MC33078: MC33078

Hello Dan,

Consider adding diodes that limit the reverse polarity on supplies. Ensure that inputs do not have low impedances paths to ground; so reverse input current will be negligible. During the time that input common mode range is not valid then output voltage will likely be incorrect. 

Therefore '1)' could be damaging and '2)' is unlikely to be any problem.  

Evaluate what is seen on device pins during this "fault" time so make sure the application is low stress. 

Hey Dan, 

Welcome to e2e!

For scenario 1, 

• How long does it take the +15 rail to power up? Therefore 2x for the -15V rail is...?
• What is the voltage at the negative supply rail (Pin4) of the op amp when -15V rail has not powered up? 
• Is there an input signal (Pin 2,3,5,6) at scenario 1? 
  • If so, what is the voltage range?

For scenario 2, 

• Is there an input signal (Pin 2,3,5,6) at scenario 1? 
  • If so, what is the voltage range?

All the best,
Carolina

Thanks very much, Ron.

I'm a little confused by the diode suggestion.  Do the diodes simply add a little impedance in the case of latch up?

Also can you please confirm in my schematic below the positioning of the diodes.  You'll notice in the image on the right two diodes shown in RED in the Vcc and -Vcc connections, both pointing downward.

Our inputs are limited by 100k resistors in this case.  Thanks for the suggestion!

All the best,

Dan Bolda

Hi Carolina,

Thanks very much for the replies!

We haven't actually built the circuit yet.  Prototype PCBs go out next week.  There are about 1,000 parts on the board.  It's about the size of a pizza box.  The designer, "Dave", used Webbench to design the various blocks in the circuit such as the power supply rails, amps, etc..  So we only have as much as Webbench will tell us.  Unfortunately, Webbench doesn't flame up when amps latch up.  Maybe next revision?

I came in late in the project after everything was designed and am reviewing the blocks and trying to envision little gotchas when the blocks get to play together in the sandbox someday.

This particular issue jumped out at me since it just seemed dangerous to me.

Dave will return from vacation next week, so I will work with Dave to see if I can get answers to your questions.  He'll have to fire up Webbench and do this thing.  He has all the initial files.

• How long does it take the +15 rail to power up? Therefore 2x for the -15V rail is...?  (DAVE)
• What is the voltage at the negative supply rail (Pin4) of the op amp when -15V rail has not powered up? (DAVE)
• Is there an input signal (Pin 2,3,5,6) at scenario 1?  YES.   
  • If so, what is the voltage range?  The input can be -10V to +10V.

For scenario 2, 

• Is there an input signal (Pin 2,3,5,6) at scenario 1?  YES 
  • If so, what is the voltage range? -10V to +10V

Have a great weekend and Best Regards to you!

Dan Bolda

Dan,

More like this on left, two diodes help protect all devices that use +/-15V busses.

If using the right side, put caps on op amp power pins. two diode need for each op amp package and diode will reduce maximum output swing and narrow the input common mode range.  So I prefer left side

There is a data sheet absolute maximum for input to input current, but note 3 states that input voltage beyond power rails is forbidden.

So the input voltage without proper supply power violates the absolute maximum table. 

Hi Carolina, 

Dave, the designer, is looking at the information provided by Webbench in order to answer your questions.  Dave is a little doubtful that the model for the TI chip that produces -15V rail will give sufficient information.  In the meantime, Dave also asked me to share the system schematic with you with hopes that it may fill in some of the gaps (see below).  Maybe just to reiterate:

1)  Ignoring the inputs altogether (i.e. maybe they are floating), is it possible for the MC33078 to latch up or get damaged if the +15V rail is "significantly" faster than the -15V rail?

2)  Assuming the inputs could be +/- 10V, is damage possible if the +15V rail is "significantly" faster than the -15V rail?

Thanks very much!

Dan Bolda

Hey Dan, 

My questions stem from the abs max table: 

& also what the op amp "sees", picture is a little blurry but good enough to find the supply voltages, general setup:

The board has dual supplies (±15V) in reference to gnd "VSSA_LS" and has an input applied to the op amp of ±10V also in reference to gnd "VSSA_LS." 
However, the op amp doesn't know what is GND or midsupply, it can only work off of the voltages provided to the supply rails at that given time. 

Dan Bolda said:

1)  Ignoring the inputs altogether (i.e. maybe they are floating), is it possible for the MC33078 to latch up or get damaged if the +15V rail is "significantly" faster than the -15V rail?

Reference Jerry's response below. 

Dan Bolda said:

2)  Assuming the inputs could be +/- 10V, is damage possible if the +15V rail is "significantly" faster than the -15V rail?

15V is applied but -15V is not: Under the assumption that the op amp's supply rails are:

• 15V & GND, the ±10V input would violate the abs max table since -10V is much greater than 0V. 
• 15V & V- is between 0V and -15V, the ±10V input would violate the abs max table up until -10V. 

Stresses beyond those listed under Absolute Maximum Ratings may cause permanent damage to the device.

All the best,
Carolina

Hey Dan,

Yes, the device could latch up or be damaged if the +15V is applied significantly faster than the -15V rail.

If the inputs are above or below the rails, this can damage the device. You can protect them externally with diodes from the rail to the inputs. The internal ESD diodes are not designed to carry large current values, larger external diodes can accomplish the protection without being damaged. Make sure these are placed close to the device you are protecting.

Best,
Jerry