Design Con 2015
Breaking News
Engineer’s Bookshelf

Have You Ever Been Blindsided by Your Own Design?

NO RATINGS
View Comments: Newest First | Oldest First | Threaded View
Page 1 / 4   >   >>
Max The Magnificent
User Rank
Blogger
Re: Happens all the time
Max The Magnificent   12/17/2014 1:10:25 PM
NO RATINGS
@GSKrasle: Waxing the lipstick on a pig?

Pucker up! :-)

GSKrasle
User Rank
CEO
Re: Happens all the time
GSKrasle   12/17/2014 12:37:00 PM
NO RATINGS
ante,

"it was my own document formatted and looking like a professional publication"

Waxing the lipstick on a pig?

dvhw
User Rank
Rookie
"official data"
dvhw   12/17/2014 10:48:20 AM
NO RATINGS
About 20 years ago, during a talk I gave in Japan, I needed to comment (in passing) on market share of embedded OS vendors (back then VxWorks, pSOS, Lynx, QNX etc).  There wasn't a lot of good market data but it was pretty clear who the leaders were approximately how big they were; I also had a pretty good idea from what our various customers were using.

So I drew up a pie chart on foils (remember those?  Called OHPs in Japan).  I remember saying something like "The data aren't great but the market looks approximately like this."

So imagine my astonishment when at the next Embedded Systems Conference I saw professionalized versions my hand drawn chart on several vendor booths! 

I was already skeptical of marketing claims but that pretty much put an end to any residual faith I might have had in them.

antedeluvian
User Rank
Blogger
Re: I named an oscillator
antedeluvian   12/16/2014 4:25:24 PM
NO RATINGS
traneus

A year or few later, I got a copy of a new Harris Application Note for the HSP50016. In the App Note, Harris called that oscillator the "Weaver oscillator".

And I'll bet no one at Intersil knows why- they probably think that everybody called it that through the years.

 

traneus
User Rank
Rookie
I named an oscillator
traneus   12/16/2014 3:45:17 PM
NO RATINGS
In 1994, I wrote an article about using the then-new Harris (now Intersil) HSP50016 digital downconverter (DDC) in a shortwave radio receiver, and QEX published the article. In my block diagram of the innards of the DDC, I named one oscillator the "Weaver oscillator", after the inventor of the analog form of the architecture.

A year or few later, I got a copy of a new Harris Application Note for the HSP50016. In the App Note, Harris called that oscillator the "Weaver oscillator".

 

GSKrasle
User Rank
CEO
Re: Happens all the time
GSKrasle   12/16/2014 3:33:16 PM
NO RATINGS
Beware the boomerang of forms and procedures!

We have all dealt with ones annoying in demanding lots of information that's not to-hand, or just unreasonable.

I have seen how that happens. I made a simple tracking database of proto-systems, and included fields for 'location', 'user', 'return date', 'FW rev' ,etc., not expecting a thorough and complete record, but just opportunistic rough ideas of where what and when. When I turned this over to ops, suddenly a company-wide inventory had to be performed daily, and every developer had to create a detailed and accurate Project Schedule so he could enter the correct date for the system to be returned, and every time a new FW rev was tried, it had to be logged and updated..., All these reports came flooding-in although I just didn't care about that level of detail. There is no value added by onerous completeness.

Likewise with testing: if I want performance data logged at -10, 20, 40C +/-5, it's not necessary to hit -10.00°C, 20.00°C, 40.00°C, especially if it's going to take all day to achieve that precision. If I need extraordinary precision, I'll be explicit about it! Correct procedure is to get close-enough, and log how close, along with the results.

In both those cases, I try to make the instructions clear that there is no expectation of extraordinary effort, but it's hard; there is a definite tendency when playing 'telephone' for the instructions to become more elaborate and difficult at each step.

GSKrasle
User Rank
CEO
Re: Happens all the time
GSKrasle   12/16/2014 3:23:16 PM
NO RATINGS
Then there's the opposite: On a project, we found that a supplier's sensor with integrated processor had some serious functional flaws. For other reasons, we were committed to actually incorporating an additional processor into our design (the sensor chip was intended to be 'standalone'), so we were able to 'fix' the behaviour. But we never told the sensor's manufacturer, lest THEIR fix aid our competitors who might use the same part....

GSKrasle
User Rank
CEO
Re: Happens all the time
GSKrasle   12/16/2014 3:15:56 PM
NO RATINGS
zee,

That's interesting. I suppose there's nothing inherently WRONG with a manufacturer incorporating ideas from their customers, it implies a healthy lack of haughtiness on the part of their engineers, but a little generosity, at least of gratitude, would be noble.

I found bugs in the recently-acquired 'latest greatest' expensive new MSO models from the leading two manufacturers. They sent their engineers over (separately, but oh, how funny it COULD have been!), and I received firmware updates promptly, but I should have asked for thank-you certificates (or swag?)

This is sort of similar: Once, there was a problem with clock-distribution in a partner's prototype. On a con-call, they were distraught, so I suggested that there was surely a solution; 'maybe you can simply program both the codec and processor as slaves, add a [xx]MHz crystal to the codec chip, take its clock-out, put it through two flip-flops and feed it back into both chips to get the right frequency....'

When the next protos appeared, they had done EXACTLY that, in the most literal way possible: there were a crystal and two 74LVC1G74 singles wedged in between the codec and (mostly unused) CPLD. It worked, and I'm glad, but what if I had been wrong? It looked like they were taking me as the Voice of God or something! If they had thought it through, they didn't need those added chips at all, cheap as they are.

 

 

,  

zeeglen
User Rank
Blogger
Re: Happens all the time
zeeglen   12/16/2014 2:08:33 PM
NO RATINGS
@Antedeluvian A few weeks later I was given a document, laser printed and formatted nicely.

Back in the '80's I designed a fiber optic receiver board for Ethernet using an integrated connector/photodiode/transimpedance device in an all-plastic package.  The external optical connector was an-all metal SMA 906 type and when mated the business end was very close to the photodiode.  Thus any EMI picked up by the ungrounded metal connector capacitively coupled to the photodiode and clobbered the signal.

The EMI came from a TTL ribbon cable running adjacent to the receiver (not MY idea) and even though I had shielded the electronics the metal optical connector was still a problem.  I kludged in a spring that would compress from the connector and ground it.  No more EMI problem.

After discussing this with the well-known manufacturer, a few months later I was reading their latest data sheet and saw that my rough spring drawing had been cleaned up and was in their data sheet!  Later that manufacturer developed conductive plastic and metal port options for this problem.

A few years later I designed another grounding device using a flat fingered berylium copper strip for less inductance than a spring.  It worked fine when tested to the FOIRL spec.  The only problem was that a few months after a bunch of us got laid off, my former employer went to a SMA connector vendor that used different nut dimensions and the fiber cables would no longer fit.  When they called me to ask what to do, I could only say "Sorry - not my problem any more."

mhrackin
User Rank
CEO
Reverse twist?
mhrackin   12/16/2014 1:11:35 PM
NO RATINGS
I've been on both sides of this issue. I often found it politically expedient to add my boss(es) as co-inventors on my disclosures, especially at one time when my boss's boss was a member of the Patent Committee that decided what to go ahead and file! Later in my career, as the chief engineer/system architect I would do the top-level detailed design, and then turn over to my team to implement. When they came up with patent-worthy designs, i encouraged them to submit the disclosures as the sole inventor, even when they wanted me to share credit. In their own careers, they got well over 100 patents total, but their first ones were granted when they worked for me.

Page 1 / 4   >   >>
Flash Poll
Like Us on Facebook

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Times on Twitter
EE Times Twitter Feed
Top Comments of the Week