adg3300

I'm trying to use the ADG3300 to simply translate between a 3.3V and a 5V system .    Originally I had planned to

use the auto-sense capability for bidirectional signals,but have decided the part is way too unstable for that to ever work.

Is it possible to force it to not autosense, or is there a pin compatible alternative that has a direction control?

    thanks and regards,  qwik

  • Hi qwik,

    The autosense function on the ADG3300 cannot be disabled as such.

    However, if the circuit that drives the I/O pin used as input has a low output impedance while the I/O pin used as output always sees a high impedance, the autosensing will have no effect as translation will only occur from the input I/O to the output I/O pin.

    As regards an alternative part, I assume you want to do step up translation only (3.3V to 5V). Is my asumption correct?

    What is the maximum speed of the signals that need translation?

    Regards,

    FEB

  • Hello,  thx for the response;  Yes I only want 3.3   to 5.5.   I do drive the input with a low impedance, and the output sees high impedance-  back biased diode

    Protection network -  Diode to gnd, one to  +5.     The part oscillates wildly- particularly when I drive more than one input.    I need it to operate up to a speed of

    About 5 mhz digital, which means there are harmonics up to 15 mhz.

    Whats really annoying about the part is if I drive one input at 1 mhz digital, and another at 100 khz, its as if theres an interaction between the two-  it

    Oscillates. I always drive the inputs on the a side, the 3.3V side,  and the output is the y side,the 5V side.      Thx   Carl

     

    From: FEB analog@sgaur.hosted.jivesoftware.com

    Sent: Monday, January 09, 2012 7:52 AM

    To: Robbins, Carl

    Subject: New message: "adg3300"

     

    Analog Devices EngineerZone <http://ez.analog.com/index.jspa>

    adg3300

    reply from FEB <http://ez.analog.com/people/FEB>  in Interface and Isolation - View the full discussion <http://ez.analog.com/message/39834#39834

  • Hi Carl,

    To avoid oscillations, you need to have good decoupling on the supply pins of the device.

    Also, the length of the tracks to /from the device should be kept as short as possible. For 3.3 to 5V translation only, I would recommend placing the ADG3300 close to the 5V logic circuitry and terminate the track that feeds the 3.3V input of the device with an appropriate resistor. This should remove the oscillations.

    Regards,

    FEB

  • I had to put a 250 ohm resistor on the 5 V output- in series ,not a load.  Otherwise I couldn’t even drive a scope 20 pf input.

      Also,the rise time of the output is about 1 usec(not affected by the 250 ohm series resistor).

      All in all, a pretty flaky part        Carl

     

    From: FEB analog@sgaur.hosted.jivesoftware.com

    Sent: Wednesday, January 11, 2012 10:18 AM

    To: Robbins, Carl

    Subject: New message: "adg3300"

     

    Analog Devices EngineerZone <http://ez.analog.com/index.jspa>

    adg3300

    reply from FEB <http://ez.analog.com/people/FEB>  in Interface and Isolation - View the full discussion <http://ez.analog.com/message/40052#40052

  • Hi Carl,

    As the ON resistance of the one-shot generators have a couple of ohms, the rise/fall times measured across a 20pF load should be within the datasheet specifications. the 1us value is very high for such small capacitive load.

    If adding the 250 ohm resistor in series did not impact the rise/fall time, it means that you must have a large series inductance in the load that dominates. Assuming you are using a solid GND place, I would recommend checking the ground lead inductance of your probe.

    Regards,

    FEB