OPA335: OPA335 Transimpedance Amplifier Circuit resulting in op-amp being damaged

Hello Akshay

One common culprit is you might be sitting there with each channel oscillating at much higher F than you are looking - dissipating a lot of power, raising the Tj and eventual failure. 

Two missing pieces - your detector capacitance and as you go off the ADS1248, do you by chance put some big caps on that ADC at inputs - if so what value, 

Here I estimated 20pF source C and 10nF load C, only 8deg phase margin, You feedback cap is huge, so I put in a self resonance of 4nH as a guess. This is a loop gain phase margin simulation. 

Hi Michael,

There are no caps on the ADC inputs,it is being fed directly.

The op-amp output to the + differential input of the ADS1248 and the reference fed into the - differential input of the ADS1248.

Kind Regards,

Hi Michael,

Just to confirm. What you are alluding to is that any further capacitive loading could be causing the phase margin to decrease to the point where the criteria for oscillation is met?

Kind Regards,

Well you ask why are parts breaking after a few days, very unusual unless they are operating at very high Tj

Have you checked to see if they are running very hot? 

In the past, I have seen many failure mechanisms traceable to unperceived oscillations. 

If the parts are not running hot, then there is likely something else going on perhaps related to overdrive

What does your source impedance look like. 

Hi Akshay,

the 4µ7 cap stores enough energy to destroy the OPA335 when abruptly powering down the circuit. As a remedy you could decrease this capacitance or insert a current limiting resistor which limits the current spike flowing through the OPAmp.

Can you post the equivalent circuit model of your sensor so that we can carry out a phase stability analysis?

Kai

Hi Marek,

We have added a 5v1 protection diode across the 5V LDO.I have attached the datasheet of the sensor in question herewith.I see in your simulation you are using a 1k resistor and not 3k3.There is unfortunately no equivalent circuit model in the datasheet. I am not using a front end filter. Connections are made directly to the ADC inputsBA-R-Sensor-Datasheet (1).pdf

Kind Regards,

Hi Akshay,

5.1V protection diode on supply pin may be too close to 5V LDO output and result in a large current flow.  Also, do you use a Zener or TVS on supply pin ?  A Zener may be too slow to clamp the LDO output overshoot below 7V absolute maximum required by OPA335. 

With no LPF filter at the output, the circuit is pretty stable with 80 degrees phase margin - see below. 

But 4.7uF in the feedback is very large and forms a low-pass allowing only very slow signal to be measured below 10Hz - do you really need it this low?  Also, as Kai mentioned, such a large cap stores enough energy that could damage the part during rapid shutdown.  Thus, you should either greatly decrease it (e.g. 4.7nF) or add a resistor, R2, to limit the current.  This hardly affects the stability of the circuit but will limit the current spike to less than 50mA thus preventing a potential damage – see below.

Hi Marek,

I have changed the feedback capacitor to 22nF.

Also,I meant to ask,is there anything wrong with feeding the same 1.25v reference used at the positive input of the op-amp to the negative differential input of the ADC?

Kind Regards,

Adding 22nF feedback should be fine and eliminates the need for R1.  You do not show how you derive Vref but if it is a buffered reference voltage or series reference voltage, it is actually advisable to use it to drive both nodes, non-inverting input of the op amp and negative differential input of ADC, for higher precision.

Attached please find Tina_TI schematic you may use for your further analysis.

OPA335 Akshay TIA AC.TSC