Quantcast
Channel: EngineerZone: Message List
Viewing all articles
Browse latest Browse all 22625

Re: AD8657ARMZ latch-up phenomena

$
0
0

AT,

 

  The designer said a customer noticed an anomaly a few years ago and sent me an email he had sent to the product line apps engineer.  Excerpts:

 

"...This is a capacitive feed through problem.  .....  When the power supply
ramps up quickly, the bias currents are off until the current sources start ( ~
140 us for the AD8657 ).  The gate drive voltages on the output transistors
MN0 and MP0 are determined by the compensation and parasitic
capacitances.  Since the compensation caps are much larger than the Cgs
capacitance ( 16 pF vs ~500 fF ) they will have approximately 0 V across them
while the supply voltage is split between the 500 fF caps.  Once the
supply is greater than the thresholds of the two output transistors they will
start to flow current.  If there is no load attached then this current
just shoots through from Vdd to gnd.  The output voltage will ramp up or
down depending on which transistor pulls harder.  Most of this will be
determined by which Cgs capacitance is lower.  The smallest cap will have
the largest voltage across it.

          
....  Nothing happens until the bias comes up around 130 us when the comp caps start to
charge and the amplifier begins controlling its output stage.

 

This effect is present in all rail-to-rail output Miller compensated amplifiers. 

It is probably more noticeable in the AD8657 for two reasons:

 

  1. It is so low power that it takes a long time to charge the comp caps.  The
    voltage spike will only go away once the bias currents charge the comp caps to
    their proper level.  This time will be related to the slew rate of the
    amplifier.
  2. The PMOS and NMOS output transistors are the same size.  It is typical to make
    the PMOS 2-3x larger than the NMOS.  This will make the PMOS Cgs larger
    than the NMOS Cgs so the output will tend to pull down.  The single supply
    test circuit is incapable of showing pull down behavior since the output starts
    at the lowest potential, but I would expect to see it on a split supply test
    setup. 
    "

 

So if you can't change op amps, you will have to put in some sort of delay to ramp the supply.

 

Harry



Viewing all articles
Browse latest Browse all 22625

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>