OPA564: Slowly dying: I-Flag remains high, output stays close to GND, dies with excessive current consumption

Hello Andreas,

Sorry to hear you are having problems with your OPA564 application. I need a bit more information before I can make any assessment as to what may be causing the failure.

1. The normal maximum OPA564 operating voltage is +24 V, and you are using a +24 V switched supply. Have you monitored the +24 V supply with a DSO to be sure the voltage isn't exceeding the Absolute Maximum rating of +26 V? Be sure to exercise the circuit through all its normal operating cycles.
2. Monitor the output of the OPA564 during operation. Look for any evidence of oscillation. Do you observe any?
3. Please provide the connections for the various V+ and V- pins. Do the V+ pins have power supply bypass capacitors?
4. You have provided the R and L values for the transformer primary. What are the other chracteristics of the transfomer; truns ratio, secondary impedance, etc.?
5. What load is applied to output terminals MP1-1 and MP2-1?
6. D49, BAV99W, is only rated for a continuous forward current of 300 mA. When the OPA564 fails do you know if that diode pair has been damaged too? Does any other component fail?
7. Is the OPA564AIDWP underside PowerPAD soldered to an appropriately sized PC board pad that is connected to V- (ground)?

Regards, Thomas

Precision Amplifiers Applications Engineering

Hi Thomas,

 

thanks for your reply, here are the missing details:

 

1)      The 24 V supply is safe, never more than 24.4 V, rise time on power on is ~ 200 ms, no spikes.

2)      No oscillation other than intended behavior.

3)      There are C7003/7004/7005 beneath the pins (see schematic, they are offset from the main symbol),
and (not shown on schematic) and a 220u al cap (C42) close by. See attached layout snapshot:

4)      Transformer turns ratio is 1:10, secondary @ 900 mH. The secondary is tuned into resonance for the
operating frequency using various Ls that get jumpered into the circuit for given configuration.
This part of the system drives a piezo actuator.

5)      See 4)

6)      The BAV99 is always ok as are all other components. We never exceed a Vpp of more than 15 V and
the resonant circuit does not get interrupted while in operation.

7)      The underside is soldered to the pcb and connected to all available GND layers, see drills plan. A heatsink is
soldered to the pcb (AAVID 573400) on the sides of the chip:

 

Hope that helps…

 

Cheers,

 

Andreas

Hello Andreas,

Thank you for the very complete information regarding your OPA564 application. I am not finding anything about the way you are applying the device that would explain why they are dying over time. I have a few more questions that could help me better understand the scope of the issue.

1. Is this failure mode unique to the OPA564, or was it ever observed when the OPA561 was being used?
2. Is this failure happening with just some of the OPA564 devices, or do all of them eventually fail?
3. What percent of the OPA564 devices are exhibiting the described failure mode?
4. Are the failures from one particular OPA564 lot/date code, or have they occured from multiple lot/date codes?
5. You mentioned "We had at first suspected the power supply sequence as possible cause for this failure and thus implemented
   the very circuit from the datasheet, made up from Zener and resistor. It seemed to take longer for the Amp to fail (weeks instead of days)." This indicates that the first group of failures occured in days, and then after the VDIG biasing change to the circuit that time moved out to weeks. Is that correct?
6. Is the V+ and VDIG power being cycled on and off when the application is in use?

Once I have the answers we'll decide where to take this next.

Regards, Thomas

Precision Amplifiers Applications Engineering

Hello Thomas,

1) This failure mode was never observed with the OPA551. We have never seen any problem at all with > 10k units produced.

2) 3) 4) We had a first production run of 20 boards without the zener diode and just software power sequencing, that had 5 out of
5 boards fail within a few days. We then patched in the zener diode solution and had 10 boards fail within two months. The date
code from the first run is 6CCNC6T. We now have a production run of 20 boards here that have 6CCNC4T date code and the
zener diode and resistor on board. These boards are undergoing stress tests now, i will come back with results later.

5) Yes, see above. Note: All of the first production run boards were powered up once without the zener diode patched in, to
perform factory firmware programming and automated tests.

6) Yes, there may be an input signal present, when power is cycled.

Cheers,

Andreas

Hello Andreas,

I think we should take a two pronged approach to finding the cause of the OPA564 failures:

1) We are going to run some indepth simulations on your application circuit to see if an unsafe voltage condition is being developed during the OPA564 during operation. To accomplish this we need more information about the transformer and load. You mention the transformer has primary inductance of 9 mH and resistance of 2 Ohms, the secondary inductance is 900 mH:

a. Can you provide the secondary resistance and also some idea of the mutual inductance, or the primary/secondary coupling coefficient (k) for the transformer?

b. You mention the load is piezoelectric transducer. Is it a purely a capacitive load, and if so what is its capacitance?

c. Is there a cable between the transformer and piezo load, and if so what are its impedance characteristics?

2) Please contact the authorized TI distributor that supplied the OPA564 product to you. Tell them you wish to have a Failure Analysis conducted for the failed OPA564 devices, from lot code 6CCNC6T. Tell them you have been in contact with the Precision Amplifiers Applications team regarding the failures and we have agreed to this course of action. The distributor then should provide you a Failure Analysis Request form, which must be filled out completely by you. It will provide instructions regarding the preparation and return of the devices.

One additional question I have about the application circuit is with regards to the +15 V supply used to bias the non-inverting input divider, and the +24 V switched supply. Do these supplies power up and down at the same time?

The OPA564 is sensitive to an excessive differential voltage applied across the inputs and can be damaged that should occur. Page 16 of the datasheet states in the Input Protection section, "Electrostatic discharge (ESD) protection followed by back-to-back diodes and input resistors (see Figure 43) are used for input protection on the OPA564. Exceeding the turn-on threshold of these diodes, as in a pulse condition, can cause current to flow through the input protection diodes because of the finite slew rate of the amplifier. If the input current is not limited, the back-to-back diodes and the input devices can be destroyed." We need to be sure there isn't a condition being presented where this could happen.

Regards, Thomas

Precision Amplifiers Applications Engineering

Dear Thomas,

1) If have just checked some samples of the transformer: Primary R is ~ 1 Ohms, secondary ~ 40 Ohms. Short circuiting secondary yields 15 µH on the primary, same on the other side 1.7 mH. Capacity between primary and secondary ~ 80 pF (subject to errors).

Capacitance of the piezo including coax cable is ~ 800 pF, inductance unknown. Piezo itself is ~ 400 pF, cable (RG179) 240 pF plus some stray capacitance. Inductance unknown, but should be few µH range.

2) I will get in touch with our PCB manufacturer re source of OP564.

3) The 15 V supply is derived from the +24 V switched using a linear regulator, so the +15 V are always below the +24 V for both power up and down situations,
so in my understanding we should be safe:

Thank you for your help,

Andreas

Hi Andreas,

Previously, you had mentioned the 9 mH primary and 900 mH secondary inductances, and now the 15 uH and 1.7 mH inductances, respectively. I am not sure which inductances to use in a model. No matter what I arrange for the transformer I am not able to get close to a 130 kHz resonance. That is the frequency you told me you are using to drive the OPA564. Do you have part number for the transformer?

Anything you can do to help with an equivalent circuit for the transfomer would be most helpful.

Regards, Thomas
Precision Amplifiers Applications Engineering

Hi Thomas,

you asked me for the stray inductance and that is what i measured, short-circuiting the other winding. Says so in the text...

primary main inductance is still 9 mH, secondary main is 900 mH, stray primary (note that what i measured is the sum
of both stray inductances, opposite one gets transformed) is 15 µH and secondary 1.7 mH.

cheers,

Andreas

Hello Andreas,

I think that I may have a reasonably close transformer model, and for now I am using it to run simulations on your OPA564 circuit. If the simulations are close to producing the actual operating conditions, they raise more questions.

When I run a sweep across frequency the Bode response indicates several frequencies where resonances are occuring. Resonances can result in dramatic phase shifts that can compromise the phase margin and lead to instability.

Please provide answers to the following questions:

1. Have you observed the waveforms at the OPA564 at output pins (15, 16) and at the piezoelectric load when the the circuit is in operation? I really need to see DSO images of the waveforms present at the OPA564 input, output, and at the load when the circuit is operating.
2. Is there a possibility to set up the OPA564 input where it can be driven with a small-signal square wave? By small signal I mean a 10 mVpk square wave having a low frequency such as 100 Hz, or 1 kHz. DSO images of the OPA564 input drive signal and the output pins (15, 16) would need to be captured.
3. Is the output transformer the designated transformer for use with the particular piezoelectric transducer, or something else? Would it be possible to get the manufacturer's part number for the transformer?

Regards, Thomas

Precision Amplifiers Applications Engineering