Switching Output of Operational Amp

I am using AD8055 as an inverting amplifier in a set-up similar to Fig. 35(I used 1k instead of 402 Ohm).

What I need to do is switch on/off the output of the amplifier.

First, I put a switch ADG451 to the output of the op-amp, then it made a huge spike (roughly 1V) when the switch turns on. (This seems a well-known phenomena. Could you explain why it happens?)

Next, I put the switch parallel to the feedback resistor, so that when the switch is on,

the amplifier would work as a gain-0 amplifier.

However, the op-amp oscillates around 100 MHz when the switch is on.

It was not clear how this can happen.

I also tried AD844 instead of AD8055, in which case AD844 oscillates at 25MHz.

If I put OP27 instead, it works well, although the circuit is not fast enough to my purpose. 

The circuit is on a PCB board, and AD8055 and ADG451 are DIP packages,

which is one of my concern.

However, just putting ADG451 into the feedback loop seems fine,

because if I put another resistor(1k) in series with the switch (G=-1 with the switch on),

the circuit works well without any oscillation.

Is there any trick I can use to solve this problem?

  • ts1984actc,

    Above a few MegaHertz, it is very difficult to breadboard without a clean pc board layout.  If you are using DIPs with

    one of the white breadboarding strips with holes, you will have oscillations.

    Read pages 11 and 12 of:

    http://www.analog.com/static/imported-files/tutorials/MT-099.pdf

    Surface mount packages on a well laid out pc board with bypass caps VERY close to the package will result in

    a repeatible design.

    "similar to fig 35, except..."  can make a big difference.  Can you post your exact schematic in .pdf format?

    You may get better results if you switch the input to your circuit, rather than the output of the op amp.

    Harry

  • Dear Harry,

    Sorry for my unclear description.

    I attached the schematic and the part of the PCB layout.

    Bypass capacitances are not shown in the schematic.

    I used 100nF and 10uF for AD8055 and also put 100nF's to the supply for ADG451.

    I did my best in my PC board design(the layout and the picture);

    The bypass caps are closest to the op-amp for non-surface mount components.

    The only reason I used DIP packages was just to make soldering easier.

    My circuit needs to be as fast as only 5-10MHz,

    and the other part including other AD8055's was working very well,

    so I didn't think I need to use surface mount packages.

    However, I am not fully sure if this problem can be solved with surface mount components.

    What I suspect now is that there is some stray inductance in the feedback path through the switch.

    But I am not sure if it can be big enough to cause oscillation.

    Naively thinking, it should not oscillate because the gain is almost 0.

    Finally, it seems difficult to put the switch to the input in my case.

    There is a chain of op-amps before this part, and

    the input of this AD8055 is connected to the output of another op-amp;

    probably I would get a spike if I do it...

    Thank you for your help.

    TS

  • Hi ts84'

    Sorry for so late answer but I just stumble upon your post and first thing that i noticed is that cause of that issue not so much board, DIP package or any PCB layout (as people generally tend to assume) but resonant loop value.

    To be precise R3, R5 and C1. I do lot of designs with 8055 lately and push it to +30MHz in Trans-impedance amp. mode and when I saw 1uF for C1 fist thing that come to my mind was "this is gonna make a mass, and this R3 is to low for 10MHz amp bandwidth".  So I do some analyze and find out that R5 and C1 make "just fine" self RC oscillation circuit in parallel on feedback poop. I even simulate circuit and find out that it generate harmonics on 50MHz, 100MHz... even with 1MHz input. 

    I suggests you remove R3 and trim value of R5 and C1 to better match targeted bandwidth of your circuit.

    Try to change value of R5 to 12-14kOmh and  C1 to 0.6 or 1 to 3pF, then you can even rise R11 over 1k you will get spiked but stronger signal that can trigger logic.

    Your (old) design even in OFF mode do not shutdown amp, look at that B channels -1.2mV in next illustration, it will get back over closed loop to input and can even trigger next circle of avalanche amplification.

    Best regards,

  • 0
    •  Analog Employees 
    on Aug 2, 2018 4:30 PM
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin