Which university produces good analog EEs?

Well, if the sky is the limit in terms of cost ...

Ok, Chinese 4-bit micros exempted. Those can be had under 10c. But the programmers should master Mandarin.

I don't use them, I use 74HC14 or CD40106 sections. Even as PWM chips.

What's the most expensive FPGA you buy?

It's fun, now and then, to design for flat-out performance, and damn the cost.

What? Another case of asynchrophobia?

John

Seems every year they come out with a new fat cat FPGA that almost has to be delivered with a Brinks truck. Plus a 2nd mortgage for the new design suite because, of course, the last release from eight months ago is now totally outdated and obsolete.

I get to do that about once per decade. But for me the real fun begins when the client says "Nice. Now, can you do this for under seven bucks fifty?"

ROFL!

No trimpots in my designs. At least not in the last 20+ years. Watch crystals and resonators are dirt cheap and the 4060 is, too. For faster stuff I often use NTSC color carrier crystals but I guess that party is going to be over when the US goes digital. Not looking forward to that. Darn, I though it'd last until I retire.

It makes me sick to see a trimpot used where one could use a cheap micro with a *calibrated* RC oscillator on-chip and get 2-3% guaranteed accuracy over temperature. The extra gates are just to make this on-chip *analog* oscillator do what you want. ;-)

Best regards, Spehro Pefhany

Most of my jobs have been that way[*]. Thinking back, I can't remember one where the parts cost was in the top three engineering problems (schedule was usually #1). The engineering costs always swamped the production costs (or were pretty much fixed, as in the case of the processors). Thinking harder, I take that back, there was one project where cost was a huge driver (a series of encrypting LAN adapters), but that project was canceled due to lack of interest.

Manderin in 4-bits? ;-)

Did that once. My copy of the databook for a 87C51 didn't have the paragraph saying that an interrupt had to be asserted, then deasserted to bring the processor out of sleep. I had no room left on the card and a couple of spare sections of 74HCXX stuff laying around.

The most expensive I personally spec'ed was an SCX600E in, IIRC '99, while they the Virtex-E family were still engineering prototypes. They were $1200ea. As expensive as those things were, I had far more invested in software to program them than the chips themselves. Performance was an issue on that project.

I'm pretty sure the things I'm working on now are quite a bit more than that; 10x the size, twice the pins, and rated for UncleS.

Often I've had projects where performance wasn't even the real issue. Like I said, cost rarely, if ever, was.

More acousticophobia, but I'm not a NASCAR fan either. They go off in weeds. I don't like it when my designs do. It usually causes long weeks (though I've been pulling 60hr weeks for a month or so).

Brace yourself for 70 hours or more if the universities don't start cranking out engineers with more useful hardware knowledge.

Nah, I'm "retired". I picked up a contract job because I got bored, this sounded like fun, and they pay well. ;-) I just got extended until May, so I'll have something to keep me busy for the winter. ...and I'm out of the most leftist-weenie state in the union (my house sold Monday)! :-)

Congrats! Did you decide on TN or KY?

...Jim Thompson

Same

just

Thanks. Haven't decided. I've been in NE OH (well, OH is on the way) since early August. My wife will come out when the house closes next month. We're still trying to get things firmed up.