I would like some assistance on building this circuit before I purchase the components from Mouser Electronics.
Just wanna make sure it works, according to the circuit I found online.

I understand how to hook it up.
I don't understand why I need the resistors and capacitors, or why those specific values are used on the IC with the crystal.

The main question I need answered ASAP is... Will it work with the components?

Some references and reading material listed below.
1. https://assets.nexperia.com/documents/data-sheet/HEF4521B.pdf
2. https://www.mouser.com/ProductDetail/Texas-Instruments/CD4521BE?qs=sGAEpiMZZMu03P%2bZhyvnn3PvsqNpY3r3bvroM1i%2bTU4=
3. http://www.ti.com/lit/ds/symlink/cd4521b.pdf
4. JK Flip Flop and T Flip-Flop (Video)
5. The Crystal I've chosen to use.
5. My Cart on Mouser, so far

I don't see many other people that wanted to do this on All About Circuits, but I did see this post Lazy 4521 Chip - Odd Problem

The data sheet is so old it seems, it's just a lil difficult looking at it. I'll try to give it another read.



Base : 2 and Exponent :18, 19, 20, 21, 22, and 23.
Q18 output is.. 2^18 = 262,144 Hz
Q19 output is.. 2^19 = 524288
Q20 output is.. 2^20 = 1048576
Q21 output is.. 2^21 = 2097152
Q22 output is.. 2^22 = (4,194,304 Hz / 4,194,304 Hz = 1 Hz, H:500ms, L:500ms
Q23 output is.. 2^23 = 8388608 / 4,194,304 = 2 Hz (0.5Hz) H:250ms, L:250ms
Q24 output is.. 2^24 = 16777216 / 4,194,304 Hz = 4Hz (0.25 Hz) H:125ms, L:125ms

So I'd really need the output from Q23 instead of Q22 then.. it seems.

 

R1 "linearizes" the internal circuit of the oscillator, in effect biasing it so that what is normally a digital on-off circuit behaves more like a high-gain amplifier. Without the resistor, the input side of the internal circuit can be used as a normal digital clock input.

R2 limits the drive power to the crystal. Some crystals require care not to apply excessive drive. At 5 volts the resistor may be unnecessary, but it is necessary to check the datasheet for the crystal to be certain. Low frequency crystals (kHz range) are usually more fussy about overdrive than high frequency crystals. All crystal oscillators take some time to "start up" - produce stable full-amplitude output. Typically the lower the drive to the crystal, the longer it takes to start. Very low power circuits can take tens to hundreds of milliseconds to start.

The capacitors are for "loading" the crystal. Crystals are designed to operate with specific load capacitance to operate at their nominal resonant frequency. This circuit operates the crystal in "parallel resonant" mode. Typically the load capacitance required is considerably higher than the capacitance associated with the "pins" of an IC oscillator circuit, so that discrete capacitors can be used to set the capacitance, allowing greater application flexibility. The load capacitance that the crystal sees is (pin capacitance in parallel with discrete capacitance) in series with (same thing, other pin). Often one of the capacitors is a variable type so that the frequency can be trimmed ("pulled") by a few parts per million. Having one 82 pf and one 22 pF cap is very odd - for a crystal in the low megahertz range both caps would typically be in the 20-30 pF range. You can find information explaining all of this in detail on the web. Most crystal manufacturers publish applications information. I also recall an old ap note on oscillators of assorted types with 4000 series CMOS gates.

EDIT: Note that the old CD4521 is much much slower than the current HEF4521 and will almost certainly not work with the intended crystal. The HEF part should be fine.

 

Crystals need loading capacitors from each pin to ground as shown in the circuit. I have usually seen them as equal values, not sure why they are different ones here. The total capacitance (which is the two in series) needs to be a specific value based on the actual crystal used. Normally, I would use two equal capacitor as twice the specified capacitance for the crystal, so, in series they equal the required capacitance.

The 1M resistor is biasing for the gate that is used in the oscillator.

The 3.9K in series is to limit the drive current to the crystal.

I cannot tell you if these specific values would work. They may need to be adjusted for the crystal used.

Where did you get the table showing 1Hz was 1000ms on and 1000ms off? That is incorrect, it is 500ms on and 500ms off.

Bob

 

The Rs and Cs plus crystal are for creating an accurate
timing clock signal. Its an oscillator circuit to feed a clock
to the divide chain in the part.

Regards, Dana.

 

Crystals need loading capacitors from each pin to ground as shown in the circuit. I have usually seen them as equal values, not sure why they are different ones here. The total capacitance (which is the two in series) needs to be a specific value based on the actual crystal used. Normally, I would use two equal capacitor as twice the specified capacitance for the crystal, so, in series they equal the required capacitance.

The 1M resistor is biasing for the gate that is used in the oscillator.

The 3.9K in series is to limit the drive current to the crystal.

I cannot tell you if these specific values would work. They may need to be adjusted for the crystal used.

Where did you get the table showing 1Hz was 1000ms on and 1000ms off? That is incorrect, it is 500ms on and 500ms off.

Bob

I've chosen this crystal ( 449-LFXTAL014640BULK ).

Load Capacitance : 12 pF

Normally, I would use two equal capacitor as twice the specified capacitance for the crystal, so, in series they equal the required capacitance.

Okay, so capacitors in series...
Let's see what two 24 pF capacitors in series are.

It's two 24pF capacitors, like you said.
The equation is.. 1/Ceq = ((1/C1) + (1/C2))
or, it's actually.. 2 x 12pF = (1/Ceq = ((1/0.000000000024F) + (1/0.000000000024F))

C1: 1/0.000000000024F = 41666666666.666666666666666666667
C2: 1/0.000000000024F = 41666666666.666666666666666666667
C1: 41666666666.666666666666666666667 + C2: 41666666666.666666666666666666667 =
83333333333.333333333333333333333
Then 1/83333333333.333333333333333333333 = 0.000000000012 Farads.

So.. Like you said, two equal capacitor as twice the specified capacitance for the crystal, so, in series they equal the required capacitance.

K.. so this/these should do the trick then?

 

Crystals need loading capacitors from each pin to ground as shown in the circuit. I have usually seen them as equal values, not sure why they are different ones here. The total capacitance (which is the two in series) needs to be a specific value based on the actual crystal used. Normally, I would use two equal capacitor as twice the specified capacitance for the crystal, so, in series they equal the required capacitance.

The 1M resistor is biasing for the gate that is used in the oscillator.

The 3.9K in series is to limit the drive current to the crystal.

I cannot tell you if these specific values would work. They may need to be adjusted for the crystal used.

Where did you get the table showing 1Hz was 1000ms on and 1000ms off? That is incorrect, it is 500ms on and 500ms off.

Bob

I see the drive level is (1mW MAX) Which is 0.001W

So doing some Ohm's Law math.. R = V^2 / P
R = (5vdc x 5vdc) = 25 / 0.001W = 25,000 Ohm's
or.. 25,000KΩ = 5Vdc^2 / 0.001W

The 3.9K in series is to limit the drive current to the crystal.

So the 3.9K Resistor isn't going to cut it.

Needs to be higher than 25KΩ

 

R1 "linearizes" the internal circuit of the oscillator, in effect biasing it so that what is normally a digital on-off circuit behaves more like a high-gain amplifier. Without the resistor, the input side of the internal circuit can be used as a normal digital clock input.

R2 limits the drive power to the crystal. Some crystals require care not to apply excessive drive. At 5 volts the resistor may be unnecessary, but it is necessary to check the datasheet for the crystal to be certain. Low frequency crystals (kHz range) are usually more fussy about overdrive than high frequency crystals. All crystal oscillators take some time to "start up" - produce stable full-amplitude output. Typically the lower the drive to the crystal, the longer it takes to start. Very low power circuits can take tens to hundreds of milliseconds to start.

The capacitors are for "loading" the crystal. Crystals are designed to operate with specific load capacitance to operate at their nominal resonant frequency. This circuit operates the crystal in "parallel resonant" mode. Typically the load capacitance required is considerably higher than the capacitance associated with the "pins" of an IC oscillator circuit, so that discrete capacitors can be used to set the capacitance, allowing greater application flexibility. The load capacitance that the crystal sees is (pin capacitance in parallel with discrete capacitance) in series with (same thing, other pin). Often one of the capacitors is a variable type so that the frequency can be trimmed ("pulled") by a few parts per million. Having one 82 pf and one 22 pF cap is very odd - for a crystal in the low megahertz range both caps would typically be in the 20-30 pF range. You can find information explaining all of this in detail on the web. Most crystal manufacturers publish applications information. I also recall an old ap note on oscillators of assorted types with 4000 series CMOS gates.

EDIT: Note that the old CD4521 is much much slower than the current HEF4521 and will almost certainly not work with the intended crystal. The HEF part should be fine.

Yeah..? Where is the HEF4521 part ? I can't find it anywhere online. All I can find is the HEF4521 SMT/SMD IC.

 

Note that the old CD4521 is much much slower than the current HEF4521 and will almost certainly not work with the intended crystal. The HEF part should be fine.

Okay, so at 5vdc, the typical input frequency is 4Mhz, not 4.194304Mhz. But at 10vdc, the typical input frequency is 10Mhz. So I'd need to up my voltage from 5 to 10vdc. Right? That should work. I can't and do not want to use the SO-16 package HEF4521 IC.

 

I see the drive level is (1mW MAX) Which is 0.001W

So doing some Ohm's Law math.. R = V^2 / P
R = (5vdc x 5vdc) = 25 / 0.001W = 25,000 Ohm's
or.. 25,000KΩ = 5Vdc^2 / 0.001W



So the 3.9K Resistor isn't going to cut it.

Needs to be higher than 25KΩ

No, you are calculating the power in the resistor, not the crystal. Unless the datasheet recommends a series resistor, I would have no idea how to calculate it. Normally, no series resistor is neede for crystals in the MHz region.

24 pf is correct for two equal capacitors. Use 22 since it is a standard value and there is probably close to 2pf of stray capacitance.

Bob

 

If you just need an accurate 1Hz clock signal, I'll throw this in. Using a scrapped out PCB from a 1.5V battery desk or wall clock. Some have a pullup on the coils and others have a pulldown output.

Ken




Added note: In the upper circuit, I think I had added a 100kΩ between common and the junction of D4/D5/R2. This is a pull down for the base of Q1.