Test Talk
Each issue of the DFT Newsletter brings another hilarious look at DFT by Alfred Crouch, author of the renowned book "DFT for Digital IC's and Embedded Core Systems."
June 2008
The Comfort Zone
Well, here I am again – the invisible glass-cubicle dwelling technical guy, Noah Bawdy, with my woeful tale about trying to find a new job. I thought this would have nothing to do with the fact that my boss gets his ideas from late-night TV, but it looks like I can’t even go to other companies in my industry without carrying the stigma and stench of my boss.
As you may be aware (but I won’t underestimate the ostrich-factor with engineers, and may have sand in your ears from burying your head in your cubicle) recent events in job outsourcing, the escalating price of gas, the changing political climate, and of course, the sordid lives of Hollywood Stars have all affected the economy negatively. Even our company, that scrapes the bottom of the shoe of the semiconductor market to find our market space, has been suffering lately. I’m not even going to go into our “identify who to layoff” algorithm, but let’s just say it has something to do with how worn out the bottom of your chair is, if you can or can’t read the finger-smudged letters on your computer keyboard, and where and how large the wear patterns in your cubicle carpet squares are. The uncertainty put me a little out of my comfort zone, so I decided to do some pre-emptive job hunting.
Being a die-hard engineer, I did as most engineers do. I over analyzed the job market and tried to find a job as close to the one that I currently had, so that my impact and value to a new organization would be directly visible in the new position (as opposed to having to learn something new and being on the same footing as, say, a fresh-out from college that still remembers what a Closed Line Integral or a Bessel’s Function is). So, after 10 spreadsheets, 2 PERL analysis scripts, 3 statistical analysis packages, 1 binary decision diagram, a coin toss, and a new MAC to run it all on (with a Linux partition, of course), I found that my best possibility of the least amount of salary and title erosion was with Acme Amazing Really Global High-Volume Semiconductors (AARGHs, Inc.).
I dusted off (okay, vacuumed aggressively with a modified turbocharged homemade wind-tunnel tornado suck-machine) my polyester suit that I only wore twice when I interviewed in college and I guess it seemed to still fit. I got the buttons to close with a little effort and the seams only opened up by a few millimeters. On the positive side, after all those years in the dry, static-filled air of a cubicle-infested semiconductor company, I have permanent “engineer hair” and so I won’t have to use any cosmetic products to perfect that “oh so technical” look.
What I have learned in all of my years at my current company is that interviews don’t really have anything to do with hiring people, they are really about collecting competitive data. So the first thing asked of me was a round-about set of sneaky questions that could have cut about 15 minutes out of the interview if I was just asked, “What is your EDA Selection Process?” I figured this out and since my company was close to the bottom of market share, I wasn’t worried about giving up any useful information.
I explained, “You see our process came from my boss’s late-night TV management idea – EDA Cage Wrestling. We put the AEs and Sales drudges for the competing EDA Tool companies in a closed cage (er, conference room with a broken thermostat and a door that had a broken lock and could only be opened from the outside), then we slide a signed PO under the door. Only one EDA company came out with the tattered and shredded remains of the PO. We figured that they fought the hardest and so we purchased their tools.”
The interviewer first looked at me as if I was lying to him. Then he gave me a shrewd look as if he thought I was playing some sort of mind game with him.
Since my resume said that I was a DFT engineer, the second set of questions was about our design-for-test process. The interviewer asked a bunch of questions that seemed to bat around, “How do you measure high coverage? How do you get high coverage vectors? How do you get these high-coverage vectors before silicon comes out? How do you manage the number of vectors so they fit in the tester and result in a test-time that is economical?”
Well, to be honest, my company has a completely unique solution to this subject, so I knew I was going to have a problem answering these questions in a way that the interviewer would understand. My first thought was, “Well, they can’t fault me for honesty!” So, here is what I had to tell the interviewer:
“You remember that I told you that my boss gets his ideas from late-night TV shows, uh, well, he got this idea that “coverage” equates to car insurance and late-night TV has all sorts of advertisements involving catchy icons like cavemen, lizards, well-known celebrities, and cartoon spies that can all save you time, effort, and money if you use their insurance. So we invented a coverage icon that was a robot with a hammer standing next to a svelte hot sports car with a dent. We stamp that icon on the top of all of the chips and put it on all of our literature and packaging. This implies that we can fix any dent in the silicon.”
The interviewer’s eyes bugged out a little at this point, but he didn’t say anything, so I continued…
“As for vectors to get high-coverage, we tell the customer that the best vectors to see if the chip works for them is their own operation vectors…so we don’t really make any vectors, we get the customer to give us their vectors after they do some simulation stuff to capture the values at the chip pin boundary – seems to be a lot of work on their part – and we test the chip in manufacturing with their vectors. We generally get these vectors about 6 months after the chip goes into production. We wave our hands a lot when we talk to them until then.”
I heard a gulp and a cough out of the interviewer…
“And as for cost management, we target all of the used idle testers that we can find at test houses that aren’t doing well financially – they are hungry for business and don’t complain very much. They do seem to balk at the volume of chips we want them to handle, though – when we find a really, really hungry test-house, we dump all of our problematic chips on them. We’ve seem to have put a lot of them out of business…”
This seemed to strike a chord or had some significant meaning with the interviewer and I saw him furiously writing something down in the folder in front of him…
“Ultimately,” I rambled on since I was on a roll, “the way we measure our coverage and quality is to see how many lawyers it takes to get our customers to quit threatening to sue us and how long it takes for the lawyers to negotiate a settlement. We used to have a warehouse for returned parts, but once it was full and we knew we weren’t going to really do anything with the bad parts we just stopped taking bad parts. We made up address and all the bad parts were sent there.”
I stopped talking when the interviewer fell over backwards in his chair. From under the other side of the table somewhere I heard, “Haven’t you people heard of Mentor and their TestKompress and YieldAssist tools…it is the correct answer to all of those questions. They deliver fewer high-coverage vectors using multiple fault models that handle the latest process issues; earlier delivery of those vectors – usually before silicon comes back from the fab; those vectors easily convert and fit in the most common ATE platforms; and with YieldAssist, you can diagnose the source of any systematic fails and yield-loss problems.”
I saw a hand at the edge of the table, then the rest of the interviewer as he pulled himself up and said, “Well, Mr. Bawdy, you are obviously way too far outside of my company’s “Comfort Zone” and so I think this interview is over. We’ll pretend that things went well here and will tell you that we will call you if anything comes up, but,” and he winked at me, “we know what will really happen.”
Oddly enough, I got pretty much the same thing from the next few interviews with different companies. I guess I had better work harder to make sure that my bottom-of-the-barrel company continues to thrive because all that I have learned in my company over the years sure doesn’t seem to be needed or valued by other companies. The funny thing is that my Boss wants to know where I got the genius idea to replace all of our existing uncomfortable and legally-questionable test process with TestKompress (since he didn’t see it on late-night TV). It turns out we are no longer the very-bottom-of-the-shoe market-share and we are climbing fast in the rankings…unfortunately, we seemed to have lost the AARGHs account for some reason, they no longer buy chips from us.