1. We will be in Read Only mode (no new threads, replies, registration) starting at 9:00 EDT for a number of hours as we migrate the forums to upgraded software.

Discussion in 'Homework Help' started by hasie, Oct 2, 2009.

1. ### hasie Thread Starter Member

Sep 28, 2009
35
0
Hi Guys
Im new to this site, and im really looking forward to explore further. By the topics I have read sofar, this forum really seems informative.

Dont really know your full rules yet, but I would really appreciate it if anybody could give me help on my problem.
I did run a search on the forum, but did not get my answer that i am wondering about.

I am currently busy trying to build a RC Phase Shift Oscillator.

I see there is alot of info available on the net, but not any info sofar that I specifically found that I am looking for.

Most sites explain the works of the phase shift oscialltor with op amps.
My instructions are:
* Must be designed with BC547 transistor
* Frequency of oscillation must be 1KHz
* Must be as close as possible to a perfect sine wave
* Signal Amplitude must be atleast 2v(p2p)
* Must be operated with a 9v supply

I am not too farmiliar with these oscillators, so this week searched for as much info as I could.

I found a circuit at http://www.talkingelectronics.com.au/projects/200TrCcts/200TrCcts.html#37 that somebody told me works perfectly.

I then used the formula to get the required resistance values and cap values for the required frequency.( F = 1/((2PI)(R)(C)(sqrt6))

an accepable value to the frequency to caps that I have with me, was 4.7nF capacitors, and 13.8K resistors(12k and 1k8 in series).

The other 2 resistors I left as is at 1M and 3k3.

On my simulation program it runs smoothly and just short of 1KHz, allthough when I built the circuit it gives a saw tooth(triangular wave) waveform.

I have no clue as to how to continue. Can anybody maybe shed abit of light that maybe know, or maybe have a better circuit for me?

ANy help much apprecated

Many Thanks

2. ### Audioguru Expert

Dec 20, 2007
11,251
1,350
The simple single transistor phase-shift-oscillator should give a clipped sine-wave output. The output is clipped because the circuit has nothing to prevent the transistor from going into saturation. The transistor does not have much negative feedback so it produces distortion.

Check the pins of your transistor and the wiring of your circuit to see why it produces a sawtooth/triangle waveform.

3. ### Wendy Moderator

Mar 24, 2008
21,839
3,049
A simple light bulb can be used to create an AGC (automatic gain control) effect. Why not post your simulation?

4. ### hasie Thread Starter Member

Sep 28, 2009
35
0
My simulation on proteus?

How do I post my simultion? WIth a screenshot?

Many thanks guys, will recheck my circuit again quickly allthough I am almost sure its right..

How do I post simulation? Sorry im new to this

Many thanks

5. ### hasie Thread Starter Member

Sep 28, 2009
35
0
And is this circuit relatively stable and accurate? Or is there a better circuit I can use to get a stable 1KHz sine wave?

Many thx

6. ### Wendy Moderator

Mar 24, 2008
21,839
3,049
An RC oscillator is never going to be as accurate as a crystal oscillator, but it is accurate enough. I was asking for the schematic locally with the adjusted values. It might give us some insight as to what's happening.

Looking at your schematic, try adding a variable resistor to the emitter. This is better negitive feedback than the base resistor going to the collector. It will control the gain much better.

This transistors biasing needs a lot of work. This is the fundimental reason the waveform is so off.

Last edited: Oct 3, 2009
7. ### hasie Thread Starter Member

Sep 28, 2009
35
0
A RC oscillator was askes so I just hope to get a good design.

Here is my circuit at this moment

Ill play around with a variable resistor on the emitter.

many thanks for the help sofar

Any other suggestions?

Thanks

8. ### hasie Thread Starter Member

Sep 28, 2009
35
0
When adding a 100ohm var resistor to the emitter the amplitude on the oscilloscope gets smaller.

And then when I remove the 1M resistor going from the base to the collector it does not oscillate at all?

Im going to rebuild this circuit again quickly on breadboard. and test it again. any input appreciated.

Many Thanks

Last edited: Oct 3, 2009
9. ### hasie Thread Starter Member

Sep 28, 2009
35
0
Rebuilt the whole circuit on different breadboard, different components. Still doesnt work.

See also the thing that doesnt make sense to me, is that there are 3 caps in the RC network but only 2 resistors. is the f = 1/((2PI)(R)(C)sqrt6) still correct?

I can use this circuit. http://www.electronics-tutorials.ws/oscillator/rc_oscillator.html

But I have no idea what to make R4, R5, R6 and C4?
Would this circuit be better to use?

Any ideas appreciated,

Last edited: Oct 3, 2009
10. ### Wendy Moderator

Mar 24, 2008
21,839
3,049
I like the second circuit much better, it addresses my points about the transistor bias. Thing is, a 3 leg RC oscillator should have matching RC legs, the first design only had 2, the third they were hoping the transistor would be close enough to the correct impedance.

The second circuit actually puts a 3rd RC leg in there. Try using similar bias resistors, and see it if works. You can adjust the gain of this circuit by putting a variable resistor in series with the capacitor, which will allow you to adjust the gain of the transistor amp. While I don't know what the resistor needs to be it will be quite low in value.

Here's Wikipedia's take on it.
http://en.wikipedia.org/wiki/RC_Phase_shift_Oscillator

11. ### hasie Thread Starter Member

Sep 28, 2009
35
0
Many thanks. Ill try this one quickly..

Will play around with values and keep you posted.

Many thanks again

12. ### hasie Thread Starter Member

Sep 28, 2009
35
0
Cant even get this one to oscillate *sigh*
....l

found info relating.
very informative, very good explanation aswell. It actually goes through the steps to design a set oscillator.

Will go through this tutorial and see if I come right.

Many thanks for the help guys wil keep you posted

Last edited: Oct 3, 2009
13. ### The Electrician AAC Fanatic!

Oct 9, 2007
2,724
496
Since you have been asked to produce a good sine wave, you would probably be better off to use the low-pass version of the phase shift network rather than the high-pass version that you have been using so far.

See figure 6 in this file:
www.ti.com/sc/docs/apps/msp/journal/aug2000/aug_07.pdf

They use an op-amp, but it's the form of the phase shift network you want to copy; resistors in series, capacitors in shunt. This version of the network attenuates the harmonics of the somewhat distorted sine wave at the collector, rather than emphasizing them, as the high-pass version does. You might have to put a large value DC blocking capacitor in front of the network because with the low-pass version there's a DC path through the network which could upset the bias.

The frequency of oscillation is different; instead of the √6 factor you will have a 1/√6 factor (if the RC sections have the same component values; see below).

Another factor to consider that hasn't been mentioned yet is the loading effect at the input and output of the phase shift network.

The formulas you will find for the frequency of oscillation assume that the network is driven from a zero ohm output impedance source, and is loaded with an amplifier with infinite input impedance. This is not the case in single transistor amplifiers; opamp versions of the oscillator much more closely approach this ideal.

This doesn't mean that the oscillator won't work, but it does mean that the frequency of oscillation will be different from that predicted by the simple formula, and the required gain for the amplifier will be different also (probably higher).

Each successive RC section of the phase shift network loads the previous section, and this is responsible for the gain requirement; a gain of -29 is required for the standard 3 section network (equal value Rs and Cs in each section).

You can reduce the gain requirement if you taper the network. That means to increase the impedance level of each successive section so it doesn't load the previous section so much. If the first section has component values R and C, then you could make the next section with values of 2R and C/2, and the next with 4R and C/4. Tapering the network like this reduces the gain requirement, and changes the frequency of oscillation. If you carry this too far, then the transistor itself will load the last section too much.

With a reduced gain requirement you could have an unbypassed emitter resistor which would reduce the distortion at the collector.

If you adjust the gain of the amplifier by varying the value of an unbypassed emitter resistor until the output is just barely clipping, you should be able to get a fairly good sine wave.

14. ### hasie Thread Starter Member

Sep 28, 2009
35
0
thanks for the replies!

You guys are really a bunch of helpfull people!

I have managed to build a working prototype the levels out at 1060 Hz! (Which I think is close enough to the asked specification) It changes with temperature, and is between 1007 and 1105 on a frequency counter, which I think isnt too bad? It is stable and changes slowly on account of temp. Earlier today it was warmer but at night it went down to 1007.

THe last tutorial I posted really helped alot, and I actually understand the circuit compared to just copying and pasting because all the values were calculated step by step..

I used a 4 RC circuit and the R values are 1k33 and C values were 100nF.

Very happy will be basing my report on all the calculations aswell.

Many thanks for all the help guys. I really appreciate it

Thanks Electrician I think I will still change the emitter resistor to set it just beofre the wave gets clipped for a perfect sine wave as you have explained....

Thanks
Cheers

Last edited: Oct 3, 2009