works at 1Mhz but not above 10Mhz

[srinivasa]

Hi

I am facing a problem in my project. i attached circuit, and datasheets.
In my circuit when i apply a signal of 1 MHZ at J2 (BNC connector) and i can see the output signal is square wave (LVDS signal) at U23 pins 2 and 3 on oscilloscope e.When i apply a signal above the 10Mhz i am not able to see the square wave but a triangular wave and voltage levels are in millivolts.the output signal should be square wave (0 to 3v with respect to GND).Help me if any one knows the solution.

Thanks in advance
srinivasa

 

[keith1200rs]

What values are R33/R34/R35? What amplitude & offset is your input signal?

Keith.

 

[karesz]

Hi,
sorry to ask, but what kind of scope probes are you used?
If its 1:1 (or switched to 1:1_has a bandwidth of ca. 20MHz...), or is a bandlimit filter switched to 20MHz?
Btw; why did you U22 from 5V supplyed?
Another question: is it not so, that for properly testability you must apply a 100 Ohm load resistor on the diff or LVDS output!?
K.

 

[design_oriented]

What is the output of the comparator for the 10MHz square wave input?

If you see that it is the same as the output for pins 2,3 then maybe the problem is your comparator.

As you see a square wave convert to a low voltage triangular wave, what you are seeing, may be your supply ripple voltage. You can verify this by measuring the frequency of the triangular wave.

The datasheet for the comparator suggests to use a 10 nF supply bypass capacitor and you have used 100 nF but this may not be your problem.

 

[FvM]

A LVDS driver is not intended to give a 3 V square wave. The specified output is about 300 mV "Low Voltage Differential Signal" into a 100 ohm load, as the datsheet clarifies.

 

[srinivasa]

R33 =18.7K and R34=1k ,votage divider gives a 0.25v(at pin 3) and voltage applied at J2 is 6db (1.262Vpp,sine wave)
I am measuring the LVDS signals(each) with respect to GND. for 1Mhz output is 0 to 3.3V square wave.
The Osilloscope has 100MHz band width.
Any techniques are there to measure the signal or any circuit or procedure to follow how to measure the high frequency signals.

My requirement is to apply a 6db, 30MHZ sinewave and to get 3.3V(with respect to GND,not the differential) at the output.

i attached the same circuit with resistors values

 

[keith1200rs]

Try adding a 100 ohm load to the LVDS outputs.

You won't get a 3V signal out of the LVDS - that is what LVDS is about. You could just use the comparator output maybe?

Do you get correct signals out of the comparator at 10MHz?

Keith

 

[karesz]

Hi Srinivasa,
I would say; apply pls a reference voltage of ca 2.5V at pin 3, but with some good bypassing as 100nF X7R type and an AC coupled input signal.
Is your sine wave signal GND symmetric?
If you say: "I am measuring the LVDS signals(each) with respect to GND. for 1Mhz output is 0 to 3.3V square wave." > you means on LVDS31/pins 2&3?
I think its inpossible_as said Frank & Keith too!
LVDS is per default/standard a signal/pulse between 1,1...1,3V (around hes threshold of 1.24V)...
Maybe is some wrong at Power & G/Gb pins

As said too: check pls first the comparator output alone (separate from LVDS31) for your 30MHz!

I think, that your comparator works not in reality_so far Uref at 0.25V is-can not. It needs a threshold what in accordance of input signal rail is.

As said, in my opinion is better to use Uref at 2.5V& AC coupled input_than is the distortion of duty cycle smaller too as in these situation ...
Good progress!
K.