LM4041C12IDCKR VHYST parameter

Hello,

During the qualification, 77 devices were temperature cycled from -65C to 150C for 1000 cycles. All device passed post testing.

Please provide more information about the customer's thermal cycling, device cathode current, and the result values.
I will repeat this on some in house samples.

Regards,
Ron Michallick

Dear Ron:

      Thanks for your reply~ The customer still have some questions:

     After confirm, the max and min Vo set point have 0.5% difference, for component LM40401 C grade, the tolerance is 0.5%, so thermal cycle result seems for these samples could inside spec. But I want to know if we use A grade, then what the VHYST spec(pls refer to the definition as below) will be? could A grade component 1.225V voltage still inside 0.1% spec after thermal cycle?

The test process:
  the temperature drops to -40oC and then raise to 25oC, the voltage is defined as: Vz,25oC(after cycling to -40oC).
  the temperature raise to 125oC and then drops to 25oC, the voltage is defined as: Vz,25oC(after cycling to 125oC).

Though they are bothed tested at 25oC, the two values are different. The first value minus the second value is defined as VHYST.

The temperature grade {A,B,C} has no effect on VHYST.  A observed 0.50% VHST is very high.
I will check a couple LM4041C12 samples in this order: 25C, measure, -40C, 25C measure, 125C, 25C, measure, wait hours, measure. This is to determine if the customer results are realistic. How long was the 25C soak time? If the time is too short ,the junction temperature may not be 25C
If conversion to the TL4051C12 (VHST typical 0.04%) is a possiblity, I will also test the TL4051C12.

Regards,
Ron Michallick

Dear Ron:

       Thanks a lot for your support!

       Here is the customer's test condition:

400 cycle, -40C~125C, the module was powered off when we do thermal cycle test, this mean no cathode current.

Our thermal cycle test set up like below:

 
The voltage will be measured two times, first time is before thermal cycle, 2nd time is after thermal cycle.
We measure the voltage after we take unite out of chamber, so the temperature will be 25C which is just the ambient temperatrue.

The first device I tested had a VHYST of +0.09mV/V; the second had +0.25mV/V.
I cycled the temperature backwards on the second unit and the VHYST dropped to -0.06mV/V.
A third unit was -0.15mV/V. The values I observed are much less than the customer's report.
My units were temperature cycled and tested in a socket. Units soldered to a board may have a different stress profile.

The TL4051C12IDCKR should be a drop in replacement and is more resistant to package stress affecting shunt voltage.

Regards,
Ron Michallick

Dear Ron:

     Pls refer to the questions fron the customer:

I have two questions:
1.      did TI do this test only one temperature cycle or many times re-cycle, as you know Delta will do 500 cycles.
2.      what does it mean of “The TL4051C12IDCKR should be a drop in replacement and is more resistant to package stress affecting shunt voltage” 

    Thanks a lot!

1) The qualification was 1000 cycles, My lab data was single cycles.

All TI testing involved sockets. We (TI) have not observed the level of shift as Delta has observed. So I focused on potential variables that may be unique to the Delta application. I suspect that the physical stress on the package (from the circuit board) has affected the voltage. This could be verified by removing the device form the board and re-testing unit either off the board or re-soldering it back on the board.

2) The TL4051 has the same pin-out and functionality as the LM4041 and the specifications are very similar. The TL4051 highest grade has better final test yield than the LM4041 of the same grade, therefore is is logical to assume that the package stress shift in voltage of the TL4051 is less than LM4041.  

Regards,
Ron Michallick