Post Go back to editing

How fast can ADG4612 can turn on and off

Hi,

  I would like to know how fast the ADG4612 can turn on and off? What happend if 1MHz clock is feed into the INx pin?

Thank you,

Khai.

  • I am moving this question to the Switches and Multiplexer community for an answer.

    Thank you

  • Hi Khai,

    The speed at the which the switch can toggle on and off is primarily governed by the Ton/Toff time.  Ton/Toff time will depend on supply voltage, temperature and the load condition.  Note, for this part Ton time over temperature can go up to 226nsec max with a 300ohm 35pF load, see page 7 of the datasheet.  Also you will need to consider the rise time of your clock edge into the logic control pin, this will be affected by the pcb capacitance and the input capacitance Cin of the logic pin.  As you can see there are a lot of system parameters that can affect how fast the switch will turn on and off in practice.  I would also draw your attention to the fault response time and fault recovery time, if in a fault condition you will have to take these times into account as well.

    Regards,

    Eric.


  • Hi Eric,

      I do noted the numbers in the data sheet, however, if I look at the timing chart the number more like the propagation delay of Ton and Toff than the time of the switch actually transition from off to on or on to off state. I'm more interesting in the transition time. This would tell me the maximum speed I can switch.

    Thanks,

    Khai.

  • Hi Khai,

    In reality the propagation delay for an on switch is very small, the switch is typically only 5ohms in series.  I don't have an exact number for propagation delay but you can take Ton to be dominated by the transition time, i.e. how long it takes the gates to charge up and turn the switch on.  Looking at the Ton times, my feeling is that if you have a small capacitive load then 1MHz switching time should be possible as long as you don't have any significant parasitic capacitances on your board.

    Regards,

    Eric.

  • Hi Khai,

    Can I confirm what you are trying to do here.  Sounds like you want to input some kind of analog signal into the switch and then toggle the switch at 1MHz and chop it up?  The only measurement setup we have was used for the Ton/Toff results and is on the data sheet, timed as in figure 25.  I will see if I can find the original evaluation data for you and if it has more detail on the rise time.  If you need really fast switching speeds I would say have a look at the ADG702 or ADG802 (high power handling) families of parts, they have low Ton times. 

    btw, there is an inexpensive evaluation board available for the ADG4612.  EVAL-ADG4612EBZ

    Regards,

    Eric. 

  • Hi Eric,

      Thank you the response. I try to figure out if I can modulate my analog signal with 1MHz square wave to conducting current so if the transition time of the switch is less than 20 or 10ns I think my design can work with it. But I have no way to grantee if it event works with single part or if it work with any parts of ADG4612 by using information provided from the data sheet. Would you mind to provide me some measured data which give me some level of confident of using this analog switch in my design. I'm not worry about time of charging the 35pF load capacitor, but time from off to complete turn on for conducting full spec current (80mA) and time for on to complete turn off while conducting 80mA. The output voltage level isn't a concern in my design.

    Thanks,

    Khai,

  • Hi Khai,

    I can't find full timing plot data unfortunately.  The fastest route here for you to prove out the performance is for you to order an evaluation board and test the performance parameters of the part in own your specific application use case. In addition, I would recommend using a faster Ton switch like the ADG7xx or ADG8xx families if you need really fast transition times. 

    Regards,

    Eric.

  • Hi Eric,

      Thank you. I will try the eval board. Other switch are fast, but my application need +/-5V, and 220mA (So think if ADG4612 fast enough, I can parallel four of them). Otherwise, my last resource is using discrete FETs.

    Thank,

    Khai.