1 Q2 \c |
| | | +9V
+--------------------+-------------------+
How it works: At power on, C1 (.01uf) and the 15K resistor send
a (+) reset pulse to IC1, which forces pin 3 on (+). That in
turn resets IC2, which in turn resets 1C3, so all three 4017's
are set to 0 upon power on. There is _no_ LED connected to
pin 3 for IC2 or IC3, but there is an LED (LED1) connected to
IC1 pin 3. Clock pulses are applied to IC1 pin 14 and to the
collectors of both transistors. These transistors (both are
2N3904's, but you can use almost any NPN) will not pass the
clock pulses through to the emitter, until the base goes positive.
With each clock pulse, IC1 advances one count, lighting
LEDs 2,3,4,5,6,7,8 and 9 in sequence. The next clock pulse
extinguishes LED9 and causes pin 11 to go + . That places
+ on the base of Q1 and on pin 13 of IC1. With pin 13 positive,
IC1 is frozen until it gets a reset pulse. With the base
of Q1 +, the same clock pulse that moved IC1 causes IC2 to
increment from 0 to 1 and light LED 10. (Since there is no LED
connected to the count 0 pin (pin 3), IC2 must increment to 1
before it can light an LED.) IC2 is wired to LEDS 10 through
17, as mentioned earlier. It operates identically to IC1.
When IC2 places + on its pin 11, it places + on the base of Q2
and freezes as IC1 did. Because the base is +, Q2 can now pass
clock pulses to IC3, and it can advance. It lights LEDs
17, 18, 19 and 20 in sequence. At the next clock pulse it places
a plus on pin 10, which sends a reset pulse to IC1, and the entire
operation starts over.
_LED wiring recap:_
All 20 of the LED cathodes connect to a resistor (R3 at the url
you posted) - make it 3.3K and use the LEDs John recommended.
LEDs 1-9 connect to IC1 as shown on the url you posted. _NO_
LED is connected to pin 11 of IC1. LEDs 10-17 connect to
IC2. _NO_ LED is connected to pin 3 or pin 11 of IC2. Otherwise,
the LED wiring is the same as that shown at t6he url. LEDs
18, 19 and 20 are connected to IC3 pins 2, 4 and 7 respectively.