INA326EA Output Railed - Sometimes

Here's a minor update.  I have a batch of 15 boards.  8 of the boards power up and test OK with about 5 power cycles each.  The other 7 boards have mixed results and the power-up failures DO NOT only occur on the circuit with a different layout. 

Hello NHSoxFan,

I have a few questions:

a) What is the desired gain of the INA326? I assume it would be 0.54. However, the presence of R7 changes the gain as discussed in Figure 6. In your case, G=2*(113k||1M)/432k=0.47.
b) Have you tried disconnecting the 22uF load? Why is there such a large capacitance on the output?
c) Have you minimized stray capacitance at pins 1 and 8 of the device?  Perhaps you could post a screen capture of all of your PCB layouts.
d) Since you have the inverting input grounded, have you considered using a traditional op amp for this application?
e) What is the source of your input signal? What is the source impedance, frequency, and amplitude?

Can you please provide links to the other forum posts that you reference?

Hello Pete,

a) The gain would be 0.54 if I were directly scaling 5V to 2.7V, however don't forget about the 0.3V up shift (0V input = 0.3V output).  Therefore gain = 0.47.

b) The load cap is actually 2.2uF (you might not see the decimal point in there).  Heavy filtering on the output into an ADC.

c) I can't get the 0402 resistor between pins 1 and 8 any closer.  See attached.

d) I believe that due to the level shift, a traditional op amp would not work unless I had a negative supply voltage.  I don't have this negative supply voltage available.

e) The input signal comes from a hall effect sensor.  Very low frequency.  Amplitude can range from 0 - 5V.

Here are the other forum posts of people experiencing the same problem:

http://e2e.ti.com/support/amplifiers/precision_amplifiers/f/14/t/134999.aspx

http://e2e.ti.com/support/amplifiers/precision_amplifiers/f/14/t/98099.aspx

Hello NHSoxFan,

The second post states that the issue is related to the ramp time of the supply voltage.  Have you looked at the ramp rate of the supply voltage to see if there is any correlation with the output railing?  One idea may be to increase the size of the bypass cap.  Can you also show us the layout of the bypass cap? 

Hi Pete,

Here's in image of the entire circuit.  The bypass cap is C50.

I'll measure the rise time of the 5V power supply, and then I'll see what I can do to increase the rise time. 

Tom

Hi Pete,

I've taken some scope images at powerup.  Here's the legend:

Ch 2, green = pin 6 of INA326 (output)

Ch 3, purple = pin 1 of INA326 (gain resistor)

Ch 4, red = 5V supply to INA326

The 5V supply is a boost regulator made from 3.3V, thus the stepped startup.

Here is a powerup with the failiure.  Notice the oscillation in Ch 3.

I tried increasing the capacitance on the 5V supply with no change in the oscillation behavior.

Next, I tried reducing the capacitance on the 5V supply.  There was a 2.2uF and 0.1uF on the output of the regulator, so I removed the 2.2uF just to see what would happen.  The problem then went away.  The scope image shows that the 5V supply's first step is now around 2V instead of the previous ~1.75V.  See below.

I then measured one of the boards that hasn't been producing this issue, and its 5V supply even with the standard 2.2uF load always rises to around 2V during that first step of powerup.  It is looking like the INA326 doesn't like a slow rising power supply on startup.

Any comment?

Hello Tom,

I need to consult with a colleague on this. I will get back with you early/mid next week.

Hello Tom,

Can you please try changing Ro and Co to 100ohms and 1uF, respectively?