Op amp when powered down

I have an application where I want buffer cell battery voltages with op amp voltage followers. Easy.

I want to shut down the op amps when monitoring is not required. So an op amp could have 17V on an input. Another in the same pack could have 4V. Is there an op amp that is specified to "work" in this mode?

In practice it works with an LM324. However it needs to be specified to work.

Reply to
Raveninghorde
Loading thread data ...

Are there any modern-process OpAmps that got to 17V... except expensive ones?

How do you shutdown an LM324? Cut its power and rely on the input stage lateral PNP's for isolation? Probably works OK. I don't think those old dudes had any ESD diodes to fret about, but I'd curve trace the inputs to verify. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

None I know of. CMOS switches to isolate the inputs would be my instinct, = but messy. =20

LM324's have PNP darlington inputs--you'd think those could break down with= 17v reverse Vbe. In your favor is two Vbe zeners in series, with the main= leakage path being through two forward Vbe's to the other input. So, if th= e differential input voltage is small you might get away with it (can't tel= l--TI's datasheet doesn't have the old Nat Semi internal schematic), but ce= rtainly not guaranteed.

Is the goal saving power? An ultra-low Iq op amp might solve that.

--=20 Cheers, James Arthur

Reply to
dagmargoodboat

messy.

reverse Vbe. In your favor is two Vbe zeners in series, with the main leakage path being through two forward Vbe's to the other input. So, if the differential input voltage is small you might get away with it (can't tell--TI's datasheet doesn't have the old Nat Semi internal schematic), but certainly not guaranteed.

The laterals break down at the BVCBO of the NPN B-C junction, or around 36V. That's why I previously posted that it should work, but check for sneak ESD paths... which I don't think exist on an LM324. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

e?

nct, but messy.

ith 17v reverse Vbe. =A0In your favor is two Vbe zeners in series, with the= main leakage path being through two forward Vbe's to the other input. So, = if the differential input voltage is small you might get away with it (can'= t tell--TI's datasheet doesn't have the old Nat Semi internal schematic), b= ut certainly not guaranteed.

That's pretty neat. It oughtta work then, though not guaranteed, of course.

-- Cheers, James Arthur

Reply to
dagmargoodboat

but messy.

17v reverse Vbe.  In your favor is two Vbe zeners in series, with the main leakage path being through two forward Vbe's to the other input. So, if the differential input voltage is small you might get away with it (can't tell--TI's datasheet doesn't have the old Nat Semi internal schematic), but certainly not guaranteed.

Guarantee! You want a guarantee ?:-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Do you mean you want to power the inputs without powering the supply pins? when you want "off mode" ?

you say the 324 suffers no ill effects I'm going to look at the circuit on the datasheet and try to explain why.

In the 324 powering the inputs with the chip itself unpowered depends on will have the the B-E junctions of Q1,Q2,Q3 qnd Q4, reverse biased. (see ti/natsemi datasheet )

This looks bad as BE junctions can normally only withstand about 5v howeve this is just ballpark figures for generic BJTs,

You can get a figure for the breakdown of Q1+Q2 or Q3+Q4 by looking at the differential input voltage limit for this part, the datasheet says "32V" for that measure.

looking at he datasheet. with 32V in in+ and 0V on in- and 32V on V+ and 0V on GND

In this state Q2 and Q3 emitters will be at about 0.6V , this means the other input is withstanding over 30V total reverse bias

so the inputs should withstand over 30V unpowered with no ill effects.

17V should be no problem at all
--
?? 100% natural

--- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net
Reply to
Jasen Betts

The inputs on an LM324 are _lateral_ PNP's. See my recent post regarding same.

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Thanks guys. Helpful answers.

Reply to
Raveninghorde

?
?

Good point. So, the input transistors won't zener. As long as there aren't any ESD protection sneak paths at the inputs, the internal stuff seems guaranteed by design...I think.

-- Cheers, James Arthur

Reply to
dagmargoodboat

Just what I said in...

Newsgroups: sci.electronics.design Subject: Re: Op amp when powered down Date: Mon, 19 Mar 2012 11:10:47 -0700 Message-ID: ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

So

ode?

ins?

on

ee

e
e

Jasen pointed out that the input specs guarantee the part can withstand 32v differential, something not previously noted. Was there any harm in acknowledging that?

-- Cheers, James Arthur

Reply to
dagmargoodboat

Nope. I thought you had missed it. Crikey! Why is everyone so cranky? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

ith

=A0So

uld

s mode?

y pins?

ds on

(see

at

+

the

here

Obama.

Reply to
dagmargoodboat

pins?

That'll do it to you ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

There are such things as channel protectors that go ultra-high impedance wh= en powered down, and will not significantly influence accuracy, especially = for someone using an LM324. If you go with and ADG467 octal, they come in a= t about $0.67 per channel in small quantity from DigiKey. But the single ch= annel ADG465 are not so cheap, might as well go with an ADA4096 as use them= .

Reply to
bloggs.fredbloggs.fred

Yeah, exacpt I think none of us understand the signifiganse of PNP being 'lateral' when I tried to look it up, my DSL quit. and now that it's back up google only finds me a thread on edaboard interspersed with targeted adverts where people are mainly interested in frequency response.

I understand that laterals are spread across the surface of the chip, wheras vertical are diffused, by over-doping the base onto the collector and then the emitter onto the base,

That would seem to make the laterals more symmetrical which would give an E-B junction simmilar to the C-B junction both having a high breakdown voltage?

DSL's down again :(

--
?? 100% natural

--- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net
Reply to
Jasen Betts

'lateral'

In....

Newsgroups: sci.electronics.design Subject: Re: Stable simple dc amplifiers Date: Sun, 18 Mar 2012 08:49:16 -0700 Message-ID:

I said to follow this link...

formatting link

This is a typical bipolar process cross-section. CMOS? Still a P-type substrate.

[You should particularly take note of this paragraph]....

A lateral-PNP is like a vertical-PNP except there are two P-base emitters side-by-side... plus a feeble attempt to kill the vertical device by adding buried layer ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

You can occasionally get into trouble in bipolar designs where components break down due to operating conditions not found in normal operation. [Translation: start up or ESD, mostly in bicmos]

Breakdown tests are two port, but the components have three ports. For example, you can measure the collector emitter breakdown with the base floating versus the base tied to the emitter. So it is a matter of knowing the breakdown voltages that end in "o" (open). I've seen rookie designers screw up by not watching for this kind of gotcha. [This is the kind of problem spice won't find.] The parts may work, but have reliability issues or odd start up behavior. When you sell parts by the millions, this stuff can ruin your day.

Note that absmax conditions are with the device powered. They are not the same for a powered down device where only an input is stressed.

If you want to power down a component, just buy a part with a power down pin. It has been engineered to behave well under that condition.

Reply to
miso

15 months later.

The LM324 works fine. Unless you fit ST.

I have found this before with ST copies. I used to use an LM311 with inputs above the + supply with 0 input current. Worked OK for years until purchasing bought ST parts.

My database also bans ST LM317Ts.

Rule of thumb, don't use ST LMxxx parts.

Reply to
Raveninghorde

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.