Low volt indicator for LDO regulator

[Martin Evans]

I an running a battery powered circuit that uses a 3.3V LDO regulator.
The drop out voltage is 4.3. I would like to add a LED indicator to
the circuit to show when the battery voltage has fallen below about
4.5V (ie. LED goes on of off). Can anyone please offer a simple
circuit to achieve this?

Martin Evans

 

[Martin Evans]

Even if you mean that your regulator drops out of regulation when your
battery voltage is at 4.3V, that's still a 1V drop out voltage -- and
that's not very low.

Yes, that is what I meant.

Farnell has the LP2951 which has a low voltage error flag. The
datasheet says it can be "pin strapped to 3, 3.3 or 5V", but I cannot
find any further details there or online.

http://au.element14.com/national-semiconductor/lp2951acn/v-reg-ldo-adj-1-2429v-2951-dip8/dp/9488553

Is anyone familliar with this procedure?

Martin Evans

 

[Martin Evans]

http://lmgtfy.com/?q=LP2951+datasheet

Yes, why didn't I see that.

Mark

 

[Spehro Pefhany]

On Tue, 25 Oct 2011 01:19:15 GMT, the renowned

I an running a battery powered circuit that uses a 3.3V LDO regulator.
The drop out voltage is 4.3. I would like to add a LED indicator to
the circuit to show when the battery voltage has fallen below about
4.5V (ie. LED goes on of off). Can anyone please offer a simple
circuit to achieve this?

Martin Evans

BD4945G (plus a resistor & LED), provided 10V abs. max. input voltage
is okay. Will switch at 4.5V +/-1%.


Best regards,
Spehro Pefhany

 

[Joerg]

I an running a battery powered circuit that uses a 3.3V LDO regulator.
The drop out voltage is 4.3. I would like to add a LED indicator to
the circuit to show when the battery voltage has fallen below about
4.5V (ie. LED goes on of off). Can anyone please offer a simple
circuit to achieve this?

Martin Evans

One of the cheapest methods is to use a reset chip such as this:

http://www.onsemi.com/pub_link/Collateral/NCP302-D.PDF

They make a 4.5V version if you don't want a resistive divider. These
come in open drain and push-pull versions. That would reduce the whole
effort to one lone part of about 15 cents, so forgive me if I don't draw
a schematic

A reset chip has another advantage in that you can set a time delay. In
other words you can keep it indicating "low" even if the batery just
have a wee fainting spell of short duration because some load came on.
All this feature requires is the investment into a 1-2 cent ceramic cap.

As for drop-out, as Tim wrote, 1V isn't really LDO but marketeers
sometimes call it that because it sounds cool. Be careful with real LDO
regulators. Besides being iffy in stability with low ESR caps they often
have some undocumented issues. Such as going berserk when the source
impedance (your batery) becomes too high.

 

[Joerg]

with a bit of tweaking with the output voltage you could just use
another
lp2951



if you can live with a slow rising error flag, a cap on it would do
the same
on a lp2951

I will not use that series. In fact, I won't use LDOs in general if at
all avoidable, and usually it's better to use a switcher anyhow. Had a
very bad experience with the LM2931 which a client used against my
advice. Sure, I got it stable. But it went berserk when the source
impedance became too high and this was nowhere mentioned in the
datasheet. In fact, it even surprised the manufacturer who did not have
a solution short of a redesign of the IC.

the lp2951 is a real LDO, typ. 380mV@100mA and it is picky about the
output cap

I'd rather use one that has a NPN or N-channel pass element and a charge
pump. You can get even lower drop-out that way and it's stable. That
would be an LDO I'd use but so far had to roll my own because commercial
ones were either unavailable or prohibitively costly.

 

[Martin Evans]

One of the cheapest methods is to use a reset chip such as this:

http://www.onsemi.com/pub_link/Collateral/NCP302-D.PDF

They make a 4.5V version if you don't want a resistive divider. These
come in open drain and push-pull versions. That would reduce the whole
effort to one lone part of about 15 cents, so forgive me if I don't draw
a schematic

Not 15 cents, but I like the functionality:

http://au.element14.com/intersil/icl7665sacpaz/voltage-detector/dp/1562053

Would the above IC also work as an audio clipping indicator?

Martin Evans

 

[Spehro Pefhany]

Looks like that to me ;-)

...Jim Thompson

What am I missing? - it looks okay to me (though the threshold for the
LED turning off will be lower than 4.5V).


Best regards,
Spehro Pefhany

 

[Joerg]

Not 15 cents, but I like the functionality:

http://au.element14.com/intersil/icl7665sacpaz/voltage-detector/dp/1562053

Would the above IC also work as an audio clipping indicator?

Looks like the 7665 is way too sluggish for that in terms of reaction
speed. The NCP series chips could do that. However, keep in mind an
intricacy of such chips: They do not have a separate supply voltage.
Input is their supply. This means that at very low input amplitudes
(well under 1V, usually) they "let go" because there is nothing left to
coax the output FET into a conductive state. So you'd see a false alarm
for very low audio levels unless you pre-bias the chip. For audio
clipping indication I'd use a comparator.

 

[Spehro Pefhany]

Generally the output of an LDO tracks input with a constant delta once
you reach drop-out.

...Jim Thompson

Okay, I'm kind of ignoring the numbers the OP used for dropout,
because I think he added a margin to the actual dropout.

Most of the LDOs I've used have dropout voltages of a few tens or
hundreds of mV at moderate currents (battery power). Even old bipolar
ones like the LP2950, so << 600mV.

If he wanted to increase the ~4V LED "off" to 4.5V, a BE resistor
would do it (>1V dropout, or "saturation" voltage)

But if you postulate an old semi-low-dropout like the LM1117 (>1V
saturation voltage), no it won't work properly as shown.


Best regards,
Spehro Pefhany

 

[ehsjr]

Looks like that to me ;-)

...Jim Thompson

_Off_ perpetually, assuming the LDO in/out delta is always > ~.7

Ed

 

[ehsjr]

Yep, I stand corrected.

...Jim Thompson

Well, I meant to reply to Bitrex, so I was in error. You just
agreed with his statement, after already pointing out that there's
a problem with the circuit. In any case, the on/off thing is like
a typo, very easy to type (or think) off when you mean on or vice
versa. You pointed out the important factor: the regulator chip
in/out delta.

Ed

 

[ehsjr]

It's an interesting problem. The proper way would use something like
the old National "Miser" circuit to flash the LED at low duty cycle to
conserve battery power.

...Jim Thompson

I don't like using battery power to indicate the battery is low. :-(
I prefer auto power off and "negative indication", something like
a led that goes out, or the gizmo being powered stops working.
'Course that's not always a good idea - and that does make it an
interesting problem.

Ed