Power Supply Noise

Discussion in 'General Electronics Chat' started by Management, Sep 30, 2009.

  1. Management

    Thread Starter Active Member

    Sep 18, 2007
    306
    0
    Hi Everyone,

    I.
    If I have a surface mount device (VCSEL Driver) that has a supply voltage of 2.5V and the datasheet states that the supply should not have more than +/- 400 mV peak to peak (200 mV peak) how can I or what can I do to make sure I meet this requirement?

    Do I place a cap to GND? If so what value?

    The supply current is a maximum 0.6 mA. Yes milliamperes.

    II.
    If I have a device (TIA) that uses dual supplies, 2.5V and 1.2V, what is the best way to get 1.2V from a 2.5V source? Do i just use a voltage divider (resistors)? or does this not help me meet a noise spec of 200 mV peak to peak (100 mV peak).

    Thank you.
     
    Last edited: Sep 30, 2009
  2. steinar96

    Active Member

    Apr 18, 2009
    239
    4
    I)
    The standard is usually a 0.1uF ceramic capacitor as close to the power pins as possible (between the power pin and ground).

    II) I do not recommend voltage dividers because the "voltage division" depends on current drawn by the circuitry working at the voltage of the lower resistor. If it's varying the reference voltage you are trying to set can vary alot (which is usually bad). 1.2V is so low that normal series regulator chips might be hard to find. But IF the current requirement is very low. You might get away with a simple zener regulator but the effeciency isnt very high.
     
  3. Management

    Thread Starter Active Member

    Sep 18, 2007
    306
    0
    Ok. Well I am going to do a search for some regulators then. I will use a 5V supply into two regulators to get the supply voltages.

    Even with the regulators do I still need the caps at the power supply? If so, why?

    Thanks for the help.
     
    Last edited: Sep 30, 2009
  4. alphacat

    Active Member

    Jun 6, 2009
    186
    0
    Hey,
    I wanted to ask you something please about that.
    To the best of my knowledge, Ceramic capacitors are used to filter AC signals.

    In the OP's case, A DC level is desired to be "straighten" (to decrease its noise), so why not using an Electrolyte capacitor?
     
  5. Management

    Thread Starter Active Member

    Sep 18, 2007
    306
    0
    Is this a good capacitor to use?

    Link: http://www.avx.com/docs/catalogs/tajlp.pdf
     
  6. peranders

    Well-Known Member

    May 21, 2007
    87
    0
    I It is pretty easy to get less 400 mV peak noise.

    Decoupling caps can be 10-100nF not very critical.

    II
    You can use a 431 shunt regulator with 1.2 V ref voltage. Texas Instruments has those.

    Have you read the datasheet for the driver? I'll guess most of the important is there.
     
  7. steinar96

    Active Member

    Apr 18, 2009
    239
    4
    Usually electrolytic capacitors are used when rectifying AC voltages because they can achieve high pretty capacitances.

    Ceramic capactors are usually used for defending IC circuitry from high frequency noise. The ceramics are better at the noise frequencies (fairly non inductive at high frequencies).
    If you expect big swings in voltage however you might want to also add a electrolytic capacitor to absorb bigger and lower frequency swings. But if you have to do that to keep your IC circuitry running correctly you have a big proplem with your regulation which you are better of fixing instead of compensating for.
    But let's say if he had decided to use a big electrolytic capacitor instead. Then that would mean that most high frequency noise is able to propegate into his circuitry, which might in turn not function correctly.
    The different kinds of capacitors have different applications.
     
  8. Management

    Thread Starter Active Member

    Sep 18, 2007
    306
    0
    This is a great discussion. Learning a lot. So I guess the cap above is not a cap I would need because I plan to use a regulator to provide a steady reference.

    So I guess a ceramic capacitor would be best to stay below the +/- 200 mVpp noise requirement, correct?
     
  9. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    Your cap choice will depend largely on the implied noise from surrounding devices. If your combining digital and/or power switching you'll want to best serve that. It will also depend on how deep your drawing on your supply.

    By rule, digital devices require, as mentioned, .1uf close to the supply pins minimum, with a few picos close to the source of noise.

    With the ease of using SMT devices in small packages, the more the merrier.
     
  10. ifixit

    Distinguished Member

    Nov 20, 2008
    638
    108
    What frequency are you running at? The 0.6mA seems way too low for a laser driver.

    The VCSEL driver manufacturer should have some recomendations for decoupling in the spec.

    Read the voltage regulator spec and go with their recomendation on decoupling the regulators input and output.

    Regards,
    Ifixit
     
  11. Management

    Thread Starter Active Member

    Sep 18, 2007
    306
    0
    The device is a VCSEL Driver with really low differential signaling coming in and driving a VCSEL. The part only dissapates 5mW. The TIA only dissapates 9mW and uses a dual power supplies.

    I'll check the regulator when I pick one and then look at the VCSEL specs. I'll update with what I find.
     
Loading...