interviews

No, we put parts on PC boards. We do FPGA design, but FPGA designers are relatively easy to find. It's harder to find people who can do "real" electronic design: architectures, analog stuff, power supplies, things like that.

Whatever it is exactly, here is the question:

The short answer would be "a year or two" I guess. Having a BSEE or a physics degree would help.

I don't know what you mean by that.

We interviewed one EE today. His resume mentions designing a 9V to 30V boost converter. He knows *nothing* about boost converters. Not the basic topology. Not that they usually include an inductor.

We don't see any competition from Asia. There don't seem to be small instrumentation companies there. They mostly go for volume consumer products. We do have some competitors in the UK and europe.

around here.

My impression is that Chinese and Indian engineers tend to follow rules, which is bad for inventing things.

Yes truly stunning. Over the last couple of years, I have interviewed 50 people for an analoc ic design job. Most had impressive paper resumes, all but 2 were, essentially, useless. Simple could not do anything they clamed or understand the basics. This was with typically 10 years+ experience, with all the right companies. What is also staggering is that some of these got hired by these companies afterwards.

I tell you what though, around 2000, my Current Mode Controller SuperSpice example model took several minutes to run. It now only takes 7 secs on my

Kevin Aylward B.Sc.

I guess you missed the point of that answer.

Kevin Aylward B.Sc.

Krw misses a lot of points. If an idea doesn't match up with his simple and inflexible world view, he can be relied on to reject it.

That you wanted to be ridiculous? That's the only excuse.

That is the only excuse that krw can imagine. The slightly more complicated idea that while you may have to pay more for experienced staff, the fact t hat they produce better output faster will reduce your overall costs, is wa y too complicated for whatever it is that krw uses instead of a mind.

Not good enough. I want about 10,000x the compute power that I have now. I want to be able to attach sliders to part values, slew them around, and see the waveforms change instantly. I *can* do that on a breadboard.

If this person is approaching mid-20s, don't you think he would have at least dabbled in hobby electronics by now, as a minimum, if he was truly interested in "real" electronics?

Might as well hire an orangutan...

Actually, you can do that in SuperSpice, change the values in real time. The one example actually engages an option to slow down the simulation!

Kevin Aylward B.Sc.

David Beckham was kicking balls about when he was 5 years old. Name one world class soccer player that started at 23?

Or name we any world class, ballerina or ice skater that started at 23 ?

Yeah, I was building and modifying kit when I was 11... Connecting vibrator transformers up to door handles and such like....

Kevin Aylward B.Sc.

There is a great Netflix slide show on this:

Slide 36 - "...the best are 10X better than average..." Slide 96 "...one outstanding employee gets more done and costs less than twice two adequate employees..."

ergo... (the right) experience is always cheaper.

Kevin Aylward B.Sc.

On a sunny day (Wed, 16 Apr 2014 19:06:10 +0100) it happened "Kevin Aylward" wrote in :

You were evil then, and now with slimulations! ;-)

There's something to be said for getting those neurons wired during the formative years.

Multisim also has "real time simulation", which means you get to diddle switches and pots while it's running. Of course, most circuits rarely go fast enough to call it "real time", by six orders of magnitude or so.

I don't know of any standard SPICE package that actually goes much faster than any other (LTSpice -- when its proprietary solutions are actually working -- being a notable exception on multicore machines), unless the high end (Super-, you might say..?) special ones might be doing cool stuff. I've read about partial matrix solutions before (i.e., evolving the circuit by local blocks at different timesteps), but never seen a package -- at least a cheap one -- that claims to be anything beyond XSPICE, or a derivative thereof.

Tim

That person is almost always more expensive. There is no assurance that anyone you hire will be 2x the productivity of any other.

So buy an RTDS or make your own. That should work for some subset of what you want to do.

complicated

fact

is

Thank you for the link Kevin.

?-D

Hmmmmmmm - There is a slight inconsistency in your job requirement - You say you'd settle for a guy who knows how an op amp works, and then seem to want more extensive analog design skills. ISTM you want the latter, but are so despairing of finding it, you will settle for anyone who is not a complete idiot, to put words in your mouth. I was hiring sw programmers to do OCR a few years ago in the bay area - I eventually had to come up with a test - likely candidates were only 1% of people i talked to (interviewed). A test might help you cull through the 300 people you may need to interview. If the 1% ratio is correct, you don't want to be in the loop.

Here is a question about the resources needed to find this person: How about the idea of spending $10,000 to find him/her? Is that worth it? I'm guessing it is, since an inadequate employee can suck up your capital with salary, not to mention frustration and time cost.

If you agree with that line of thinking, then you might be willing to try odd (and expensive) ways of finding that employee. How about a geographically- defined Google pop-up ad for those searching on circuits? Or, a $1000 bonus for a teacher who produces a student who knows this area? Circuit Cellar, Nuts'n'Volts? Have you maxed out your job listing possibilities?

Part of this campaign may be to create a company "image," as an exciting place to work in a company doing cutting edge stuff. That may require some thought.

Bear in mind, I'm brainstorming, and the operative value there is 100 crappy ideas may have one good one!