LDO problem

small

make

would

use

no

current.

scorching

The

My fave used to be the MAX900. RIP.

Cheers

Phil Hobbs

set

small

make

would

use

and no

current.

scorching

The

"Fav" and MAX *anything* don't belong in the same sentence.

set

small

make

would

use

no

current.

scorching

For 20x the price, maybe.

Even a quad op-amp

Some will. Some opamps go bonkers when used as comparators. Power consumption can go way up, and opamps with shared bias generators can have bizarre channel interactions.

John

current.

If I put a diode in series with it, it won't be stiff any more. If I don't put separate diodes in each pullup leg, they talk to one another. And the inputs can be outputs, externally pulled up to +40 maybe.

John

om

nce.

, set

small

o make

his

ould

r,

u use

d no

e current.

I meant putting the diode in the feedback loop. That way the pull-up bus would be stiff, a voltage source.

R ensures a minimum load.

But if there's the chance of an external pull-up to a higher voltage, well, that's a horse of a different color--discrete schottkies it is.

James

Earlier on in the thread you mentioned using 1k pullups. Why so low? That's a lot of power. Do you think anyone really needs the "speed" you'll be able to achieve at 1k?

Just curious.

steve

nuisance.

set

small

make

this

would

you use

and no

some current.

scorching

The

Nah, that was one of the ones they actually made in quantity for years. IIRC they switched foundries and couldn't make them anymore.

The best thing about them was their well-behaved output latch. I often want to use AC feedback from the output to the latch pin, to make a pulse stretcher that doesn't push crap back out the inputs the way the usual AC Schmitt approach does. Most latched comparators oscillate when you do this, but not the dear departed MAX900. *sniff* *sniff*

Cheers

Phil Hobbs

I like that "speed" in quotes. What was the pullup in RTL? I forget.

Since I don't know what a user might do with a general-purpose i/o board, I can't select the right value. 1K is about as low as I can go for various power reasons. It would be enough to drive an LED or an SSR directly. My users are aerospace/industrial, so would typically have a lot of cable between a VME crate and whatever gadget they are driving, so the lowest feasible pullup sounds good. They could use it for continuity testing too, maybe.

This is a case of inventing a product that has no known users, hoping to include enough interesting features that somebody will buy it. We have been surprised by what sells and doesn't sell, so we try to get lots of products out and see what happens. I swore that we didn't want to be in the 4-20 mA business, because everybody does that already and it's too cheap. We wound up doing a 4-20 i/o module, for a customer who said he'd buy them and then didn't, and it's looking to be a hot seller to other customers. Weird.

Maybe half of our products are money makers. But I bet that 90% of software startups fail these days. I have stock in one!

John

nuisance.

set

small

make

this

would

resistor,

you use

and no

some current.

and

scorching

The

Maxim seems to farm out a lot of their fab. We got nailed on the MAX9690 comparator. Fab problems at some little subcontractor made for failure time-bombs. Rather than fixing it, Maxim just discontinued it without warning. I rather liked the part, and we had to "make our own" replacement as a tiny PCB that soldered into an SO8 footprint.

ftp://jjlarkin.lmi.net/Break2.jpg

ftp://jjlarkin.lmi.net/OnBoard.jpg

This sort of shut down NIF. Painful, overall. So we don't buy Maxim any more.

John

snipped-for-privacy@yahoo.com

nuisance.

so, set

small

to make

this

would

resistor,

you use

and no

some current.

and

scorching

point. The

op-amp

Best looking roach wires I ever saw. ;)

Cheers

Phil Hobbs

...and that's a good thing? Max has a history of this, without even a OTB possibility. I spend time designing their crap out, not in.

small

make

current.

How about an active pulldown for that case?

|\| | +---|-\ |/ | | >-+---+| | -|+/ | |>

| |/| | | | | | | | V | | - | | | +---------------+------++-----+------+ | | | | .| | |/ .-. .-. .-. ++-| | | | | | | |< | | | | | | | '-' '-' '-' | | | | === V V V GND

(created by AACircuit v1.28.6 beta 04/19/05

set

small

make

use

current.

The user could still pull a pin up to 40 volts and toast my pullup.

John

nuisance.

set

small

make

would

use

no

current.

He could still light the building on fire, too. ;-)

I hate to throw accuracy away for nothing.

Did I say that? They were available in volume for about 15 years, which is pretty good for Maxim, and they had this very nice behaviour that you couldn't get elsewhere, certainly not for the price. If I had been designing mass-production hardware back then, which I wasn't, I might well have used separate RS flipflops for the latching function, or possibly a second layer of comparators with AC positive feedback.

I used them in the false-alarm-rate tracker in the ISICL sensor,

Cheers

Phil Hobbs