Issue with AD8039 Producing 5MHz Triangular Wave

Hi,

How does your hardware look like? Could you provide a photo? I guess on the scope the output of U1 can be seen. What is the waveform at the output of U2?

Zoltan

Hello ,

Yeah on the scope its output of U1 . The output of U2 is shown in yellow wave form  . The second image is hardware schematic.

Thanks

Hi,

I think the supply may not be decoupled properly. Since IC1B works as a comparator, the output is saturated to the supplies, but as you see, it is not a clean rectangular wave, therefore the integrator built using IC1A can’t output a nice triangular wave. In the simulation the supply is perfect. So the supply here probably has a considerable high frequency AC component.

According to your schematic, all decoupling capacitors are 10uF, the 100nF capacitors are missing (they are recommended for proper high frequency decoupling). You only shared the schematic, the photo of the hardware could show possible wiring/layout problems too. The other parts can also affect the supply.

So I recommend to use 100nF ceramic capacitors for decoupling (in addition to the 10uF ones) very close to the ICs, and to test the triangle wave generator separately to be sure that the supply is not affected by the other components (IC2 and IC3). Be sure that IC3B’s inputs are not floating, I do not see it on the schematic.

Let me make one more note. I think it would be better not rely on the supply concerning the input of the integratror. You can use an SPDT analog switch (2:1 analog MUX) that has two clean DC analog inputs and use the output of IC1B to select between these voltages. This way you can have a nice rectangular waveform as the input of the integrator.

Zoltan

Hello,

Thank you for your reply. I tried adding a 100 nF capacitor in parallel with a 10 µF capacitor, but the result was the same. My main problem is that the frequency is limited to 2 MHz. Could you tell me whether the AD8039 is suitable for use as a comparator at 5 MHz? Thanks

Nasir

Hi,

According to your simulations, the opamp should be OK (although I think it is always better to use a comparator rather than an opamp as comparator). The frequency is determined by R2, C1 but also by the ratio of R5 and R3. You can try to play with their values to get higher frequency. Don't forget to test your circuit without the additional parts (IC2 and IC3) and be sure that you have proper layout, grounding. 

I still think the main problem is with the supply, I would use an analog switch to get supply-independent square wave. 

Zoltan

Hello.

I just test with different combination of R2 and C1 . The output waveform is pure triangular wave at 800KHz as shown in figure but  when i goes above 1.5MHz signal is distorted . I think i am over driving .

Hi,

How does the rectangular wave look like at 800kHz? It must be a quite nice square wave.

Try also to increase R5, that should increase the frequency. Although the amplitude will be smaller, you can have higher frequency. Choose R2 and C1 to get the highest frequency good triagular wave, then try to change R5. 

Zoltan

Hello Zoltan.

Thanks for your suggestion . I try as you said and i am able to generate 3MHz triangular wave without distortion as shown in figure.  

But after its stuck to 3Mhz even i increase R5 until 1M ohm no change in frequency only amplitude is changing. Can you suggest something. Thanks

Farhan 

The ratio of R5/R3 matters, so you can also try to reduce R3. Keep R3+R5 between 1k and 10k, and change their ratio. This way you won't have too high or too low values. This way you can avoid the effect of parastic capacitances or overloading. 

So you mean the addition of both R3 and R5 has to be between 1K to 10K.? Or you means the ratio has to be 1:10?