Notice: This technical discussion forum has been established by the SRE to assist all reliability engineers, not just SRE members. To post your technical problem, solution, question, or answer here, send an e-mail to SRE webmaster. Your e-mail should follow the format of those already posted. Postings are accepted from anyone, as long as they relate to the reliability field. It works best if your posting contains contact information for possible follow-up. But, if necessary, postings can be anonymous - just so state in your email! Also, let us know when you do get workable solution or answer. We want to post it so everybody will benefit.

Problem/Question (December 10, 2002):
I am looking for a Standard that describes the process of creating reliable BIT/BITE diagnostic software for Health Management Systems in military vehicles.

What is needed is a step-by-step set of guidelines that clearly define the development process, similar to MIL Stds.

If there is no Standard, I would like to see Technical Papers on related subjects.

Also, my search includes terms "Bayesian Networks" and "Autonomics", in any and all applications capacity.

Robert Poltz, SRE member-at-large
Phone: 805-961-7034
E-mail: robert.poltz@gm.com

Solution/Answer: Can you help?

Problem/Question (September 20, 2002):
I am a structural engineer (buildings) with little background in reliability predictions and would like an opinion of how I would utilize reliability approach to better understand life expectancy of steel moment frame connections based upon three tests specimens of three connection type s-n curves. Thank you. Jay

Jay Allen, SE
E-mail: jayallense@yahoo.com

Solution/Answer: Can you help?

Reply:(20 Sep, 2002)
This type of problem is not as simple or straight forward as it might seem.

Emile J Gumbel ( a Nobel prize winner ) has performed many prediction of Statistical extrema. (See Statistics of Extremes by E. J. Gumbel ) there are many applications to statistical extremes of wind loads, meteorological phenomena such as floods and droughts etc.

If you give us more information, perhaps we can help. Maybe your problem is a simple one.

Barry Schlund
Visit us at SYSTEMS-RELIABILITY.com or call (928) 369-4350

Problem/Question (July 10, 2002):
I am looking for a free software which is capable to perform the tracking of my part in the chamber ( log in, log out, time calculation as well as reminder notice ).

Thanks.

Jason Yan Li
E-mail: jasonli@oluma.com

Solution/Answer: Can you help?

Problem/Question (July 2, 2002):
I am a Reliability Engineering working for the Army and have been ask to contact organizations in the private sector to see if anyone can assist with our concern. We are seeking a single matrix to measure vehicle reliability, maintainability, availability, performance and operational costs but without funding an expensive data collection system. We would like a matrix that could be used during initial user testing and follow on field application. Any ideas or experience out there?

Mike Barton
E-mail: bartonm@tacom.army.mil
or
E-mail: leitheip@tacom.army.mil

Solution/Answer: Can you help?

Reply:(July 2, 2002)
Hi,
This appears to be simple enough. In one matrix you would be able to represent all your data without any expensive data collection system.

Since this question has been posted a year back, I hope you have already got the answer. If not, I would be able to provide some samples of such calculations.

Regards,
Dibyendu De
Reliability Management Consultant
www.rmcglobal.com

Problem/Question (June 27, 2002):
Good day.

I would like to know the procdure of going about doing PCB testing on electrical transformers. Can you please help?
Thank you.

Renash Rampersadh
Electrical Technician
Steelplant
Iscor Newcastle
Phone: (034) 314 8362
E-mail: Renash.Rampersadh@Iscor.com

Solution/Answer: Can you help?

Reply:(27 June, 2002)
3M in Minneapolis has some fluids (Fluoro carbons) which are non conductive. They come in different temperature ranges, from very very cold to very very hot.

They are used throughout the electronics industry for life testing and specialized tests.

Barry Schlund
Visit us at SYSTEMS-RELIABILITY.com

Problem/Question (June 19, 2002):
We had a failure in the motherboard PCB of a rectangular box. The chassis is Magnesium casting and two rectangular PCBS are attached to the motherboard thro long connectors. Wedge locks hold these PCBS and they are 8-layered PCBs with three horizontal stiffeners.
The unit had undergone ESS also with high temperature of 700 C and low temperature of -400C with 20minutes exposure time at each temperature for each cycle and total number of cycles are 5. The unit had gone thro random vibration before and after the thermal cycling. The unit then undergone limited acceptance testing like random vibration, classical shock, high and low temperature tests. After some time track discontinuity in the six-layered motherboard is observed. Our static and dynamic analysis indicate that the shear stress on insulation laminate is with in safe limit.
May I request you give some guidance to understand the cause of such failure?
Secondly, what is maximum Dynamic load (in terms of g rms) a PCB can with stand?
We know that a PCB can withstand max. of 1250 C to 1300 C but I do not know the limit in negative temperature side. Can you clarify Please?
You may kindly indicate the possible reasons for track failure.
With warm regards.
Prasad Vizag
E-mail: kr_guntuku@yahoo.com

Solution/Answer: Can you help?

Reply:(19 June, 2002)
It sounds like you are the victim of " cold flow ", or " plastic deformation ". The thermal cycling you described can produce extreme stresses, particularly on rigid (stiffened) bodies. It results in a residual (or built up) stress which you could easily observe with laser interferometric technology. Solder (and the PWB material itself) will tend to stress relieve themselves through creep and relaxation.

Can you be more specific about where and how the failure(s) occurred? A plated through hole? A Broken chip cap or chip resistor? ????

Visit us at SYSTEMS-RELIABILITY.com

You can provide further detail on the "Contact Us" page.
p.s.
Your PCB may be able to withstand 1250°C but semiconductors self destruct at 425°C when the Aluminum metallization desolves into the Silicon. Be careful !

Problem/Question (May 27, 2002):
I am trying to find reliability data (or failure statistics and I'll calculate) for High Voltage Porcelain Insulators. I would hope to disregard early (fitting stress related) failures.

Usual web searches have failed to identify, Bellcore is no help and I cant find a referance in Mil-217.

Can anyone help with this?

Regards,
Mark Bonthron
E-mail: Mark.Bonthron@OLEAlliance.co.uk

Solution/Answer: Can you help?

Reply:(27 May, 2002)
You may use Thermal Imaging as a primary tool and then conduct some field tests and use an equation to find out the desired reliability.

Regards,
Dibyendu De
Reliability Management Consultant
www.rmcglobal.com

Problem/Question (May 22, 2002):
I am performing an MTTR analysis according to Procedure V in MIL-HDBK-472N1 (as reprinted in Appendix D of MIL-HDBK-470A). It doesn't seem as though the equations specified for MTTR, Rn, and Rnj are correct when the process is followed step by step. Has anyone had experience performing this analysis?

Thanks for your help,

Nancy M. Pyron
Reliability Engineer
Coastal Environmental Systems, Inc.
Phone: 206-682-6048, ext. 153
E-mail: npyron@coastalenvironmental.com

Solution/Answer: Can you help?

Problem/Question (April 10, 2002):
Hi, I'm trying to design an uncomplicated model for reliability prediction in Commercially Available Off The Shelf (COTS) Automatic Test Equipment (ATE). Suppose a user has a requirement of DEMONSTRATED 1000hr MTBF for a test suite. I'm going to evaluate reliability for a series of tech proposals. Most proposals will primarily use COTS for a proposed suite solution. A suite is characterized by a chassis (VXI, PXI, etc), slot zero controller(s), chassis power supply, and discrete instruments (scope, function gen, dmm, DC supplies, etc). A failure is described as any event which results in a maintenance action, repair action, or degradation in performance. The suite must run uninterrupted in the field for 1000hrs.

I want to provide a non-complex model to use as guidance to offerors. I suggest that any proposed suite should be evaluated as a serial system. If any one component fails, and the design doesn't offer redundancy, then MTBF is breached. My model uses the exponential distribution for failure of electrical components. My EXCEL sheet lists by column, instruments, number of instruments, vendor spec'd instrument failure rate, offeror proposed derated failure rate. I then multiply instrument failure rate by # of like instruments, I then sum the failure rates, and apply R(T) = e^-(sum*T), where T = 1000. Likewise, I do the same for the derated rates. I don't want to complicate the model at all, I want it to be simple to use, I have other aspects of proposals to evaluate.

In summary, offerors will propose predicted reliabilities for their suite, but they'll have to describe instrument failure rates in terms of (column 1) vendor posted spec sheet rating and a (column 2) derating factor. Do I let the offerors proposed their own derating factors for each discrete instrument? Or should I try to "band" the derating for the instruments...digital inst = X, analog = Y, RF = Z and suggest the factors myself? I'm not a reliability guy. Is this a reasonable approach to a simplistic model for an application like this? Don't want to buy S/W, don't want to train up eval team or constrain offerors w/ one S/W application, rather use EXCEL and make it simple.

Scott Dotson
Program Manager
Warner Robins Air Logistics Center
Warner Robins, GA 31098
E-mail: SCOTTLDOTSON@aol.com

Solution/Answer: Can you help?

Reply:(10 April, 2002)
The simple thing would be to ask the suppliers propose the MTBF of the instruments that they supply and the conditions under which the instruments may likely to fail and then put it in the contract that the instruments would be accepted after a Reliability Qualification/Acceptance testing with say 85% confidence limits (using Chi square) method.
That makes it simple and you may still use the Excel sheet. I have used this method for a major aerospace company. So, I know that it works.

Regards,
Dibyendu De
Reliability Management Consultant
www.rmcglobal.com

Problem/Question (March 22, 2002):
My organization is performing a design-phase FMECA using MIL STD 1629A and we have a discussion going regarding the proper use of the "Failure Detection Method" column on the 1629A format. The MIL STD specifically says that that column is used to describe how the equipment operator will discover the component failure. Another reference we are using says that -- for a design-phase FMECA -- what is important is to identify what part of your design effort is intended to detect failure of a hypothetically-fielded component. In other words, 1629A identifies that "Failure Detection Method" for an Operation-and-Maintenance-phase FMECA and that the appropriate focus of a design-phase FMECA should be on the "Failure Detection Method" that is pertinent to a product's design phase.

Comments or Suggestions?

Robin E. Franke
Principal Engineering Analyst
Teledyne Energy Systems, Inc.
10707 Gilroy Road
Hunt Valley, MD 21031-1311
E-mail: robin.franke@teledynees.com

Solution/Answer: Can you help?

Reply:(March 22, 2002)
Would be helpful if you may kindly say what exactly are you trying to find out.

Regards,
Dibyendu De
Reliability Management Consultant
www.rmcglobal.com

Problem/Question (February 26, 2002):
I'm working on a reliability analysis for a Permanent Magnet Motor. I was looking at MIL-HDBK-217 model but the model is for motors below 1 HP. Also I was trying the model from David Taylor Research Center (Reliability Prediction Procedures for Mechanical Equipment) but the model is too sophisticate. Is there any other method and/or reference that I could use for my analisys. I'll really appreciate your help.

Thank

Brenda Ortiz
E-mail: bov_rum@yahoo.com

Solution/Answer: Can you help?

Problem/Question (February 20, 2002):
Hi,

I am looking for information related to the MRB Process (Maintenance Review Board) and the MSG-3 Analysis (Maintenance Steering Group) that are based on the RCM Concept (Reliability Centered Maintenance) and used in the development of an initial maintenance plan. I would like to make a contact in order to discuss some points ( like the position of the supplier, vendor in a MRB Process. A new aircraft must have an initial maintenance plan. The MRB Process is used to develop that Initial maintenance plan using the MSG analysis. I have some data on the tasks, responsabilities and the place of the operator and the manufacturer but i did not found any data on the vendor and its place on the MRB Process.)

Thanks for all to answer me.

Sylvain Bonnel
E-mail: Sylvain.bonnel@ixi.fr

Solution/Answer: Can you help?

Problem/Question (February 15, 2002):
We design & build custom (one of a kind) manufacturing automation systems. I am trying to calculate the "reliability projection" of our system using the series probability formula, Rs=(Ra)(Rb)(Rc).... , to meet QS 9000 TE R&M requirements.

I have received performance data from suppliers, e.g., 1,00,000 hrs or 1,000,000 cycles, etc., and calculated a failure rate (lambda) based on Chi Square (X^2) where:

lambda=X^2(1-a)/2t
a=1-percent confidence desired (.05)
r= number of failures = (0)
dof=degrees of freedom= 2(r+1)
t= device hours= supplier data

I want to calculate system Reliability from the reliability equation: R=e^(-lambda(T)), where "T" is a specified interval of time, but don't know:

1) Is there an industry standard definition for "T", e.g., machine life, shift duration, preventive maintenance interval, etc., or can I make and state an assumption?
2) Is this even a statistically valid way to calculate a "reliability projection"?
3) What is a "practical" reliability engineering reference book? (I am not a statistician nor a reliability engineer, but I have taken a couple intro level college statistics courses.)

Thanks,
Steve Lardner
QA Manager
Wes-Tech Automation Systems
720 Dartmouth Lane
Buffalo Grove, IL 60089
Phone: 847-419-7937
E-mail: slardner@wes-tech.com>

Solution/Answer: Can you help?

Reply:(15 Feb, 2002)
T is the demonstrated Test time summed over all test units. Chi-Sqr = 2 x T / M where M is the desired MTBF. Nu is the number of degrees of freedom = 2 x n + 2, where n is the number of failures encountered during the T test hours.

Plug these into the Chi-Square table and use simple multiplication / division to obtain your answers.

Barry Schlund
Visit us at SYSTEMS-RELIABILITY.com

Problem/Question (January 22, 2002):
I am a senior at Indian Institue of Technology, New Delhi and looking for a sampling plan that can minimize the total cost (inspection cost + faliure cost). The data available is the relevant cost factors and the target is to minimize the total cost. the project is a part of my final project.

Sharad Mittal
Department of Mechanical Engineering
Indian Institute of Technology
New Delhi, India
E-mail: sharad_iitd@yahoo.com>

Solution/Answer: Can you help?

Problem/Question (January 16, 2002):
I am in charge of getting our company QS 9000 certified. Part of that system requires that we make MTBF calculations for the specialty machines that we build. I have been able to gather MTBF information for individual components of our machines (sensors, drill units, etc.). Unfortunately I do not know how to take that information and turn it into a MTBF calculation for the entire machine. Can anyone be of assistance.

Thank you,
Chris Adams
E-mail: cadams@manntool.com>

Solution/Answer: Can you help?

(September 15, 2003)
I came to know of your problem thru the web. There are basically two approaches to do the calculations.

One is the traditional method. In this method you would have to find the way in which the components are arranged (series, parallel so on) and then apply the formulae to obtain the answer. Refer any standard text on the subject.

However the problem is that the answer does not quite reflect the reality in the sense that the probability of success is too small to be of any meaning.

This leaves us with two other approaches which we have developed and applied in industries with great success. We obtain practical and meaningful results. In brief, the one of the method is to run your machine for sometime and then monitor the health of the machine. You may then calculate the general MTBF of the machine by obtaining the data where the first deterioration is observed.

The other is to segregate the components into three classes in order of importance. Then find out the least probable from each of the classes. The least count gives you the practical MTBF of your machine.

Since you have posted your query long ago it seems that you already have some answer or worked out something. Would you be still interested in any alternative solution that would be of practical interest?

Regards,
Dibyendu De
Reliability Management Consultant
www.rmcglobal.com

Reply:(16 Jan, 2002)
If you are using LOTUS 123, Quattro Pro or Excel as your Spreadsheet then your problem (and the solution) is a simple one.

Each of the spreadsheets has an optimization option, where you can choose from varying sets of parameters, minimizing some, maximizing others, or achieving target values for some of them.

Hopefully, you have a user's manual for your spreadsheet.

Barry Schlund
Visit us at SYSTEMS-RELIABILITY.com