Sorry, Yogi

John Larkin, as out of touch with reality as ever - it's all connectors, and always has been.

You got to screw up royally with the semiconductors and the passive components to create the same kind of disaster that a single ill-chosen connector can turn on.

--
Bill Sloman, Sydney

Blimey Bill, you must have fell out of the wrong side of bed today!

,

This is a reaction to John Larkin, who wastes our time with links to idioti c denialist rants about climate change from well-known denialist sources.

Even if I'd gotten out of the right side of the marital bed this morning (a s in fact I did) I'd still be feeling peevish with him - not to mention tha t connectors are 90% of the packaging problem, and tend to define the rest of the package once you've found connector solutions that might work.

--
Bill Sloman, Sydney

you're just jealous, admit it!

At least he can produce what comes out. You on the other hand simply pollute and create S/N ratio exceeding tolerable levels.

Jamie

her

tors,

osen

iotic denialist rants about climate change from well-known denialist source s.

g (as in fact I did) I'd still be feeling peevish with him - not to mention that connectors are 90% of the packaging problem, and tend to define the r est of the package once you've found connector solutions that might work.

He's got himself a nice little racket, but producing a new design roughly e very two weeks isn't exactly stretching the design space.

If I was going to jealous of anybody, it would be of Jeorg - but he does st uff that's rather beyond my reach.

Or so Jamie thinks. Anything complicated - like 2+2=4 - goes into his too hard basket, and he sustains his self-image by not paying any attention to it at all.

--
Bill Sloman, Sydney

her

tors,

osen

iotic denialist rants about climate change from well-known denialist source s.

g (as in fact I did) I'd still be feeling peevish with him - not to mention that connectors are 90% of the packaging problem, and tend to define the r est of the package once you've found connector solutions that might work.

He's got himself a nice little racket, but producing a new design roughly e very two weeks isn't exactly stretching the design space.

If I was going to jealous of anybody, it would be of Jeorg - but he does st uff that's rather beyond my reach.

Or so Jamie thinks. Anything complicated - like 2+2=4 - goes into his too hard basket, and he sustains his self-image by not paying any attention to it at all.

--
Bill Sloman, Sydney

John, please go somewhere to post your off-topic BS.

Sincerely

Haven't you learned there is *no* off topic in this group? Except if you post something electronic that isn't along the line of analog design. Some here have criticized people for posting digital electronics. lol

--

Rick

Electronics is off-topic here now?

You'd rather talk about great-grandkids and flu shots and Firefox?

Besides, this is somewhere.

Sorry, I should have been more positive. John, I enjoy your on-topic comments very much.

Well, there is some overhead of off-topic stuff; that's the way people are.

We've recently seen a pretty hefty amount of bad electronics packaging. I think bigger companies have somebody draw a schematic and toss it over the wall to another group that does PC layout, and then to some other organization that does mechanical design.

I designed a waveform digitizer box to acquire optical waveforms in a biggish laser. 12 bits at 250 MHz, so it dissipates a few watts. The customer's MEs, in spite of my protests, packed six of them on thin metal brackets in a 3x2 array. The two in the middle have no cooling. The enclosures are running almost 60C case temperature at normal ambient. The internal temps are over the rated parts specs. There is no margin for operation above room temp.

So, they had a bunch of meetings and told us to fix it. The fix is to install this:

Even funnier, the "fix" is invisible when the boxes are installed.

Maybe you can sell the extras to McDonalds for their coffee cups.

But aren't connectors part of packaging?

I once designed a circuit that would measure the power in a high power shor t duration pulse. It really was the domain of the test equipment engineers , but I was directed to do this. I was told that otherwise the test equipm ent guys would say it could not be done.

So I designed it with a load for the pulse, then a shield with the signal g oing thru the shield on a coax cap. Then an integrater and another shield and feed thru cap to the level detection circuit. And it worked. So gav e a sketch to the test equipment department.

A month or so later I got asked to figure out why the box the test equipmen t guys had made did not work. It turned out that they had carefully shield ed the integrater , but had the load right next to the level detection circ uit.

Dan

To do good hardware/software/packaging/thermal design, people have to do it together. Things don't compartmentalize well. Physical presense helps a lot.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

hort duration pulse. It really was the domain of the test equipment engine ers, but I was directed to do this. I was told that otherwise the test equ ipment guys would say it could not be done.

l going thru the shield on a coax cap. Then an integrater and another shi eld and feed thru cap to the level detection circuit. And it worked. So gave a sketch to the test equipment department.

ment guys had made did not work. It turned out that they had carefully shi elded the integrater , but had the load right next to the level detection c ircuit.

I've got this instrument/electronics redesign on my horizon. I don't want to change all that much circuit wise. (It's gotta be backwards compatible.) Goals are

1.) Reduce power loss, heat dissipation. I've got several current sources driving ~1-2 ohm coils at 1 amp max, from too high a voltage. 2.) Get rid of the fan! and put all the "class A" pass elements and other power on the back panel, with the right sized heat sink. (I will have to do a worse case thermal budget.) I've got this image of all the power coming into and out of the back panel, front panel is control. 3.) simplify wiring. There is this mid pcb, (in my mind), at the moment laying along the bottom of the box. Mostly it will be doing inter connect, but there will have to be some electronics there too.

At the moment, I'm thinking about how to do the inter connects between the various boards. Given tolerances, I can do hard connections at the front end, (I think, or would you use wire at both ends?) With ribb on cable (or something else?) to the back panel.

The only circuit thing I might do is make a better thermal couple temperature controller. T-type, copper and constanan. (small enough and it's not very magnetic.) I've not done made one before, so it's a pretty sure bet, that I'd screw up, have fun, and learn something.

George H. (to prove your point)

,

Since they are the interface to the outside world, they could be seen as th e only bit of the packaging that does anything electronic.

But in my dialect, the packaging is what provides mechanical and electrical shielding and support for the device, and - incidentally - accommodates th e connectors.

First pick your connectors, then chose the packaging that works with them. If none does, pick a new connector.

There are exceptions to this approach. Jeorg's amazing ultrasound scanner t hat fits into the tip of a catheter and gets shoved up inside arteries, sca nning the plaque on the arterial walls as it goes, is an exception, but it was exceptional in all kinds of impressive ways.

John Larkin was never going to be talking about something like that. Most l ikely John doesn't rate connectors because his customers tell him what they want used.

--
Bill Sloman, Sydney

nd

short duration pulse. It really was the domain of the test equipment engi neers, but I was directed to do this. I was told that otherwise the test e quipment guys would say it could not be done.

nal going thru the shield on a coax cap. Then an integrater and another s hield and feed thru cap to the level detection circuit. And it worked. S o gave a sketch to the test equipment department.

ipment guys had made did not work. It turned out that they had carefully s hielded the integrater , but had the load right next to the level detection circuit.

m

bbon cable (or something else?) to the back panel.

Think DIN-41612 mixed signal connectors. I've used them with SMB-sized coax inserts in the big holes, but there are also high-current connectors that fit in the same holes.

I don't know if DigiKey stocks them - Farnell certainly did and Element 14 (think Newark) probably still does.

--
Bill Sloman, Sydney

and

en

er short duration pulse. It really was the domain of the test equipment en gineers, but I was directed to do this. I was told that otherwise the test equipment guys would say it could not be done.

ignal going thru the shield on a coax cap. Then an integrater and another shield and feed thru cap to the level detection circuit. And it worked. So gave a sketch to the test equipment department.

quipment guys had made did not work. It turned out that they had carefully shielded the integrater , but had the load right next to the level detecti on circuit.

tom

e s

ribbon cable (or something else?) to the back panel.

ax inserts in the big holes, but there are also high-current connectors tha t fit in the same holes.

Yeah something like that. I'll keep all the high current stuff off of the connector, only control. (Well that could change if it doesn't work well e nough.)

George H.

4 (think Newark) probably still does.