Calculating MTBF for an electrical component

[jag53]

Hi,

I am trying to calculate the MTBF for a certain electrical part that is being specially made. I have the bill of materials for this component and have purchased the Electrical Parts Reliability Data (EPRD) to obtain failure rates for each individual component. If I am correct once I have all the failure rates for each individual component, from there I can calculate the MTBF for each component (MTBF=1/Failure rate). Is the correct equation to calculate the MTBF for the electrical part = 1/(1/MTBF1+1/MTBF2+...+1/MTBFn)? Will this equation give me the MTBF for the electrical part as a whole?

Thanks,
jag53

[Marc]

Any takers on calculating MTBF for an electrical component?

[Jim Wynne]

There are a lot of factors to consider, such as serial vs. parallel component failures (the former assumes that a single component failure will render the assembly nonfunctional; the latter assumes that the assembly will continue to function so long as 1 of n components continues to function.

I recommend the OP have a look at MIL-HDBK-217F, the military standard for electronic assembly reliability testing.

Military Standards

[Tim Folkerts]

Jag53,

Yes, your equation is correct, at least for a basic mathematical prediction.
MTBF = 1 / (1/MTBF1+1/MTBF2+...+1/MTBFn)

This assumes 1) the parts are used in typical circumstances (e.g. no high temperatures) and 2) failure of any one part will lead to the failure of the whole (e.g. no redundancy). Also, if the design puts a heavy load on one part, it could fail more quickly (and if it has a light load, it might last longer).

Tim F

[Jim Wynne]

Originally Posted by Tim Folkerts

Jag53,

Yes, your equation is correct, at least for a basic mathematical prediction.
MTBF = 1 / (1/MTBF1+1/MTBF2+...+1/MTBFn)

This assumes 1) the parts are used in typical circumstances (e.g. no high temperatures) and 2) failure of any one part will lead to the failure of the whole (e.g. no redundancy). Also, if the design puts a heavy load on one part, it could fail more quickly (and if it has a light load, it might last longer).

Tim F

Another thing to remember is that the published MTBF values for components are not always reliable (pun intended) themselves. I've always thought that it's OK to use them as baseline values, but it's best to develop assembly reliability data independently.

[Sherin MR]

Hi,
Can anyone please provide the parameters to calculate failure in time of
1.Resistor
2.Capacitor
3.Diodes
4.Relays
I have gone thru the replies that i got but it doesn't satisfy my needs.

Regards,
Sherin.

[Jupitor]

OK, let me see if I can be of some help. Here goes!

1. Get hold of a copy of Mil-HDBK-217 F.
2. Identify your component type for which you want to calculate failure
rate, e.g. Resistor
3. What type of resistor? say wire-wound resistor. Identify the table for the
selected component type
4. In the table you will find a formula that goes like this -

Lambda b (Base failure rate- e.g. for WW resistor it is 0.0024/million hrs,)
x Pie T (Operating temperature factor)
x Pie P (factor for rated power at the selected op. temp)
x Pie E (factor for Enviromental conditions of usage, e.g. used in an
aircraft or in an airconditioned lab.etc.)
x Pie Q ( quality factor, ie, whther the component is a high reliability part
or is of commecial grade etc.)esistor.
x Pie S (the stress factor-i.e the percentage of rated power you actually
apply to the component in the interest of higher/useful life)
The table gving values for all the above factors are available in Mil-HDBK-217F.
Does this make sense to you?

[Sorry for not putting in the symbols in stead of writing Lambda etc. I don't have them in my character map].

Jupitor

[Stijloor]

Originally Posted by Jupitor

OK, let me see if I can be of some help. Here goes!

1. Get hold of a copy of Mil-HDBK-217 F.

Mil-HDBK-217 F can be found here.

Stijloor.