LMV761: LMV761 Output jitter issues

Hello Matei,

If you get  'jitter" in a certain range of frequencies, this usually points to some sort of system resonance in either the supply, grounds or output trace. Problems like this are usually layout related.

The LMV761 has fast output rise/fall times (2 ns), so it must be treated as a high-speed digital device, with all the precautions you would take for a 500MHz digital device. Tight layout. good supply bypassing, solid grounds, short traces..the works.

Try adding a 1000pF cap DIRECTLY between the V+ and V- pins. If you are driving a long output trace (>10cm), then add a 10 to 50 ohm series resistor at the output to help dampen any trace reflections.

And what are the input voltages? The input range is only 0 to 2.1V on a 3.3V supply. Make sure none of the inputs, including any "reflections", exceed these limits.

Can you provide a schematic?

Hello Paul,

Thank you very much for your comments!

The V+ has a maximum of 1.4 V or 1.5 V and the triangle has a peak value of 3.3 V, so it seems I am exceeding here with the triangle, right? Need to decrease the triangle.

Besides that, please find attached a snip with the layout area of the comparator, I tried to point out all the elements here, pretend there are jumpers for the 2-pin headers here.

This is a 2 layers design, so underneath there are only corresponding ground planes.

Thank you! Kind regards!

Matei

Thank you for your support, Kai.

Regards,

Joe

Hello Kai,

Thank you very much for your suggestions!

As a matter of fact, there is already a board based on the layout screenshot I showed so there is not much I can do right now to correct. I totally agree with the use of only one GND plane, it was clearly a mistake. I will change it in the next revision. But I could add series resistance on the comparator output and the triangle input, as you suggested.

Another thing to mention would be the fact that the triangle is generated from a DAC and due to resolution the small steps of the triangle rise and fall could be visible (only 12 bits) to some extent and maybe this could lead to parasitic switches of the comp out. Would this be a valid theory?

Thank you! Kind regards!

Matei

Hi Matei,

I am looking into an answer for your DAC question. I hope you will get a response by end of business Wednesday 11/18.

Regards,

Joe

Hi Matei,

for a simple test you can join the two ground planes together by the help of some wire connections, every centimeter or so :-)

Best do it on both sides of the printed circuit board.

Kai

Thank you for your input Kai.

Matei,

I am still looking for an answer to your question. In the meantime, I recommend trying Kai's suggestion.

Regards,

Joe

Thank you Kai, Joe! :)

I will certainly do that. 

Looking forward to hearing an opinion, would be interesting to see if the DAC, with its resolution (12 bits, 3.3V reference) can cause some jitter, instability at the comp out.

Kind regards!

Matei

Hi Matei,

it's difficult to answer this question without knowing the schematic. Having said this, it's well known that an ADC or DAC within a control loop showing a monotonicity error of more than 1LSB can even cause osillation :-)

Kai

Thank you for your support Kai.

Matei, I also think that your question is better answered if the schematic is known. I recommend simulating your circuit to better predict if the DAC will impact your comparator output.

Regards,

Joe