Today’s guest is Nat Ozarin, Senior Software and Reliability Engineer with The Omnicon Group Inc. He has decades of experience troubleshooting a wide range of issues with a particular emphasis on aircraft electrical systems. During the interview Nat reviews his academic and corporate background. He details his frustration with attempting to do bent pin analysis in a manual fashion. Nat expands on the logic behind troubleshooting and some proprietary tools that The Omnicom Group, Inc. can provide.
This a lively 22 minute interview. If you wish to scan the conversation, the transcript is provided below.
John Gilroy Welcome to the Lectromec podcast with John Gilroy. Lectromec provides you the turn-key solution for aerospace wiring assessments, testing, and certification. The focus of this podcast is to keep you informed of trends in the rapidly changing aerospace industry. Today’s guest is Nat Ozarin, Senior Software and Liability Engineer with the Omnicon Group. In 2009 he was named Reliability Engineer of the Year by the IEEE Reliability Society and some say he is reputed to be the world’s foremost authority on bent pin analysis, and that’s our topic today. Nat regularly presents at technical conferences, most recently at the 62nd Annual Reliability and Maintainability Symposium on January 25th, 2016 in Tucson, Arizona. How are you doing today, Nat?
Nat Ozarin Hi, I’m good. How are you John?
John Gilroy Wow, that’s, uh, quite a strong background. I looked you up online, you have an MBA and you’ve got a master’s degree in electrical engineering. Wow, that’s a lot of studying, isn’t it?
Nat Ozarin Uh, yeah it is. I rather enjoyed it, too, but, uh, enough already.
John Gilroy (Laughing)
Nat Ozarin Concentrating with, uh, more practical things now.
John Gilroy Yeah, we’ve got to establish your reliability and credibility before we move on. At Lectromec here, we do regular blog posts and videos and oh, a couple weeks back we put together a blog post on bent pin analysis and, Nat sent me an email and he had some comments on that and I figured, hey he’d be a great guest to have on. So that’s what precipitated this discussion, was Lectromec’s blog post on bent pin analysis and so in order to learn more about Nat, what I did is I went online, in fact, this morning, I went to Google and typed in bent pin analysis. First thing that popped up was your white paper, isn’t it?
Nat Ozarin Uh, yeah, I was (laughs), very pleased to see that. Actually I think that’s a very good paper.
John Gilroy I think so too, I’m going through it and so if you’re- if you’re a student, or if you’re involved in this industry, or you think you know about this industry, this is a free white paper available online. You just go to Google, type in ‘bent pin analysis’ and you can download it and, uh, I think it’s very thorough and well-done, it’s got illustrations. I think most of the people listening would benefit from downloading this, so it’s…good job. I like sharing information with the people in the industry and that’s what we are doing this podcast, trying to, share information.
My first question for you, Nat is, we know you have a strong background in academic areas. What made you emphasize aerospace and then go into this little area of bent pin analysis?
Nat Ozarin My first job out of college, a hundred years ago, was in aerospace industries and, about half of my years since that time has been in aerospace. In my current job, I was given the assignment to do a Failure Modes Effects Analysis (FMEA) on some circuit boards and connectors for a commercial aircraft and, when doing the bent pin analysis, I got through maybe 5 of the 7 connectors and I said “This is ridiculous. There’s got to be a way to have a computer, uh, take care of this; it’s very routine.’ And just to finish the job, I came up with a crude way of doing it using Excel, but I realized that was not the right way so, I began developing a geometric approach, which would use double precision trigonometry and represent the geometry of connectors by physical sizes, locations, pin diameters, and so forth, and let the machine, in this case a very elaborate, visual basic program in Microsoft Excel, come up with all the permutations of bent pins, the probabilities of each permutation to generate the worksheet. The goal was to….everything that a computer could do, let the computer do, and leave only for humans, what only humans can do. And the algorithms were developed over a period of years, but I think they worked out quite nicely and we have used them.
John Gilroy Yeah, I think of…if I’m doing a project around the house, do I want to use an electric drill or a manual drill? That’s how old I am. I can remember, you want to use electric drill. So you want to use automated tools when you can, I think that’s kind of the, the genesis of a lot of your work here, is that right?
Nat Ozarin Yeah a machine obviously makes less mistakes than a human, is a hell of a lot faster than a human, and it provides repeatable results, if there’s any doubt, you can always revisit what went on and verify it.
John Gilroy In the first paragraph of your white paper, you talk about a human doing some analysis and sometimes it’s incomplete, it could be inaccurate, and maybe it doesn’t have a real good basis for doing any prediction of future problems, and what you found is automating it using some of your software, you can improve all 3 of those areas, is that right?
Nat Ozarin That’s right. Typically, you know, there aren’t that many reference sources as to how to conduct a bent pin analysis and usually it’s done graphically, some human looks at a layout of the connector and decides whether or not pin A can reach pin R or pin K, but wrongly assuming that one pin can reach another is just as incorrect as assuming that it can’t when in fact it can. So there’s all kinds of opportunities for errors here and the point of the analysis is to identify real hazards and you want to be as thorough as possible on that end.
John Gilroy So, what I find interesting is that, really from a human perspective, an engineer’s perspective, what this does, is it automates some functions and it allows you to focus in on circuit the systems instead of the pins and wire, so it actually makes you more productive.
Nat Ozarin It does. Very often the dog work of just trying to figure out what can short to what and then filling out repeated information and worksheets can take weeks. A machine can do it in minutes and in the time that you save, you can now concentrate on what’s really going on in the circuits and not worrying about geometry or, things that you really shouldn’t be worrying about.
John Gilroy So this is a piece of software that you buy from your company?
Nat Ozarin No, we use it, for our customers. We run it for our customers. We will consider, leasing it, but as it is, because it’s an engineering package, it doesn’t have all the manuals and so forth that you would normally expect with a shrink wrapped package. It does have a user manual and we will consider leasing it to another company that has a need for it and providing training for it, but we’re not selling it as a package.
Very often the dog work of just trying to figure out what can short to what and then filling out repeated information and worksheets can take weeks. A machine can do it in minutes and in the time that you save, you can now concentrate on what’s really going on in the circuits and not worrying about geometry or, things that you really shouldn’t be worrying about.
John Gilroy Now from Lectromec’s perspective, we usually diagnose some problems with bent pin with installation or with maintenance, is that where you see the problems popping up typically?
Nat Ozarin There are two ways that connectors fail. One is mishandling, the other is, particularly in aircraft, wires get pulled out of their crimped pins, or corrosion sets in. The analysis looks at both cases where mishandling causes something to physically bend and also the situation where nothing is bent, but corrosion has gotten into the connector and it’s caused a path to open.
John Gilroy That’s interesting. Just trying to expand this move beyond aerospace, are there other industries that have this kind of problem?
Nat Ozarin Any kind of an industry, medical industries and so forth where the possibility of a shorted path or an open path can lead to undesirable results. Certainly this is applicable to those industries as well.
John Gilroy It’s just amplified much more in the aerospace industry, because bad things can happen. If a car doesn’t start or something, that’s one thing, but if for example, a circuit fails in an aircraft, it can be critical. I guess that’s one thing, once you gather your data and you make an analysis, do you separate between critical and non-critical issues?
Nat Ozarin Yeah, because if there’s a bent pin or an open connection something’s going to go wrong. And hopefully in most cases, it would be benign or at least be a situation where you can operate in some degraded mode, but safely. What you’re particularly interested in is an unexpected short or open that can cause some catastrophic result, or really any other result that is just not acceptable.
John Gilroy So, I’m trying to view this from a novice’s perspective, or someone outside the industry. It looks like you’ve taken a mechanical process and automated it, kind of moved it from the 21st century that’s what it is, it’s moving into a more automated phase.
Nat Ozarin That’s the idea. There’s no reason for humans to spend time either second guessing what appears to be feasible. And certainly it’s not feasible for humans to do trigonometry on connectors that have 60, 80 pins to try to figure out what can short to what. This is something that is just made for a machine.
John Gilroy Yeah, that makes sense. So let’s say my son has friends who are engineering students and they graduate and they start working on these kind of systems, what typical problems, or typical mistakes do novices make when they try to diagnose a circuit problem?
Nat Ozarin I think the biggest problem is a failure to anticipate unexpected things. Typically when you’re designing and testing something you’re concentrating on what it’s supposed to do. Very often during testing, there’s inadequate emphasis on what happens if something that nobody ever dreamed of happens.
John Gilroy Hm.
Nat Ozarin It happens all the time that something fails that you get unexpected input data. You may get unexpected software behavior, which affects hardware. And I would urge people who are developing equipment to spend more time thinking about what could go wrong even though it isn’t anticipated.
John Gilroy Yeah, I once had a car problem where the fan in my car, one of the blades just broke off (laughs) and the mechanics said “I’ve never seen this before,” and I think that’s how you have to train novices, it’s that strange things do happen, not only in cars but in aircraft as well.
Nat Ozarin A friend of mine had that happen, or similarly to it, where the alternator was held in by two bolts and one of the bolts broke and the fan belt then pulled the entire fan assembly right through the radiator.
John Gilroy Wow.
Nat Ozarin And that’s not something you would think of when you think of one bolt failing, but it has happened.
John Gilroy Yeah, yeah. I guess in engineering school you learn that bolts should actually hold. There’s regulations about that.
Nat Ozarin (Laughter) They should.
John Gilroy If you’re listening to this podcast and maybe baseball season’s here 6, 8 months from now and you want to do some more research and learn more about this, where do you suggest my listeners go to get more information on troubleshooting and maybe troubleshooting aircraft electrical systems?
Nat Ozarin That’s quite a variety and covers a broad area. Certainly bent pin analysis is something you can plug into Google and get information from. There’s a variation that this software tool does, which is, to examine possibilities, permutations, and shorts among connectors without anything being bent, without any paths being open. My understanding is on aircraft, insulation over a period of years tends to become brittle and flake away causing shorts between two paths. A standard kind of analysis is called cable matrix analysis, where any conductor has the same, usually 6 neighbors. And the analysis looks at the connection between any pair of these, which normally should be insulated. But the tool also has a variation, where it will look at permutations of shorts between any one path and any other path and based on the same information, the information being what kind of a signal it is: Is it a ground? Is it an input? It is bidirectional? Is it a spare? And so forth. The algorithms, the software, will either come up with a description of what’s wrong or it will say, if there’s no effect, this particular permutation is benign.
John Gilroy Hm.
Nat Ozarin Other places to look would be to Google “hazard analysis.”
John Gilroy Mm-hmm (affirmative), that’s a good term, yeah, good, good.
Nat Ozarin Which covers aircraft and any other systems.
John Gilroy Are there any stories that you’d like to share with our listeners about any particularly difficult troubleshooting incidents? Or it gets you too mad (laughing)?
Nat Ozarin Well, you know, when I was developing hardware, I’ve had (laughing) my share of problems banging my head into the wall and I won’t go into that, so I’m going to pass on that question.
John Gilroy Sure, sure, sure. So, let’s say someone runs your software and they get some findings, what kind of a reaction does a systems center have when you report back to them? Do they say “that’s impossible,” or what do they say?
Nat Ozarin We’ve discovered a number of things, not only bent pin analysis, but, my company, the Omnicon Group, does a lot of analyses. We’ve done software failure modes analysis, failure modes analysis. And in one case we found that it was possible, for the war head of a missile to become armed inadvertently while it’s in the silo-
John Gilroy That can’t be good (laughs).
Nat Ozarin And they weren’t at all pleased, because they were reduction.
John Gilroy They weren’t at all pleased (laughs), that’s a minor understatement there.
Nat Ozarin Yeah, but they did change the software. In another case with a bent pin analysis, we discovered that if a certain pin bends, it would cause a catastrophic effect.
John Gilroy Mm.
Nat Ozarin And we let the manufacturer know, large aircraft company, and the manufacturer pointed out that this configuration has been in this particular line of aircraft for decades and they recognize the fact, and they’re going to live with it, because if such a short did exist, it would become immediately obvious to the pilot and the plane would never leave the gate. So they decided to live with it.
John Gilroy If you go online you can find, the FAA DERs for electrical systems and equipment. Is bent pin analysis part of the requirements or understanding to get that certification, that compliance?
Nat Ozarin Not necessarily, but the FAA does list bent pin analysis as one of the types of analyses that may be applicable to a system. Bent pin analysis is a Failure Modes and Effects Analysis (FMEA). If a company is doing Failure Modes and Effects Analysis on components, then, bent pin analysis should be considered, because connectors do fail. I felt, in learning about this, that there’s a lot of emphasis on failures of high-tech equipment, but there’s not all that much emphasis on low-tech configurations, such as wiring, and wiring has caused aircraft to fall from the sky. So I believe there should be more emphasis on wiring analysis and bent pin analysis.
…there’s a lot of emphasis on failures of high-tech equipment, but there’s not all that much emphasis on low-tech configurations, such as wiring, and wiring has caused aircraft to fall from the sky. So I believe there should be more emphasis on wiring analysis and bent pin analysis.
John Gilroy In fact, not far from where you live, TWA was 800, was that just 15, 20 years ago wasn’t it? It was due to some electrical troubles with that aircraft.
Anyway, what I normally do at this point in the show is, I ask my guests, what their favorite aircraft is. We’re talking about aircraft and aerospace, so what is your favorite aircraft anyways, Nat?
Nat Ozarin I’m not sure if I have one.
John Gilroy (Laughing)
Nat Ozarin But if I had to choose one, I would probably say a Boeing 737 for a number of reasons, not the least of which, the Omnicon Group, where I work, has done analyses on products that are now a part of Boeing 737.
John Gilroy So you feel safe, huh (laughs)?
Nat Ozarin I feel perfectly safe on a Boeing 737.
John Gilroy Great. Well, I’d like to thank my guest today, Nat Ozarin, Senior Software Reliability Engineer with the Omnicon Group. Thank you very much, Nat.
Nat Ozarin Yeah, thank you so much, John. Take care.
John Gilroy And thank you for listening to the Lectromec Podcast. If you would like to learn more about aerospace wiring, please visit our website at lectromec.com. We have free white papers, newsletters, videos, 173 articles, and much more.
If you enjoyed this article and would like to read more, please consider: