Electronic Design

Bob's Mailbox

Hi Bob: Re: "What's All This Spicey Stuff, Anyhow? (Part 3)."* What can I say? Hallelujah! Balm to my soul! (Amen! /rap) To explain: I work for a large power semiconductor manufacturer. Here we are pretty much split into two camps—the modellers and the empiricists—with myself firmly in the latter camp.

(Okay, then I tend to be on your side. Mostly. /rap)

I increasingly find myself in the frankly bizarre position of having to defend my bench results against Spice simulations that disagree with my results. This is "Alice in Wonderland" to my mind. ('Zackly!! /rap)

Sure, you can criticize my methods. But if you're happy with my methods, then the results I produce must be "The Truth," right? If those results differ from the results of simulations, then it's the simulations that are wrong, not reality! (In most cases, I agree with you. But, these examples are in the early design stages. Surely, when you get closer to final designs, the breadboards will indicate who is right—right? /rap)

If I look at the people who are in the two camps, I would say that the empiricists, without exception, come from practical "hands-on" backgrounds. Typically, they are people who have been building circuits since they were kids and whose first approach to a new idea is to go build it. The modellers, on the other hand, tend to come from purely academic backgrounds. And how come I rarely (if ever) see any of those guys in the lab? Hmmm... Anyway, please continue to fight the good fight. (Absolutely! /rap)

Name Withheld (via e-mail)
Pease: The proof of the pudding must soon be in the eating. And the proponents of simulation may soon be eating crow—at least in most cases. I will concede that in rare cases simulation does tell the truth better than breadboarding. Also, we will certainly agree that breadboards do lie in some cases. So we must be wary of situations where the breadboards could be wrong!

Dear Bob: You can never preach enough about the dangers of Spice—or any simulator tool. As one who went from two-way radio tech to EE in the wireless sector, I laugh at the number of engineers who bet their paycheck on software! Nothing beats engineering with all senses: touch (feels hot), sight (resistor red hot), sound (snap of the IC blowing up), and—of course—smell (of the burning pc board). And when I go back and see where I went wrong, I learn more and retain more. Let's hear more and maybe even management may come to understand that.

David Donaldson (via e-mail)
Pease: Hello, David. We sure agree a lot!

Dear Bob: "Why be difficult when, with a little bit of effort, you can be darn near impossible!" Is this the latest industry mantra? Refer to the pinout of the SOT23-5 version of the LMV2001 and the LMC7101: pin 1: output, pin 3: noninverting input, pin 4: inverting input. Okay so far. But, pin 2: negative supply for the LMV2001 and positive supply for the LMC7101! (THAT is the paddle. /rap) It's the same with pin 5: positive supply for the LMV2001 and negative supply for the LMC7101. Is there a coherent reason for this? I could almost believe it for a very high-frequency item. But these are both definitely not gigahertz-bandwidth devices. Nor are they from different manufacturers where you may expect the "we must be different" attitude. It's hard enough trying to keep up with the wide variety of new offerings without having this sort of nonsense to cope with. Comments please!

Peter Pohl (via e-mail)
Pease: In the SOT23-5, the die has to sit on the paddle, and the substrate of the die has to go to the middle pin of the three. Thus, all CMOS devices need the middle pin of three to go to +VS. And on bipolar amplifier chips, the −VS has to go to the middle pin of three. I am sorry as heck. But if you want to go argue with silicon, and with copper, count me out of the argument because I cannot help on this.

Comments invited! rap@galaxy.nsc.com —or:

Mail Stop D2597A, National Semiconductor
P.O. Box 58090, Santa Clara, CA 95052-8090

*Electronic Design, Sept. 1, p. 24.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish