How to become a Hardware Architect(system engineer)?

I'm a electronic engineer who has three years experience on hardware design. So far, my work is draw the schematics follow the Technical Director's architecture. Nowaday, I hope I have more initiative in my work and I want to become a hardware architect.

What should I do? And any book helpful?

Thanks in advance! Shawn

Reply to
Electronic Elephant
Loading thread data ...

Start designing and building your own stuff in your own time. Come up with your own ideas and implement them. Show them what you can do, and if they don't let you step up then move somewhere that will.

Dave.

Reply to
David L. Jones

ge

Thank you Dave! Thanks for your kindly reply.

Reply to
Electronic Elephant

Andy adds: Well, I'm a retired communications sytem engineer, and I can tell you for a fact that David's advice is right on target....

If you are in RF or communications, start building radios. Maybe get your ham license, your FCC license (GMDSS,GROL, etc), do some home data link projects that you can talk about or show off.

If you are in digital signal processing, start designing some one- chip machines using stuff like the PIC series --- things that actually do something that you can sit on your desk and people will ask you about.

As long as you are used as a clerical tecknician, nobody will EVER suppose that you can do diddly-squat on your own. Regardless of your degree(s) or how well you make view foils....... The path that worked for me was GETTING MY HANDS DIRTY....

Heck, my home workshop was better than the ones I had at Bendix, Texas Instruments, or Raytheon......

Obviously, tho, David and I are "old school" ---- guys who stole drive-in speakers while our buddies were all stealing hub caps --- and we grew up building our own toys.....

I don't know if that approach still works for everyone, but I don't see why not. It was certainly the profile I was looking for when I hired engineers to work on some of my own teams.

Andy in Eureka, Texas aka ( W4OAH - for 45 years)

Reply to
AndyS

The "your own ideas" part needs a little stressing here. You need to fully think through some design ideas. Start with something simple. You don't need to implement all or even many of your ideas, only the good ones. You need to think about ideas and realize that they are bad ideas. To be good you have to both be able to come up with ideas and be able to throw away the bad ones.

Reply to
MooseFET

Learn software. Implement something. Buy a PIC development kit and put something useful together, like a lawn sprinkler controller with a web-addressable front end.

Or, a middle course - lear VHDL. Buy or borrow a development kit and build the above-mentioned controller in a FPGA.

Reply to
Richard Henry

--
When you get an assignment from your boss, figure out a
different/better way to do it on your own time.
 Click to see the full signature
Reply to
John Fields

Excellent advice from everybody else

Learn to do a bit of microprocessor work. I'm just learning the Atmel AVR series. Their Studio 4 software works quite well, and you can do it with C if you download the WinAVR/ GCC software. All free, and the chips are cheap as well. Another option would be the 8052, but it's getting a bit long in the tooth. I won't mention PICs

martin

Reply to
Martin Griffith

What everyone else said, plus:

Think about what a systems (or hardware) architect does, and aim yourself in that direction.

The hardware architect needs to identify all of the problems that the hardware must solve, and be able to either propose solutions to these problems or (more importantly) be able to recognize when he/she just can't solve the problems. Identifying the insoluble problem means that you can push the problem back into the larger system architecture arena, or out of the product altogether, or get someone's attention to help you out with a solution that didn't occur to you.

Then the hardware architect needs to be able to solve those problems, or find folks who can and state the problem and the proposed solution well enough that _they_ can solve the problem, in a way that fits into the overall architecture.

The _worst_ designers are the ones that promise solutions, then string their colleagues and managers along for months, then fail to solve anything. That sort of behavior puts projects into serious trouble, or kills them altogether.

You need to be able to view the whole system (I'm assuming it goes beyond mere hardware), understand where the hardware fits into the puzzle, understand how _all_ the hardware is going to work together, how it's going to work with the rest of the system, and know who does what. You need to be able to talk to and understand the software and mechanical engineers on your team (even if you're just putting a rectangular board into a box there'll still be heat and mechanical mounting issues that will constrain the electronics design). If you're doing some specialized system such as a sensor or controller you'll have to understand, at least at the "Scientific American" level what it is the whole system is doing in the outside world.

To move into such a job you need to:

  • Speak up. Quiet Mark, who design perfect schematics but never questions them won't get promoted to hardware architect, ever. He may keep his job until he's 55 and gets laid off, but he won't move up very fast. Guys who show an understanding of the whole system and how their stuff works in it have a chance, guys who catch system-wide errors just from looking at the inputs and outputs of their parts have a better chance. Guys who _don't_ design perfect schematics the first time every time, but _do_ have a positive impact on the system design have a far better chance than guys who do quiet perfection perfectly.

  • Speak up. Blow your own horn. Better yet, notice who's good and play duets. "I figured out that there was something wrong in the frogrozzle circuit from looking at the data coming from the wingplaster, but Quiet Mark over there was the one who figured out it was the glimplip regulator that was the real problem". That gets you credit from higher ups for being able to work with Quiet Mark _and_ for instigating a fix. It'll probably earn you some gratitude from Quiet Mark, too.

  • Be right. "Speak up" is right next to "stick your neck out" in the dictionary (I have a small dictionary). It's best to not stick your neck out and then make the ax fall, at least not all the time. If your work environment is sane you can speak up, be wrong, and not get axed _immediately_, but you have to show consistent wins if you want to move up.

  • Expect tumult. Architects aren't always the most popular person on the planet. They're the ones saying 'no' to schedules, or whose stuff messes up the schedule down the road when it doesn't work, they're the ones demanding high $$ ADC converters when the product manager wants to use the "12-bit going on 8" converter on the * Expect tumult. Architects aren't always the most popular person on the planet. They're the ones saying 'no' to schedules, or whose stuff messes up the schedule down the road when it doesn't work, they're the ones demanding high $$ ADC converters when the product manager wants to use the "12-bit going on 8" converter on the $0.20 micro that he proposed even though he's never designed a working circuit in his life. If you can't stand being a lightning rod, reconsider your desires.

    .20 micro that he proposed even though he's never designed a working circuit in his life. If you can't stand being a lightning rod, reconsider your desires.

  • Asses your chances at your current gig. If your company, division, or just your manager, are too small to let you advance without some 45 year old dying from old age, consider finding a job at a different company (this advise may not apply in countries where job mobility isn't as encouraged as the US -- use your judgment). If you work for a five-man engineering group, you and Quiet Mark will _always_ be juniors, until you retire or unless the company undergoes some serious growth.
--
Tim Wescott
Wescott Design Services
 Click to see the full signature
Reply to
Tim Wescott

Oh, and of course hope that your stuff does NOT work first go. You'll learn a *lot* more if your circuit/system does not work first go and you have to troubleshoot it and figure out the what, why, and how.

Dave.

Reply to
David L. Jones

I miss the days when I designed and built circuits that didn't work. :P

Now my projects blow up in software. Wow! 4000 amps going through a

1/4 watt resistor and it doesn't burn a hole in the screen. :)

D from BC British Columbia Canada

Reply to
D from BC
[snip]

I hold the International Fry Baby Award... earned during my youth ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Hey, anybody ever bring their own audio amp to the drive-in and try to connect it to the speaker and backfeed audio into all the other cars?

Mark

Reply to
Mark

"Back then" we were self-sufficient and able to skin leopards without killing them ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

But did you have leopard skin upholstery on your first car?

--
Keith
Reply to
krw

I never cared for those. Yuckee!

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

t
d
)

! Cavemen

=A0 =A0 ...Jim Thompson

=A0 | =A0 =A0mens =A0 =A0 |

| =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

uoted text -

Thank you all for your kindly and elaborate reply. Thanks Dave,Andy,MooseFET,Richard,John,Martinm,Tim Wescott,D from BC,Jim. You really inspire me a lot.

Reply to
Electronic Elephant

Wow, a satisfied customer!, and a Google Grouper to boot.

Dave.

Reply to
David L. Jones

I know. You wrote the book: '101 ways to skin a live leopard, and live'. ;-)

--
http://improve-usenet.org/index.html

If you have broadband, your ISP may have a NNTP news server included in
 Click to see the full signature
Reply to
Michael A. Terrell

You want to become a Design Engineer. The hardware/software tradeoff changes daily. You have to be able to do both. You can get there thru hardware, but you gotta eventually understand both.

The first thing a design engineer does is wear the customer's shoes. If you're designing test equipment to be used by real engineers with advanced degrees, you may make different tradeoffs than if your customer is a Phone technician who worked at WalMart last week. The hard part is understanding that it's necessary to interpret the answer and sell that to management.

You have to wear the builder's shoes. Where/how is the thing being built. Something as trivial as using the same size screw for everything in the subassembly can save you a lot of grief. Design the mistakes OUT of it.

Then try on everybody else's shoes: management, engineers, technicians...know what everybody does and why. You have to be curious about everything.

There are some critical things that you have to have and others that you have to find.

You MUST have curiosity and fearlessness. If you took apart your first tricycle to see what made it work, you might be an engineer. Doesn't matter that you lost the bearings and couldn't get it reassembled. What matters is that you hung in there and took apart your SECOND tricycle...and your car...and your TV. Doesn't matter what you messed with, but it's critical that you messed with something. That got to be my first interview question, "what did you take apart lately?" An expert mathematician who can't fix his microwave oven is just a mathematician. We need 'em, but we don't let them architect systems.

If you spent your youth messing around with stuff, it's possible that your education "took". I found I could teach a person electronics. I couldn't teach curiosity or inguenity. But I sure as hell could select for it.

The next thing you need is a mentor. A decent mentor is VERY difficult to find. Mentors have to be selective in where they put their effort. You gotta stand out. When you find one, bake 'em cookies and stick close. Make their life easier. Make them proud. They will make the difference between a mediocre engineer and a stellar one, assuming you've got the right stuff.

There are lots of people who know HOW to do stuff. Many fewer know WHAT to do. Even fewer know how to get the how people to do the right what.

All that electroinc crap falls into place if you've got curiosity and a good mentor.

Reply to
mike

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.