ltc6409 oscillation

Hi dyoung1a ,

Can you share your board setup and scope shots?

Thanks,

Mae

Hi Mae

I've attached a scope shot showing the oscillation at the output of the LTC6409.  Depending on the board this level may change anywhere between 100mVpp to 400mVpp.

The top layer and cutout under the LTC6409 are shown below.   The next layer under the cutout is a solid ground plane.

The microstrip transmission line starts at the D4/D5 clamp diodes, passes through a via from the top to bottom then runs along the bottom layer to the R128 termination resistor.  Channel B 100 ohm differential pair is 2248mil long shown on the left and Channel A is 1620mil long shown on the right terminating at R127. 

We found all channels oscillate with the stock schematic shown above but if we increase the LTC6409 gain to x3 (150 ohm input with 450 ohm feedback) and remove the 1.3pF feedback capacitor the oscillation stops.  Any idea on what causes this behaviour?

Hello, 

Can you provide your scope shot for these highlight nodes?

Thanks,

Mae

Hi Mae

I've attached the requested scope shots.  We don't have a differential probe so I've attached single ended measurements.  The first two are at R53/R54 followed by the ADC channel B input at pins 7/8.

Hello,

Thank you for sharing your scope shots. May I know how did you come up with this configuration? We are suspecting that the RC filter is the one causing the oscillations.

Thanks,

Mae

Hi Mae

The filter was recommended by Figure 55 in the AD9694 datasheet.  I've used a single 100 ohm termination instead of two 50 ohm termination resistors because of the differential transmission line.  We also require dc coupling so have removed the 0.1uF dc blocking capacitors.

Hello dyoung1a ,

Thank you for sharing this information. May I also know your input and where does it come from?

Thanks, 

Mae

Hi Mae

The LTC6409 input is driven by a AD8336 VGA at R36.  The overall circuit amplifiers TIA pulses but for all this testing the TIA output was dc (no light).  We also opened the VGA output pin and grounded R36 -- this had no effect on the oscillation behaviour.

We found in general the LTC6409 is very sensitive to any reflections on the transmission line.  The oscillations can be stopped as I mentioned above by pulling the 1.3pF capacitor and increasing the gain to x3.  But cooling the LTC6409 to around 10C causes the oscillations to start up again on some parts.   More testing found replacing the D4/D5 clamp diodes with 44pF capacitors stabilized the LTC6409 across all temperature.  This is not ideal however as it impacts high speed performance of the signal chain.

Hi dyoung1a ,

This configuration is for AC coupling using a BALUN.

You may need to adjust the anti aliasing filter depending on your needs. Please check this circuit note as an example: CN-0227 (Rev. A) (analog.com)

For DC coupling applications, were you able to follow these steps from the AD9694 datasheet?



Best regards,
Peevee

Hi Peevee

Thanks for your response.  CN-0227 uses ac coupling and a lower speed ADC so would need to be re-designed for our application.  This is not a problem but I'm more concerned with what causes the LTC6409 oscillation.  Keep in mind of the four identical analog driver circuits into the AD9694 ADC one of them (channel D) has the LTC6409 driver right next to the ADC and has no problems with oscillation.  We initially thought transmission line reactive loading due to imperfect matching was the culprit but increasing R53/R54 isolation resistors from 10 ohm to 200 ohm had no effect. 

Regarding dc coupling we have followed the recommended procedure to export VCM onto the LTC6409 common mode input.