Simulation software for DC/DC converter

who buys

to

letting

that

from

I could have sworn that was your point. Then we've been discussing nothing.

"Getting the answers" does not always involve the designers.

So problem solved.

TI would have given me an number I needed, if they had them, but it'll be a month before they're available. I should have designed in a different regulator but no one raised a red flag until I was kiting the parts.

Sometimes? Ya' think? They are good for a laugh.

There

Then we've been wasting a lot of bits.

The ones that do provide models that are an embarrassment. They're usually behavioral so not valid in the corners or even for oddball measurements. Even LTC, though at least they try to document the assumptions.

Rubbish. Cutting edge design is done all the time without the IC developers even knowing it happened.

over

Nope. I have designed in 8051's where the clients have later successfully released 2nd sources. I am not familira with ARMs so I don't know if it's possible there. But you must stay with the standard

BT :-)

Seriously, this has cost me a summer vacation trip once.

over

You flatter the 8051 by claiming that it has an architecture.

In a lot of my cases it does. Usually when I ask a question it ain't in the datasheet.

I wouldn't go that far just yet. That was one positive event a few weeks ago versus numerous negative ones before.

You guys still do kitting? Wow. I've put a stop to that at EndoSonics 13 years ago.

If it wasn't so sad. Why are they compromising a good reputation that way? I mean, otherwise they are rather commendable. For example, I can't praise their open learning program loud enough.

if

Some stuff for switchers has to be behavioral or each sim could take hours. And I have to hand it to them, except for unorthodox stuff like dithering or sync'ing LTC's switchers come out on the circuit board almost verbatim as simulated. That is a major advantage they've got over others. For mission-critical jobs I wouldn't do it any other way if I have the choice.

Not in medical or aerospace.

over

Well, it has about the flair of a VW Beetle or maybe an AMC Gremlin, it ain't a race car. But there are reasons why Cypress uses that architecture on newer PSoC series.

over

A long time ago, in a galaxy far, far, away.

BT, made 0$, not even refunded travel. :-(

over

If you mean "well thought out", well, no. I propose that many buildings are just as well thought out, so "architecture" isn't a bad description. ;-)

over

I rather like the 8051, though it's not at all symmetric. Silicon Labs has some pretty nice 8051s, still. PSoCs come in ARM varieties, too, for those who feel insulted by the 8051. ;-)

FAEs have almost always gotten the answers quickly.

I haven't dealt with Infineon so wouldn't know.

buy

I do pre-production platform design - engineering prototypes. We have to kit the parts for them, yes. For manufacturing? No.

Because

...

Because 99% of the US population is scientifically illiterate. MIT's BS sounds good and gets good press. 99% of the other 1% doesn't pay attention to anything the popular press says. ...and then there's Larkin, to give us all, here on SED, a good laugh. ;-)

if

As long as you're simulating something they already thought of...

Rubbish.

over

shot if

Even

That's always the risk and I've found a bug in one of their chips. But they fessed up. Honesty goes a long ways in business.

I have the impression you sometimes judge stuff without knowing much about it. So tell us all, when was the last time you did a medical or aerospace design? For me: This week.

over

Program a 32x32 signed fractional multiply on one and get back to me.

Unless some radical CS concept takes over, ARM is probably *the* CPU architecture of the next decade or three.

parts.

sleep over

parts

It is not made to do that. In the same way that you would not use pliers to adjust the alignment of an IF filter. It is made to do other things and do them rather well.

It all depends on the application. MSP430 has a good future as well, because of the low power modes.

parts.

sleep over

parts

It's getting harder or maybe they just don't admit to it. ;-)

...and the morons program 8051s in C. ;-)

parts.

sleep over

parts

Done it. After 16x16 it's just loops.

Like alternate energy, I've been hearing that for a couple of decades already.

parts.

sleep over

parts

We use AVRs for exactly that reason. They can be left on.

products

Exactly. Democrats love it.

with

Not a pickup.

Well, it's also proportional to what the dealer wants to make. Some will use them as loss-leaders, some it's their only profit other than financing kick-backs.

shot if

Even

But the simulators (models) are rarely useful in places they haven't already beaten to death.

IC developers don't sit holding Boeing's hand. Get real.

shot if

Even

Having a behavioral model sure beats having none. Sending a non-sim'd design into DO-160 testing sure would be a white-knuckle ride. That's because in this business delivery dates are fairly firm and when such a deadline isn't met the big Klieg lights turn around, towards the guy who dunnit. Remember Dilbert's visibility cartoon?

developers

They do. Because they know that that's the only way to remain a driving force in higher end markets. Meaning high margin markets.

Why do you think LTC provides such excellent support to guys like me? They know that I use their parts only in high-end lower volume designs.