Variable frequency sinusoidal wave oscillator

Status
Not open for further replies.
The original design is using two current sources, a permanently enabled negative source and a positive source switched by diodes. For a symmetrical triangle, the positive source has to be programmed to double the current of the negative one.

In addition, the generator is self-oscillating by feedback through a comparator with hysteresis.
 
Thanks FvM For your help. What kind of comparator i can use , here?
And how can i get feed back from that to obtain auto triangle oscillations?
Thank you
Goldsmith
 

Selecting a comparator is a detail engineering question. Something you can play around with after deciding for a basic oscillator structure. I already mentioned, that the discussed FG1617 circuit uses a line receiver component, not a typical comparator but apparently serves the purpose.

To emphasize the principle operation, I sketched a LTSpice simulation circuit with ideal current sources, the diode current switch, and a simplified OP model used as comparator. You can adjust slew rate and GBW accordingly to make it act as a fast comparator. In the simulation example, it's operating in a slow kHz range.

 

Attachments

  • triangle gen.zip
    647 bytes · Views: 118
Dear FvM
Hi and thank you for , your helps to me .
I simulated , that circuit with programmable , current source , but it created , ramp wave , and when i tried to change the currents to improve that , it stopped oscillations . ( at LT spice , file that you attached , when i changed the current of current source , the wave form , changed , from triangle , to ramp . but with real current source , that i used , that problem occurs) . where is the problem and how can i optimize , that?
Help me , please.

Appreciate
Goldsmith
 
Last edited:

According to the symmetrical comparator hysteresis, the lower current source must pull-down to -15V, not ground. The reference voltage should be adjusted.
 

Dear FvM
Thank you for your help . but as you can see at my last attachment , i pulled it down with R8. or probably is not your mean that ? i can't understand your accurate mean.
Thank you
Goldsmith
 

R5 needs to be connected to -15V as well, otherwise the current source will never reach -5V.

I have to add a comment about the positive current source. The LM358 input voltage range ranges to Vcc-1.5V. So the positive source can't work with shunt resistor voltage drops below 1.5 V, which seriously restricts the available current respectively generator frequency range. FG1617 is solving this problem with different power supply for the OPs.
 

Dear FvM
Why When i connected , the R5 to the -15 v , the circuit started the oscillations? ( i corrected the circuit as you can see at below: )


Thanks in advance
Goldsmith
 

Why When i connected , the R5 to the -15 v , the circuit started the oscillations? ( i corrected the circuit as you can see at below: )
Click to expand...
I guess, you meant, why did it not oscillate?

There are additional conditions, that must be met. The setpoints for both current sources should be referenced to positive and negative supply voltage, otherwise programmed currents are too high and the oscillator node is blocked.
 
Dear FvM
Thank you for your helpful guides , really .
As you said , that the currents are high , i decreased them , but again it didn't oscillate . where is the problem ?
With best appreciates
Goldsmith
 
Last edited:

Dear FvM
From last night until now , i trying to find the problem of my circuit , but i couldn't . is it possible that you guide me again , please?
Thank you
Goldsmith
 

I guess, the setpoints are unsuitable. Can you please post the zipped PSpice project?
 

Dear FvM
Again Hi
What is your idea about that ( Zipped file ) . could you find the problem?
Thank you for your helps to me.
Best Regards
Goldsmith
 

There are two points:
- current source setpoints
- a special SPICE simulation problem

For the setpoint problem, please review the modified schematic. I already mentioned it previously, but possibly wasn't clear enough.

The simulation problem is probably less obvious, at least I didn't think about it before. Normally, we would assume, that comparator is switching reliably due to positive feedback. It actually does in a real circuit. But in the simulation bias point calculation, the positive feedback of 1/3 is canceled by a negative feedback of nearly 1 through the diode switch path. The integration capacitor which delays the negative feedback is functionless in bias point calculation.

Applying an initial condition to the capacitor node solves the problem.

As an additional remark, a real LM118 isn't well suited as comparator, because it has an input voltage clamp circuit, that makes the positive feedback path feeding current to the negative input, distorting the triangle waveform. But you don't see it in simulation because the model misses the voltage limiting diodes.
 

Attachments

  • modified.pdf
    29.3 KB · Views: 90
  • IRI.zip
    16.1 KB · Views: 77
Dear FvM
Thank you very much . that worked very well . you are a very expert Engineer . Again thank you .
I have another question about the circuit of FG1617 . as i can see in its circuit , that used some digital gates , as a comparator ! ( with negative , supply !) what should i do about that . i used LM319 comparator , but it's behavior at this circuit was not good. can i use some gates , as i can see at FG1617 ?
BTW , why when you connected the end of that supply voltage , to the op amp supply , the problem solved?
Thanks in advance
Goldsmith
 

Dear FvM
Is it possible , that , you again guide me , please?
Best Regards
Goldsmith
 

The original FG1617 cicrcuit uses 75105 line receivers as comparators, I guess, because the designer didn't want to use an expensive high speed comparator. It's additional digital inputs are cross-coupled to form a set/reset latch. You can do similar things with other parts, but the circuit is rather sophisticated and you have to consider many parameters.

why when you connected the end of that supply voltage , to the op amp supply , the problem solved?
Click to expand...
You mean the current setpoint voltage sources V7 and V9? The usable setpoint range is limited, when the setpoint sources are connected to ground, you are outside this range. Please consider, that the setpoint voltage equals the voltage drop across R2 and R1, in so far it's reasonable to refer them to the +/- 15V rails. This is also done in the FG1617 frequency control circuit, see page A1.
 

Dear FvM
Again Hi
the behavior of frequency changer circuit at FG1617 is not clear to me , is it possible that you simplify it , please?
Best Regards
Goldsmith
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…