DCO/VCO Fine clock output?

[[email protected]]

Hey there,
I'm trying to figure out how to output a clock frequency depending on
the position of a potentiometer slider.

I need the clock frequency to range from 39KHz to 49KHz and the
mid-point to be 44.1KHz. I've been informed a DCO would be the best
option but I'm not sure if their output range can be that finely tuned?
All the DCO's I have found go up to around 20MHz and do not seem suited
to a lower frequency task like this.

Does anyone have any advice as to how to finely tune an output
frequency of one of these devices?

My original thoughts were to use an ADC to sample in the position of
the slider and then use a DCO. Are VCO's not stable enough? I need the
clock to be accurate and precise...

Any help greatly appreciated! Many thanks, Alex.

 

[Joerg]

Hello Alex,

Look at the National LM331.

Regards, Joerg

 

[Jon]

Use an op-amp. Feedback resistor R1 from output to Inv Input
Shunt resistor R2 from Inv input to gnd.
Positive feedback R3 from output to Non Inv Input.
Capacitor from Non Inv Input to ground Inv input to ground.
Let Vsat = op amp saturated output voltage.
Let Vref = voltage at Inv input with op amp saturated.
Then Vref = VsatR2/(R1+R2)
Let Vs = Vsat + Vref.
Let f = frequency of oscillation.
f = 1/{2 Tau ln[(Vs-V)/Vs)]}
Wwher ln(X) means "natural logarithm of X".

 

[Jon]

Ignore my last post. I accidentally clicked "Send" before I was
through editing.
~
Use an op-amp. With +/- supplies:
~
Feedback resistor R1 from output to Inv Input.
Shunt resistor R2 from Inv input to gnd.
Positive feedback R3 from output to Non Inv Input.
Capacitor from Non Inv Input to ground.
Let Vsat = op amp saturated output voltage.
Let Vref = voltage at Inv input with op amp saturated.
Then Vref = VsatR2/(R1+R2)
Let Vs = Vsat + Vref.
Let Tau = R3C
Let f = frequency of oscillation.
f = -1/{2 Tau ln[(Vs-V)/Vs)]}
Where ln(X) means "natural logarithm of X".
If the +/- saturation voltages are equal, the ouput frequency will be
independent of the supply voltage.
~
A pot can be substituded for any of the resistors.
~
Ignore my last post. I accidentally clicked "Send" before I was
through editing.
Regards,
Jon

 

[Chris Jones]

Hey there,
I'm trying to figure out how to output a clock frequency depending on
the position of a potentiometer slider.

I need the clock frequency to range from 39KHz to 49KHz and the
mid-point to be 44.1KHz. I've been informed a DCO would be the best
option but I'm not sure if their output range can be that finely tuned?
All the DCO's I have found go up to around 20MHz and do not seem suited
to a lower frequency task like this.

Does anyone have any advice as to how to finely tune an output
frequency of one of these devices?

My original thoughts were to use an ADC to sample in the position of
the slider and then use a DCO. Are VCO's not stable enough? I need the
clock to be accurate and precise...

Any help greatly appreciated! Many thanks, Alex.

I'm guessing you want this for a digital audio clock. I would suggest a DDS
chip such as AD9850 or one of the slower versions. They go down to 1Hz or
less, and can be programmed with a PIC microcontroller such as a PIC16F84A,
and you would also have to hook up a serial output ADC to digitise the pot
position. I think that this solution would produce a less noisy (lower
phase noise) clock than any analogue method that I can think of, and it
would be very repeatable and manufacturable. If lower performance is OK,
or lower power consumption and lower cost interests you then some kind of
VCO like the one out of a 4046 IC might do what you want.

Chris

 

[Joerg]

Hello Chris,

... If lower performance is OK,
or lower power consumption and lower cost interests you then some kind of
VCO like the one out of a 4046 IC might do what you want.

That one has quite poor linearity. The LM331 is much better. AD also
offers a few V/F converter chips that are top notch but cost a bit more
than the LM331.

It is also possible to linearize the 4046 via a feedback loop but then
the cost approaches that of the LM331.

Regards, Joerg

 

[Fred Bloggs]

Hello Chris,



That one has quite poor linearity. The LM331 is much better. AD also
offers a few V/F converter chips that are top notch but cost a bit more
than the LM331.

It is also possible to linearize the 4046 via a feedback loop but then
the cost approaches that of the LM331.

Regards, Joerg

Doesn't like being pushed beyond 10KHz...

 

[Fred Bloggs]

Hello Alex,

Look at the National LM331.

Regards, Joerg

Doesn't like being pushed beyond 10KHz...

 

[Joerg]

hello Fred,

Doesn't like being pushed beyond 10KHz...

It can do up to 100kHz but you'd have to spring for an opamp if
linearity errors really have to be under 1/10th of a percent.

Regards, Joerg

 

[Joerg]

Hello Fred,

Doesn't like being pushed beyond 10KHz...

Does. Dunnit...

Regards, Joerg

 

[[email protected]]

Thanks guys for all the advice - I'm going to try a few of the methods
suggested and see what works best.

I will update you with news of my progress! Best regards, Alex