How to calculate the MTBF of Devices for designing..??

[deepakchikane]

Hello scholars,

i would like to know the Information on MTBF of the devices.

also what are the steps taken for the Product designing stages..?

Kindly share your ideas..!!

Thanks with Warm Regards
Deepak Chikane

 

[andre_luis]

The MTBF of a product is obtained by the direct sum each one the individual MTBF of either assembled modules, as well whole single components. This can be obtained either at manufacturers specifications or at standardization handbooks such as the MIL-HDBK-217 ( reliability prediction of electronic equipment ). There are some environmental factors that impacts on the lifespan, but the one which is the most expressive is the temperature of operation/storage.

A systematic method that is quite advisable during drawing of electronic boards, is creating additional component fields at the schematic libraries. Once you fill the standard values for BOM ( part number, footprint, manufacturer, etc.. ), you also insert all the correspondent coefficients ( pi’s, lambdas, etc… ) so that by a further analysis of the exported text/spreadsheet generated file you will be able to perform calculation of each part.

 

[dick_freebird]

A tricky thing to do well, because the MTBF numbers
you will receive are from some declared worst case
set of conditions which probably don't bear much
relation to your board level application of the parts
(for example, a MOSFET's FIT number might be at
rated Vds(off) and rated Ids(on) at some switching
frequency - are you using the same voltage and
current in the same rough frequency range, or if
not did anyone care to provide you the voltage /
current lifetime fitting coefficients? Usually the
power guys derate heavily taking themselves well
away from the manufacturers' rating basis.

You can really get into some deep quicksand if
you apply the generic reliability numbers to a large
system that applies components more benignly than
reliability lift tests do (and you'd be vigorously
wire-brushed for designing anything up against
those conditions / limits, anyway).

 

[hobbyckts]

The stress analysis, de-rating analysis, reliablity analysis altogether will give you answer on MTBF. Generally for IC's the manufacturer will give all the worst case condition at which it can operate in their respective website for any IC's as a reliability report. There is also tool available to calculate the MTBF by entering the above said values will provide you the final MTBF.

Somebody correct me if I'm wrong....

 

[deepakchikane]

The stress analysis, de-rating analysis, reliablity analysis altogether will give you answer on MTBF. Generally for IC's the manufacturer will give all the worst case condition at which it can operate in their respective website for any IC's as a reliability report. There is also tool available to calculate the MTBF by entering the above said values will provide you the final MTBF.

Somebody correct me if I'm wrong....

How to calculate stress Analysis..?

 

[hobbyckts]

For example when you take up the resistor.

You connected in a place where you will be knowing the value of voltage. now calulate max.current & max. current. Now decide the power rating of the resistor to be used in your circuit. The power rating of the resistor has to be approximately 1.5 times higher than the what you have got in your calculation. If the criteria is not matching means the stress analysis is fail at the particular stage where you have used the resistor.

Hope I explained clearly

 

[deepakchikane]

Not cleared yet.

if i am taking a 230 volts is the place where i am going to place a resistor.
v=iR
R=230 OHM (i took random for calculation)
so i will be 1 amp

now, Power=i*i*r=1*1*230=230watt

what is next..??

What is best stress analysis in case of semiconductors ..?

 

[andre_luis]

The stress analysis is somewhat implicit at the standards hadbooks for electronic components reliability prediction, if considering it is working within the absolute maximum ratings conditions. It essencially take in account how much a device is operating beyond its rated specifications, such as external factors ( temperature, humidity ) as well intrinsic factors ( nominal value, number of gates, etc… ).

I don’t think a good idea to perform a precise calculation for each device, particularly if involves a big amount. Keep in mind that the focus of the MTBF analysis is to make the thinks simplest, not complicated. Moreover, the result carries a certain level of uncertainty, due it is in the field of a statistical science, not deterministic.

Take a look here: http://www.reliabilityeducation.com/intro_mil217.html

 

[jay1691]

Hello All,

As i have used 4Gb DDR3 SDRAM (H5TQ4G63AFR-PBC-448A) in one of my product and while calculating MTBF of that board based on MIL-HDBK-217F-Notice2 i found some issue.

In MIL-HDBK-217F-Notice2 document " Die Complexity Failure Rate - C1 " factor is given only for up to 1M but there is no information for 4GB SDRAM... so i have confused on below two points,

1) is this document is applicable to SDRAM ???
2) how to calculate MTBF for this SDRAM ??

Please help me to solved this out.

 

[hobbyckts]

can you provide the full partnumber for both?

 

[jay1691]

Dear hobbyckts,

I have already mentioned part number. It is H5TQ4G63AFR-PBC-448A from SK hynix manufacturer.

 

[andre_luis]

It is plausible that a document aimed to provide an estimate does not cover all the possibilities of the real world, so in this case, you can perform an extrapolation of the data and calculate the value of the parameters, extending the model to the size of the product that you have.

 

[hobbyckts]

Check out the document. Micron website will be useful for technical documents of memory devices.



https://www.google.co.in/url?sa=t&r...PO7Y2Yp1ve8-HV2n4aqGBw&bvm=bv.103388427,d.dGo

 

[dick_freebird]

The newer the part the less likely it is that any reliability
work to support these rate predictions, in environments
that represent the military's interests, has been done.
You're looking at a Method that probably hasn't been
updated in a decade, judging by the upper bit count.
1M SRAMs were '90s vintage, 1M DRAMs late '80s? And
there has been much movement in DRAM art between
then and now, with everything much closer to the bone
than "back in the day", different failure mechanisms and
sensitivities and so on because geometry and process
have moved so markedly.

What you'd like, is deeper reliability info from Hynix, not
a stack of stale paper from the Pentagon. But if your
customer (or "customer") insists you follow the Method,
then you've got a story to come up with and sell, about
how you're bridging that knowledge gap.