good cheese

Scroll down to Casu Marzu

"This cheese from Northern Italy has a special ingredient. No, not a special herb. It's Piophila casi, the "cheese fly," which lays thousands of eggs in this sheep's milk cheese. Casu marzu is served with the maggots alive, as once they die the cheese becomes toxic. Diners are encouraged to brush off the maggots before eating the cheese, given the larvae can survive human digestive tracts long enough to lay eggs. Hatching maggots then try to burrow through intestinal walls, causing pain, vomiting, and death. Not likely to put you in the mood for dessert."

--- If you weren't so stupid you'd realize you're abusing the first amendment and just adding to the noise non-idiots have to bear.

-- JF

messagenews:VPadna9St7ioFCvSnZ2dnUVZ snipped-for-privacy@earthlink.com...

Please, please, it's *Barrie* Gilbert.

-- "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." (Richard Feynman)

OK, I admit you did talk about power plugs in the basics group. You were wrong by 2:1, but you did talk about, well, not electronics, but electricity. Close enough.

We're eagerly awaiting JT's masterwork on LDO compensation, and he's promised us that.

It works, rev A, and we're shipping them and the customer is sending us checks. That sounds right to me.

The regulator that I posted isn't outside the box. It's plain and dull and quite ordinary. Its transient response is good enough for the application. Jim has no rational reason to slag it, but then Jim isn't rational.

I designed it, well, not on paper but in my head, values and all. But because things are nonlinear, and the fets run in their ohmic range, I couldn't mentally evaluate load-step response, so I Spiced it and tweaked the comp capacitor. It all took about an hour, and it works fine.

Jim is being an ass as usual, and he has committed to writing for us a masterwork on LDO compensation, with my circuit as a bad example. That would be fun, but he'll probably never do it. He's made empty promises like this before.

His excuse will be, of course, that I don't deserve to benefit from his brilliance.

You did claim to be a "master circuit designer." OK, show us your stuff.

Gosh, can't wait.

When you stay arm chaired, you tend not to know what is really happening out in the real world. Any bells ringing yet?

That's why the wonderful invention of fly swatters is such a nice commodity, one good whip and the useless parasite either exits or is wasted.

Filters are at times, just not enough.

Jamie

So you now have elected yourself to speak for the others?

Figures. I should not have expected anything but.

If you're trying to be one of the herders, I am not a sheep for your grazing field.

Jamie

messagenews:VPadna9St7ioFCvSnZ2dnUVZ snipped-for-privacy@earthlink.com...

Yes, didn't you know I have friends I talk to all the time? Multiple minds can excel beyond your imagination, and lately, it has been a doozy. ( your imagination, that is).

With me, I can actually admit to that, it does not bother me one bit. I find it beneficial.

------------------

Sheep, a whole room full.

Jamie.

But I did. It's not uncommon, with a box of this complexity, for there to be several iterations of prototypes, and for one or more PCB versions to be built and thrown away. I've heard of five. There are seven PC boards in the box, and every one is working properly at rev A. There was no prototype: we went directly from design to manufacturing. Many of the parts were new, so we had to get the PADS footprints and pinouts right the first time, too. It has PCI Express, eight lanes of 2.5 GbPS serial data with active cable equalizers. I only simulated the LDO and one ADC driver diffamp and the fast ramps (which was a waste of time, since the AD8014 model is junk.)

Really, three jumpers in the entire box ain't bad. All the mechanics was right first time, too; everything fit. That was because it was all modeled in SolidWorks.

The worst part was getting the PCIe to be reliable. It would power up erratically working on four or eight lanes, depending on temperature, and we wasted a bunch of man-weeks trying to find out why. It turned out to be a complicated tangle of powerup resets and PLLs and stuff inside the FPGA. Luckily, running on four lanes was enough to keep the software developers busy.

Except that I am going to write it up, because it caught my attention... why are LDO's "difficult"? They must be, so many complain of them here.

But your example shows classic misunderstanding of what needs to be compensated.

It made me start thinking about how _would_ you design a good LDO.

I'm not sure you can with off-the-shelf parts.

But I have a theory. ...Jim Thompson