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.

Register Log in

Why this oscillator does not work?

Status
Not open for further replies.
Z

zeller

Member level 5
Joined
May 9, 2010
Messages
84
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,286
Location
london
Activity points
1,959
Hi,please help .I have designed this colpitts osscillator (27 mhz) It worked well when simulated on multisim i.e there were oscillations. I applied the schematic on a circuit board yet when I used an oscilloscope to measure , I couldn't see any oscillations.

How can a schematic work on simulation and then not work in practice?

Is Multisim a useless programme? I have checked and I'm sure 100% that the components have been correctly placed on the circuit board!
 

Attachments

  • oscilllator2.doc
    23.5 KB · Views: 255

Those are highly suspicious values. The collector and emitter resistor are out by a factor of about 100. Try making R2 1K and R3 5.1K although swapping them around looks more realistic. You will also get better results by center tapping the tuned circuit, in other words, make C3 and C4 about the same value. You may have to adjust the inductor to re-tune back to 27MHz.

Brian.
 
  • Like
Reactions: zeller

    Z

    zeller

    Points: 2
    Helpful Answer Positive Rating
betwixt said:
Those are highly suspicious values. The collector and emitter resistor are out by a factor of about 100. Try making R2 1K and R3 5.1K although swapping them around looks more realistic. You will also get better results by center tapping the tuned circuit, in other words, make C3 and C4 about the same value. You may have to adjust the inductor to re-tune back to 27MHz.

Brian.
Click to expand...
Thanks for the reply!
So why did the simulation work on Multism?
 
Last edited:

Likely because you did not put the coil Rs and stray capacitance in the Multisim model. It does not take much stray capacitance when you have such high value resistors. I am surprised it even worked in sim with less then 8 uA of collector current. Your loop gain will be extremely low.
 
  • Like
Reactions: zeller

    Z

    zeller

    Points: 2
    Helpful Answer Positive Rating
betwixt said:
Those are highly suspicious values. The collector and emitter resistor are out by a factor of about 100. Try making R2 1K and R3 5.1K although swapping them around looks more realistic. You will also get better results by center tapping the tuned circuit, in other words, make C3 and C4 about the same value. You may have to adjust the inductor to re-tune back to 27MHz.

Brian.
Click to expand...
Okay I have changed the resistor vales R2 and R3 as you suggested but I haven't changed C3 and C4. The circuit seems oscillate (on simulation) to about 27.7 mhz which is okay for me, do I still need to change C3 and C4 for it to oscillate in reality? I got a he folmula from a book which recomended that C3 should be at least 3 to 10 times greater than c4.
 

zeller said:
Okay I have changed the resistor vales R2 and R3 as you suggested but I haven't changed C3 and C4. The circuit seems oscillate (on simulation) to about 27.7 mhz which is okay for me, do I still need to change C3 and C4 for it to oscillate in reality? I got a he folmula from a book which recomended that C3 should be at least 3 to 10 times greater than c4.
Click to expand...

Increase base capacitor to 1000 pF or more. If still does not oscillate increase 39 pF coupling cap at collector.
 

Hi,

Your oscillator is supposed to be a colpitts with common base configuration.

At 27Mhz, C5 and C2 must be seen as shorcuts but 39pF is a too much low value for that (they are acting like 154 ohms resistances; use for example 100nF).

Current in emitter is quite low (about 7uA). This increases the small-signal resistance rbe, and so decresing the amplifier gain (Ao = R2/rbe). If I am not wrong, your gain is lower than 1 !

In this kind of oscillator, Ao must be greater than (1 + C3/C4) = 6 and this is absolutetly not the case.

In conclusion your original circuit should not maintain oscillations. Be carefull with the simulations parameters. How long are you simulating ? Note that oscillations will anyway start because of the LC configuration but they should not be kept oscillating for long time. May be you are watching the starting oscillation phase.

Let us know about your progress on your project !

Efron
 
Last edited:

efron said:
Hi,

Your oscillator is supposed to be a colpitts with common base configuration.

At 27Mhz, C5 and C2 must be seen as shorcuts but 39pF is a too much low value for that (they are acting like 154 ohms resistances; use for example 100nF).

Current in emitter is quite low (about 7uA). This increases the small-signal resistance rbe, and so decresing the amplifier gain (Ao = R2/rbe). If I am not wrong, your gain is lower than 1 !

In this kind of oscillator, Ao must be greater than (1 + C3/C4) = 6 and this is absolutetly not the case.

In conclusion your original circuit should not maintain oscillations. Be carefull with the simulations parameters. How long are you simulating ? Note that oscillations will anyway start because of the LC configuration but they should not be kept oscillating for long time. May be you are watching the starting oscillation phase.

Let us know about your progress on your project !

Efron
Click to expand...

Thanks ERON for your insight which is very helpful, I had tried to change values of those resistances as suggested by betwixt (above) but although i got oscillations when simulating with the programme (multisim) - The circuit did not work in reality(on a circuit board)

Now you said 'Be carefull with the simulations parameters. How long are you simulating ? Note that oscillations will anyway start because of the LC configuration but they should not be kept oscillating for long time. May be you are watching the starting oscillation phase'

So how do I 'correctly' simulate on Multisim? I usually run the circuit and attach the 'tetronix' oscilloscope, which display virtual sine wave, and I simulate for on average 5 minutes. So how do I simulate correctly as you suggest? I got the capacitor values from a certain book which gives folmulas for appropriate capacitors for given frequencies.

Please help as I want this project to work!!I kind of nearly giving up now!

Many thanks.
 

View attachment osc27.doc

I changed your circuit just enough to make it such that it should oscillate.See the link above to the attached doc.
I analysed it open loop to look at a bode plot of gain/phase and it is a good starting point.

You should be able to improve the basic circuit to get a sharper phase response etc but I'll leave that up to you...
 
Last edited:
  • Like
Reactions: zeller

    Z

    zeller

    Points: 2
    Helpful Answer Positive Rating
GOHZU,
Thanks for this, so which programme did you use on your design? And what should I do to get a sharper phase respone?By varying capacitor values?
 

I used genesys to quickly check a bode plot.

However, because you are running from 9V I think you can get better performance if the emitter resistor is a larger value than in my first circuit.
eg set it to 2k Ohm to save on DC current (was 560 ohm)

Are you worried about DC power requirements and noise performance?

I only looked at this very briefly so you still have work to do to optimse this oscillator but at least it should oscillate if you build it as per my changes in the doc.
 
  • Like
Reactions: zeller

    Z

    zeller

    Points: 2
    Helpful Answer Positive Rating
zeller said:
So how do I 'correctly' simulate on Multisim? I usually run the circuit and attach the 'tetronix' oscilloscope, which display virtual sine wave, and I simulate for on average 5 minutes. So how do I simulate correctly as you suggest? I got the capacitor values from a certain book which gives folmulas for appropriate capacitors for given frequencies.

Please help as I want this project to work!!I kind of nearly giving up now!

Many thanks.
Click to expand...

Dear Zeller,

You are saying that your are building your project on a circuit board (I hope it is not a breadboard as over 1 Mhz it should never be used !).

By selecting betwixt's proposal above (R2 = 1K, R3 = 510K), emmitter current (in DC) is around 0.78mA and rbe ~ 32 ohms, which leads to a gain in open loop of about Ao = R2/rbe ~ 31, which is now much bigger than (1+C3/C4) = 6. Therefore the circuit must now oscillate, but only if C5 and C2 are seen as shortcuts at 27 Mhz. Please use 100nF values if not yet done.

Because of the great difference between Ao=31 and (1+C3/C4), the output is supposed to include much distorsion (a lot of harmonics) --> what are you seeing in your simulations ?

Where are you taking the output signal, at Q collector or emitter ?

You may show to us your simulation graphics !

Note that when measuring with real oscilloscope, you are loading the circuit, usually with a 10K resistance in serial with a 20pF capacitor until ground. But for the moment I can not say if this is having a negative effect on your circuit.

Waiting for good news !
 
Last edited:

efron said:
Dear Zeller,

You are saying that your are building your project on a circuit board (I hope it is not a breadboard as over 1 Mhz it should never be used !).

By selecting betwixt's proposal above (R2 = 1K, R3 = 510K), emmitter current (in DC) is around 0.78mA and rbe ~ 32 ohms, which leads to a gain in open loop of about Ao = R2/rbe ~ 31, which is now much bigger than (1+C3/C4) = 6. Because of this, the circuit must now oscillate, but only if C5 and C2 are seen as shortcuts at 27 Mhz. Please use 100nF values (if not yet done).

Because of the great difference between Ao=31 and (1+C3/C4), the output is supposed to include much distorsion (a lot of harmonics) --> what are you seeing in your simulations ?

Where are you taking the output signal, at Q collector or emitter ?

You may show to us your simulation graphics !

Note that when measuring with real oscilloscope, you are loading the circuit, usually with a 10K resistance in serial with a 20pF capacitor until ground. But for the moment I can not say if this is having a negative effect on your circuit.

Waiting for good news !
Click to expand...

Thanks for this, I changed the resistors but did not change the 39pf capacitor, I implemented the circuit on a circuit board but upon checking with a real oscilloscope, it did not oscillate, I will now change the 39pf capacitors to 100nf and see what I get.

We keep you posted.
 

zeller said:
Thanks for this, I changed the resistors but did not change the 39pf capacitor, I implemented the circuit on a circuit board but upon checking with a real oscilloscope, it did not oscillate, I will now change the 39pf capacitors to 100nf and see what I get.

We keep you posted.
Click to expand...

Hi Zeller,

Any progress ?
 

efron said:
Hi Zeller,

Any progress ?
Click to expand...

Unfortunately not much progress yet.I have changed the circuit as suggested (see attachment) and tried to measure oscillation with a real 35 MHZ oscilloscope (I implemented the circuit on a circuit board).When I switch on the circuit, the green line on the oscilloscope moves up(as if indicating that there is a sort of a voltage) but the line remains there(straight) it does not oscillate! I tried to change scope parameters, but still there are no oscillations.I thought I made a mistake on a circuit board but all looks fine,is there a problem with my scope?By the way this is a second hand scope I bought recently although so far I have only been able to measure DC currently accurately(never measured a sine wave) so I assume it works!!!
 

Attachments

  • oscilllator3.jpg.doc
    24 KB · Views: 136

R3 seems too high. Try around 220-470Ω
Make sure you have the transistor pin-out correct.
 
Last edited:

The oscillator should be able to work with the current resistor values. However, the L/C ratio seems too low -- I would use 1...2 uH (or more) for the inductor. Also, the 100 nF capacitors may have a relatively large parasitic inductance at 27 MHz, resulting in loss of signal; therefore, 10 nF (or less, maybe as little as 1 nF) might give better results.

The gain most likely would be greater if the 1 k collector load resistor were removed and L1 moved to its place. The power supply then would need to be properly decoupled for the frequency of oscillation.
 

5k will be a very bad match to the capacitive tap point in the tank circuit resulting in low loop gain and low oscillation amplitude.
 

Your oscillator coil has probably a Q-factor of only 20 if it is SMD type in 0603 or 0805 package. So simulate with such a low Q-inductor with 470nH (Z= j76 Ohm @ 27MHz) and add a series resistor with 4 Ohm in your simulation. You get better Q-facor with more inductance and bigger size. I would use about 1µH. Couple the inductor not full parallel from collector to + supply. The transistor reduces the unloaded Q too much and by this the phase noise becomes bad. I suggest following circuit: The best circuit is collector direct connected to +9V and a parallel LC resonant cicuit loosely coupled (for example with 39pF) to the base. Apply 68pf between base and emitter and 56pf from emitter to ground to achieve negative input resistance at the base. Use about 2,7 kOhm from emitter to ground for the DC. This gives you about 1mA current and the oscillator will work fine. For base bias use 22kOhm to gnd and about 33kOhm to +.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top