[Coco] Something of interest to the design people here
Gene Heskett
gheskett at wdtv.com
Sat Oct 25 18:51:58 EDT 2014
On Saturday 25 October 2014 06:32:46 Mark McDougall did opine
And Gene did reply:
> On 25/10/2014 9:00 PM, Gene Heskett wrote:
> > TV engineering is a funny business, and I have in my time designed
> > and built several very unique things to do a labor saving job at a
> > tv station. Unfortunately, the market for such an individual device
> > is also limited to other tv stations facing similar technical
> > problems, which is a roundabout way of saying that to sell even 100
> > such devices is a resounding success.
>
> I'm in a very similar position myself. I have a 'contract R&D
> engineering' company and we often design and develop prototypes for
> customers. That could mean as little as a single device. What's more,
> as far as the 'estimated production quantity per annum' that vendors
> and distributors are so keen to extract from you are concerned, we're
> completely at the behest of the client. If we're designing a
> specialised or niche product, production quantities could be measured
> in tens per annum. With these numbers it's very difficult to get any
> type of design support.
>
> Fortunately for devices that aren't bleeding-edge, Digikey and Element
> 14 usually come to the rescue. Where we're stuck is when we want to
> use a device that is new to the market. So generally, no samples, no
> inclusion in our design. Of course such small volumes means nothing to
> the vendors...
>
> ...but then there's the scenario where I want to use a bleeding-edge
> device in a prototype for a client in a niche market and a do get
> samples. No pay day for the vendor then, but what if our next client
> comes along with a commodity market design and we have the opportunity
> to use the same - tried and tested by us - device? They need to look
> at the bigger picture.
>
> I was doing a small design years ago and was looking at using a
> particular analogue video chip that looked like it was exactly what I
> needed. Not only could I not get samples, I couldn't even get the data
> sheet for it, unless I was projecting to ship hundreds of thousands of
> units.
>
> It's an unfortunate trend that I fear will only continue.
>
> Regards,
So do I. To use a furinstance, the now ancient Grass Valley Production
Video Switcher, a 3 bus, 24 input device, when it was new equaled about
2/3rds of a 7 foot rack to hold it, and the NEC digital effects package
doubled the price and took up most of a 2nd 7 foot rack.
When GVG designed that thing, there were no truly video speed op-amps, so
they designed their own in discreet transistors that just barely had
enough bandwidth to handle NTSC video.
When they started failing, we were well into the far end of the infamous
bathtub curve, and well past the 5 years since they built the last one. I
won't say support wasn't, but what they had on the shelf as priced out at
$1700 a copy & no guarantee they were still good.
So I started researching video speed op-amps, and finally settled on one
in a to5 can for $1.37 each, with a gain/bandwidth product north of a
gigahertz!
No samples were contemplated since the thing used 72 of them. I bought a
5 pack, had a heck of a time making the legs fit in old circuit boards
inline hole pattern in the module, and wondered if I'd have to kill some
of its hf gain with ferrite beads. But it worked, drawing 10% of the
power their OEM board drew. Just one problem though.
This thing was delay matched, and the channel it was in was so much faster
the color was shifted CCW about 15 degrees, well beyond the ability if the
circuits trimmers to match. Since pulling one module to work on it and
doing the work would have put production out of business, I had two
choices, 1 let it be, 2 go on the graveyard shift for about 2 weeks (after
ordering and receiving enough of them to do the whole thing) and do them
all.
Since at the time I was the sum total of the engineering staff, the 100
bucks for the parts, and me being out of service days, the then current GM
said no.
6 months later another one failed, and running down which one and
replacing it took most of a day, with production at first, and the GM
next, haunting me all day, distracting my train of thought and just that
probably cost me half a day. I finally blew up and told the production
manager that the quickest way to do it was to keep Mike the GM, the hell
out of my hair. Worked for about an hour. But that was enough to identify
the guilty party and get started on its replacement. Took about half the
time that time.
Not too long after that, the annual visit by corporate the accountant
found something not exactly kosher & Mike was given 5 minutes to clean his
stuff out of the desk, he was going to be replaced. Forcibly.
I found out years later that he had called the owner 3 times wanting to
fire me for insubordination, but the owner told him "not no but hell no".
Thats a warm fuzzy feeling at 11,000 feet over Lake Michigan when I was
told that. In the owners airplane, a twin Cessna. :)
Then I retired, and the next failure forced its replacement, which in turn
had way more than its share of infant mortality at $1500 a replacement
board.
Cheers, Gene Heskett
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS
More information about the Coco
mailing list