Opamp that _really_ does Voffset < 1mV

I need an amp, to operate from +22V and 0V rails. I need it to have an offset voltage of less than 1mV, and an offset bias current of less than

750pA -- preferably at the same time.

Aside from that, and low price, I don't really need anything to write home about. It'll be low-pass filtering the result anyway, so GBW >

100kHz is fine; it doesn't need to pull down to the negative rail (or up to positive), so rail-rail-ness isn't needed, etc.

I'm looking, and I'm getting RATHER TIRED of Digikey's propensity to list _typical_ offsets rather than _maximums_. It makes comprehensive searches _tedious_.

--
www.wescottdesign.com
Reply to
Tim Wescott
Loading thread data ...

OPA130

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

formatting link

Reply to
Vladimir Vassilevsky

Il Thu, 28 Jul 2011 16:37:53 -0500, Tim Wescott ha scritto:

What about this?

formatting link

M
--
Frustra fit per plura quod fieri potest per pauciora
Reply to
Michele Ancis

$1.71/ea vs. $0.60/ea for other parts, _max_ Voffset is still 2mV.

--
www.wescottdesign.com
Reply to
Tim Wescott

$4.76/ea? Ouch! My customer would whack me with a broomhandle, and I would deserve it.

--
www.wescottdesign.com
Reply to
Tim Wescott

Your "whole question" does not include your maximum price.

Reply to
John S

Il Thu, 28 Jul 2011 16:59:20 -0500, Tim Wescott ha scritto:

LF411A, which is the proper part for your supply range, has 0.5mV max if I read correctly. Sorry, I am not english mothertongue: $xx/ea means "each"? Where do you get your pricing? And what's the target price then?

M
--
Frustra fit per plura quod fieri potest per pauciora
Reply to
Michele Ancis

Il Fri, 29 Jul 2011 00:28:31 +0200, Michele Ancis ha scritto:

Ooops, I didn't read carefully ;-) You can use the LF411 as well between 22 and 0, of course.

--
Frustra fit per plura quod fieri potest per pauciora
Reply to
Michele Ancis

$xx/ea requires that business be your mother tongue, actually.

Target price is cheap cheap cheap! I think I found a solution that'll let me use a crappy op-amp, though -- see the "Is this circuit idea crazy?" thread.

--
www.wescottdesign.com
Reply to
Tim Wescott

formatting link

Scroll to the bottom of the page.

Hope This Helps! Rich

Reply to
Rich Grise

Il Thu, 28 Jul 2011 18:00:32 -0500, Tim Wescott ha scritto:

Well, if you do business in/with US, probably. However, you're right indeed: business is not my mother tongue ;-)

I am reading it, however I don't know what a "high-side current monitor" is so I'll need to cover the basics first. I'll take it as an occasion to learn something new (which is by the way the reason why I am reading SED).

About the business side, if you say "cheap cheap ch.." any undefined amount of times, this gives no actual clue as I (and probably others) don't know what "cheap" is. If you meant compared with "crappy" OpAmps, then it would be a lost battle: you cannot have "more performance" with "same money", can you?

M
--
Frustra fit per plura quod fieri potest per pauciora
Reply to
Michele Ancis

ave

if

u
l

is

.

High side: The high voltage side of things, typically the high voltage side of a switched mode power supply.

Current monitor: measures current consumption, passes the info to some other part of the circuit

nt

d

an

Why would an opamp reading a 15milliohm sense resistor need picoamps offset? Its hard to answer a question when its clear we havent been given all the relevant info.

NT

Reply to
NT

Input error includes input resistance times bias current added to that offset voltage, of course. That means bias current drift and tempo are also important. The balancing of input resistances helps.

Reply to
whit3rd

Because 100pA of input bias offset, times 1k-ohm of input resistor, turns into 100nV of bias voltage. There is a trade that can be done with all- around lower value resistors, but then instead of worrying about the transistor beta stealing current, I need to worry about the gain resistor stealing it.

Input bias current does count -- if it didn't, folks wouldn't list it on their data sheets.

--
www.wescottdesign.com
Reply to
Tim Wescott

So, saying "way cheap" isn't better?

I can get a "crappy" (Vin Ordinare?) LM358A for something like $0.60/ea, in qty 1 -- less by the 100's, but the price tracks. So putting something in that solves all my problems and costs something like $1.20/ ea, qty 1, doesn't hurt too much. But going more than twice the price makes me think that a few lines of code and a hiccup in the current every once in a while is a pretty attractive trade-off.

--
www.wescottdesign.com
Reply to
Tim Wescott

The offset bias current is going to be a problem. It rules out the OP07 and similar amps. Does it really have to be that low?

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

No. Because I'm an idiot, I slipped three digits. 750nA. Nano. Not pico. D'oh.

--
www.wescottdesign.com
Reply to
Tim Wescott

Oh well, that's only 60dB more :-)

750nA is huge. Did you check out the OP07? Around 30c in large quantities if you take the C-version which should hopefully fit your definition of "cheap cheap cheap". Comes from half a dozen mfgs, probably some patent ran out like they do on medicines.

formatting link
formatting link

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

Well, cheap cheap maybe. But inexpensive compared to the alternatives.

I had hoped to figure out a way to use a _really_ cheap amplifier like the LM358, and let the microprocessor null out any offset -- but that runs into if I'm trying to do level translation on a high-side driver. So unless I wanted to get really really weird and do a home-rolled chopper (which may work if you've got a micro sequencing everything, but is still a bit too high on the weirdness scale for customer designs) then I think a straight-up circuit using the OP07 is the ticket.

--
www.wescottdesign.com
Reply to
Tim Wescott

The homebrew chopper could beat it in price but that only makes sense if your client will be looking at a yearly production volume in the six digits. Else your billed time would dwarf their savings.

--
Regards, Joerg

http://www.analogconsultants.com/
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.