Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

32.76 kHz Colpitts oscillator circuit

Status
Not open for further replies.

vladislovas123

Newbie level 3
Joined
Oct 25, 2018
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
41
Hello, everybody. I'm trying to construct a Colpitts oscillator circuit for 32.76 kHz. But after soldering, the circuit is not working.

What have I done:
• checked the circuit.
• checked the values to apply Barkhausen criteria (not sure about this part)

The only thing I get at the output is 5V (measuring with oscilloscope).

P.S. Maybe somebody know better working circuit to get stable 30-50 kHz oscillations (according to my needs solution should be cheap enough)

Untitled-1.png

Thanks for your attention

- - - Updated - - -

Almost forgot to mention that I need sine (or almost sine) wave oscillations. There is a huge amount of ready made circuits for 32khz square oscillations.
 

What is the NPN value? Did you simulate the circuit?
 

Thanks for your answer.

Tranzistor is 2N3904. I didn't do simulation , but I am working on it right now (will post LTspice project as soon as possible)
 

I've done a fair amount of Colpitts oscillators with crystals, and they usually work of the bat.

I have also done a 32.768Khz with a CMOS inverters, and what I have seen is that the 32.768 Khz tuning fork element (they are tuning forks, not crystals) do not like to be driven hard. I remember in one instance, I had to reduce the supply voltage from 9V to 3 volts for the oscillator to work. In other instance, I had to put a large series resistor to reduce the drive level to the tuning fork.

I don't know whether my experience can be applied to your circuit, but perhaps you could try to use a lower supply voltage, just as an experiment.
 
Last edited:

According to datasheets, 32 kHz crystals have ESR in several 10 kOhm range, means the shown circuit has too low impedance bias network.
 

The 2N3904 has the hfe in the range of 100-400 whereas at the 548C for example, this parameter is in the range of 400-800. You could replace to see if there is any change; this would increase the chance to meet the Barkhausen criteria.
 

Have not tried it with tuning forks.
Expect similar problem as with post #1 schematic. Notice that 32 kHz tuning fork crystal ESR is factor 100 to 1000 higher than MHz crystals. Need at least higher impedance bias network. NJFET instead of NPN may be a good idea.
 

chemelec, it is not a sine wave oscillator.
 

32 kHz watch crystal models have been posted along with oscillator circuits in the Yahoo Ltspice forum https://groups.yahoo.com/neo/groups/LTspice/info mostly by Helmut Sennewald.

A common attribute is the high impedance bias network and µA transistor current.

osci_32ka.PNG

The start pulse is a simulation feature and not used in a real circuit.
 

chemelec, it is not a sine wave oscillator.
I Know, But I wasn't considering the Topic of a Colpitts as being important.

- - - Updated - - -

I Know, But I wasn't considering the Topic of a Colpitts as being important.

You could add a Small Inductor and Capacitor on the output to get a Sine wave.
 

Chemelec's schematics (post #10) have the 4069 invert-gate. Although the 4069 normally changes output state abruptly, the UBE (unbuffered) version can be made to behave as an analog device. A string of them (with proper resistor values) amplify audio. I first saw a similar circuit in the Forrest Mims Engineer's Notebook.

A successful sinewave is reported in this thread (despite its title):

https://www.edaboard.com/showthread.php?327831-Can-t-make-oscillator-with-CMOS-inverter

Also discussed in:

https://www.edaboard.com/showthread.php?49829-Will-a-CMOS-inverter-work-as-an-Amplifier

https://www.edaboard.com/showthread.php?304771-why-using-CMOS-4069-inverter-as-ampilifer
 

The Colpitts oscillator uses a COIL. That is what gets you a gain greater than unity to oscillate. The junction of the two capacitors works as a midpoint tap when resonating with a coil. Thai is a 2:1 relation from base to emitter. A crystal does not do that. You have no gain here. You can try to place a coil in parallel with the crystal, that resonates close to the oscillator frequency. But 47 pF for 32 Khz??? 47 nF may be closer to what you need. I would just go to a cmos oscillator and shape the output with an LC circuit to have sine wave.
 

The Colpitts oscillator uses a COIL. That is what gets you a gain greater than unity to oscillate. The junction of the two capacitors works as a midpoint tap when resonating with a coil. Thai is a 2:1 relation from base to emitter. A crystal does not do that. You have no gain here. You can try to place a coil in parallel with the crystal, that resonates close to the oscillator frequency. But 47 pF for 32 Khz??? 47 nF may be closer to what you need. I would just go to a cmos oscillator and shape the output with an LC circuit to have sine wave.
Crystal model contain coil, which can be 10Henry or more, so capacitance range can be in the range of pFarad at 32kHz. Crystals have series and paralel resonace frequency, the circuit around it will determine that at which resonace frequency it will oscillate, at colpitts it is the series resonant frequency like in most of the cases. Tap point at Colpitts is used to get enough phase shift, and gain is generated by the resonant tank's quality factor. If you put inductor in paralel with the crystal you push up the resonace frequency, and decrease the quality factor and loop gain. At colpitts the non-linear components in its output are smaller, and it is not necessary to use additional filtering, depends on application. Shaped squre wave oscillators, like Pierce oscillator requires more filtering, which in some cases is not the simplest solution.
 

The Colpitts oscillator uses a COIL.
Not necessarily, a crystal can be used in place of the LC resonant circuit.

The circuit in post #1 is designated Colpitts oscillator in literature. It works easily with MHz crystals, however a 32 kHz watch crystal has relative high impedance in series resonance, respectively the connected circuit must use MOhm resistance level, the original poster missed this point.
 

Thanks for the impedance clarification. That is what I meant about the use of 75 pF. These have 64 K (z)ohms at 23 kHz.
For this to work, I would try a Jfet, remove R1 and use something in the MOhms range for R2 (anything 1 to 10 MOhms may work) and use a bigger value for the source (ex emitter) resistor. somewhere in the range 10 to 47k. If this still doesn't work, I would go to an OP AMP solution. MHZ crystals are easy. 32KHZ crystals are 'stubborn'.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top