# Reliability of components

• posted

I'm working on a design at the moment where the reliability of the circuit needs to be calculated. I've seen some datasheets where the reliability data (MTBF, component life time, etc) is listed, but most of the components I'm using for this circuit don't list any values.

Are semiconductor companies likely to list values somewhere? I'm not too sure how often components such as ICs, diodes and transistors fail, but surely they would eventually. I could see other components like resistors and capacitors failing more regularly but I'm really not sure how to go about finding/requesting reliability data.

Even if I can't get data from the manufacture, I thought that there may be some test data for certain types of generic components that give a rough guide as to how long, say, a generic transistor should last for. Anyone know of anything like that?

Anyone ever done calculations like this and found all the data necessary of it?

• posted

Not clear from your posting, but you are aware of the Bellcore and MIL- HDBK-217 methods, yes?

• posted

Two accepted methods of reliability prediction are included in MIL-HDBK-217: parts-stress analysis and parts-count analysis. Copies of the handbook are available online, along with some later revision pages and a cancellation notice (mil standards are, mostly, no longer permitted to be specified as contract requirements and tons of them were cancelled as a result; the techniques still hold though). The old Bell Labs had a similar prediction method, nowadays usually called the belcore method. At the end of the grinding, the number that falls out is the predicted failures per 10^6 hours of operation; invert for MTBF.

For example, the parts-stress predicted failure rate of a low frequency bipolar transistor is calculated (using some assumptions here) as: Lp = Lb * Pt * Pa * Pr * Ps * Pq * Pe, where Lb, base failure rate, 0.00074 Pt, temperature factor, 3.0 (for an example junction temp of 80 C) Pa, application factor, 0.70 (for a switching application) Pr, power rating factor, 0.83 Ps, voltage stress factor, 0.11 Pq, quality factor, 8.0 Pe, environmental factor, 1.0 for a predicted failure rate of about 0.001 failures per 10^6 hours.

Sum the calculations for every component on the board, for every board. Interconnects count, too. Mix well with a reliability block diagram and stir in a FMECA.

Relex has some software tools that automate the calculations somewhat.

My advice is to never admit to knowing what reliability is, how it's calculated, or what MTBF stands for. ;-)

```--
Rich Webb     Norfolk, VA```
• posted

circuit

life

circuit

resistors

be

I dedicate a chapter in my book to showing how to calculate the MTBF of a product using Microsoft Excel based on the Bellcore data. The book is "Excel by Example: A Microsoft Excel Cookbook for Electronics Engineers" published by Elsevier/ Newnes ISBN: 0750677562

-Aubrey

• posted

A very good book, BTW!

• posted

The Bush administration wanted to make it even easier for Halliburton to steal the Government's money?

• posted

On Thu, 20 Nov 2008 21:33:45 -0800, Everett M. Greene wrote (in article ):

You think it was HARDER to make money off the military with mil standards?

-- Charlie Springer

• posted

i have a few problems related to MTBF. can pls someone explain the solutions to me.. and if possible can anyone suggest some online reading material regarding these things? i am totally new in this field and i have no prior experience in this type of calculations. I know basic statistics though, if someone can translate those geeky problems in more mathematical terms then it ll be of great help..

1. A computer is being used as a server for a website. It has an MTBF of 15,000 hours. Maintenance personnel for the server are available only from 7AM to 7PM, Monday through Friday. Assume the website is fully operational at 7PM on each Monday, Tuesday, Wednesday, Thursday, and Friday afternoon. What are the probabilities the website will be operating at 7AM on the mornings of: a) Monday, b) Tuesday, c) Wednesday, d) Thursday, e) Friday, f) Saturday, and g) Sunday?

1. A car lighting system consists of two headlights (MTBF of each is

10,000 hours) and two taillights (MTBF of each is 15,000 hours). To be considered fully operational all four lights have to be operable. To be minimally operational at least one headlight and one taillight must be operational. (In other words, in the event of light failure, we may drive a short distance home or to an auto service center so long as we have at least one headlight and at least one taillight working. If both headlights or both taillights fail we must pull over immediately and call for roadside assistance.) What is the MTBF for the car lighting system for fully operational? What is the MTBF for minimally operational?

1. Assume a secured entry system has an MTBF for being =93fully operational=94 of 1000 hours and for being =93minimally operational=94 of

10,000 hours, and that maintenance personnel are available only during normal business hours (7AM =96 5PM, Monday through Friday, not including holidays). What is the probability the secured entry system will have full operational capability at 7AM Monday morning after having no maintenance since 5PM Friday? That it will have at least minimal operational capability at 7AM Monday morning? What are the corresponding numbers for return to work after a three-day weekend if a holiday falls on a Friday or a Monday?
• posted

No-one here is going to do your howmework questions for you, however, you should find these two links useful reading. Google and Wikipedia can also turn up other useful material.

```--
********************************************************************
Paul E. Bennett...............```
• posted

Thanks for the replies guys (and sorry for not replying earlier). I've done my reliability calculations and at the moment it looks like I will need a few high reliability/quality resistors and capacitors.

Does anyone know where I would start looking for "MIL-R-10509" or "MIL-R-22684" spec resistors and "MIL-C-11015 Non-Est Rel" spec capacitors. I've done a few searches and can't seem to find much. I'm not too sure if this is because these standards are non-existent anymore (anyone know?) but I also don't seem to see a lot to do with resistor quality when searching for components.

If anyone has some tips they would be appreciated.

• posted

SecDef policy as of 1994 is to use "performance standards" and commercial specs in place of mil-specs. Goal: cost savings and faster access to advances in commercial technology.

UNCLAS, non-restricted mil-specs and standards are available on-line through the ASSIST database, at

What you may wish to do instead of tracking down qualified MIL parts is to bump the predicted reliability up a notch by spec'ing components with higher temperature, current, or voltage ratings and see how that affects the calculations.

```--
Rich Webb     Norfolk, VA```
• posted
[[ This message was both posted and mailed: see the 'To' and 'Newsgroups' headers for details. ]]

On Sat, 6 Dec 2008 12:31:25 -0800, Paul E. Bennett wrote (in article ):

snip

ilability.htm>

Oh come on! I want to know! Do you think anyone here can give a definitive answer without obfuscations about how long it takes to repair or if they are Earth hours or Moon hours or the factory spec was right?

-- Charlie Springer

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.