Design without mistakes? It never happens…or?

20 years ago at APEC, I met a guy who had been a long tenured VP of Engineering at a power conversion company. We were talking about the power conversion design process and the inevitable mistakes we make along the way.

He said: ”I told my guys, if any of you ever designs a power supply or even a PCB that doesn’t require rework or design change I will buy you a Porsche”.

“And?” I asked.

“It never happened” he said with a smile.

“Probably because you don’t have guys like me” I said to myself.

My boss never offered such an incentive. Some years later I realized that even if he had, I would not have won the prize. Design without mistakes? It never happens. No matter how careful you are there will always be that one thing that you overlooked.

What also never happens is that you go through the power conversion design process without blowing the unit up. We all know too well that lonely feeling: sitting at the lab bench, smoke billowing out of the unit with unmistakable smell of a burnt semiconductor. You are shellshocked, your face is charred and vision blurred while the crowd of onlookers congregates around you offering fake compassion. You know they are glaring with schadenfreude but you can’t blame them either, we’ve all been on both sides. I mean what’s more exciting than a big bang that we all know will come. We just don’t know when. Or will it?

I was reminded of that feeling quite recently when I had to do my first design while fully relying on HIL. Since access to the lab was limited over the past several months, we had minimal debug time with real hardware. Still, the design verification process was complete. The unit is with the customer. Nothing blew up. Was I more careful? No, we don’t change our attitudes that easily. It’s still the same customers who want the product delivered quickly, it’s still the same limited resources that wouldn’t allow us to investigate all the details, we are still accepting failure (and long recovery from it) as an integral part of the design process. What’s different now is that when you make a mistake nothing blows up. You don’t have to go into desperation mode of both trying to repair the unit and understand what went wrong when you couldn’t even capture the event. All you need to do is clear the flag in your HIL device and you can repeat the same thing over and over until you fix the problem. What took hours, days or weeks can now be resolved in a matter of minutes.

It’s time to go to my boss and ask him: “Will you buy me a Porsche if I design this new power supply without blowing anything up?”