LT6600 oscillations with spec sheet gain


I am attempting to use some gain in the LT6600 chip to get lower noise output with a single-to-differential configuration that drives an ADC.  The spec sheet says you can use 402 ohm Rs resistors (I'm using 470) but the circuit below oscillates at about 3 MHz when measured using a spectrum analyzer.

I'm using C0G series caps and tolerances on all other components to within 1%.

Is the LT6600 only marginally stable at this gain, or, am I doing something completely wrong?

  • I forgot to mention that this was shown with shorting the input pins (to GND) and the additional AD9257 load was not attached.

  • Hi again,

    I'm still struggling with the oscillations.  Things I've tried:

     - bumped Rs up to 1k

     - removed the singled-ended input and did differential

     - Tied Vocm to Vmid

     - Used cleaner power supplies

    Nothing seems to help with the oscillations.  Any other ideas?


  • 0
    •  Analog Employees 
    on Jun 6, 2019 4:45 PM in reply to buckd

    Hi buckd,

    Are you AC coupling the LT6600-2.5 inputs as your LTspice image shows?

    I think you'd need to allow the input bias current (35uA max) to flow. So, I'd say having a DC path is required because you're AC coupling.

    You may try adding a large resistor to ground on each input and see if that resolves your oscillation?

    Also, when evaluating use low capacitance techniques and make sure there is adequate differential load on the output(s).

    Please let me know if these recommendations help you in any way?



  • Hi,

    I am AC coupling the LT6600-2.5 like in the schematic I posted.  I followed the topology shown in Figure 2 on page 8 of the spec sheet:

    I believe am using low capacitance techniques (short traces, small surface area, etc) and I have been attempting noise measurements like in Figure 5.

    I am attempting to add some gain with the AC coupling - does that topology only work with unity gain?



  • 0
    •  Analog Employees 
    on Jun 7, 2019 6:52 PM in reply to buckd

    Hello D,

    Sorry! My bad. Had not realized that the LT6600-2.5 input bias current spec in the datasheet (35uA max) can be supplied through the integrated feedback resistor (1580ohm) of the first stage and thus input AC coupling is allowed. My apologies!

    You are biasing VOCM to 0.9V and your VMID is at mid-rail (2.5V).

    There is a note under Table 1 which states VOCM > = 1.5V, as copied below:

    "NOTE: VOCM is set by the voltage at this RIN. The voltage at VOCM should

    not exceed 1V below the voltage at VMID. To achieve some of the output

    common mode ranges shown in the table, the voltage at VMID must be set

    externally to a value below mid supply."

    I recommend you increase VOCM to > 1.5V to comply with this requirement to see if your instability can be addressed? Better yet, just tie VOCM to VMID and see what happens.