techniques for sensing fairly small currents (<1A)

[Michael]

Hi - I am working on a power board that needs to monitor currents on a
number of supply lines. Currents should all be fairly small - all less
than an amp. Largest supply voltage will be 11.1V.

What methods exist for sensing currents this small? The only one which
I am familiar with is where a small shunt resistor (preferably a high
side shunt) is used and then the voltage across that resistor is
amplified, often with a current sense amplifier.

Is that it?

Thanks,

-Michael

 

[Phil Allison]

"Michael"

Hi - I am working on a power board that needs to monitor currents on a
number of supply lines. Currents should all be fairly small - all less
than an amp. Largest supply voltage will be 11.1V.

What methods exist for sensing currents this small? The only one which
I am familiar with is where a small shunt resistor (preferably a high
side shunt) is used and then the voltage across that resistor is
amplified, often with a current sense amplifier.

Is that it?

** No.

Try looking for " Hall effect sensors " and current transformers.

The former operate with AC and DC currents while the latter are AC only.

Both are non contact ( aka galvanically isolated) sensing methods.



....... Phil

 

[Michael]

** No.

Try looking for " Hall effect sensors " and current transformers.

The former operate with AC and DC currents while the latter are AC only.

Both are non contact ( aka galvanically isolated) sensing methods.

...... Phil

Hi Phil - I should mention I'm only interested in DC current sensing.

I'm aware of hall effect current sensing - but isn't that only viable
for large currents? A quick google only finds parts designed for
larger current ranges than what I'm interested in.

-Michael

 

[Phil Allison]

"Michael"
"Phil Allison"

Hi Phil - I should mention I'm only interested in DC current sensing.

I'm aware of hall effect current sensing - but isn't that only viable
for large currents? A quick google only finds parts designed for
larger current ranges than what I'm interested in.

** Most linear Hall effect sensors work down to the 1 mA level, just
ne - Lem, Honeywell etc.

All YOU need do is run a few turns of insulated wire through the hole to
increase the sensitivity.



....... Phil

 

[Terran Melconian]

What methods exist for sensing currents this small? The only one which
I am familiar with is where a small shunt resistor (preferably a high
side shunt) is used and then the voltage across that resistor is

Keep in mind that many other components, such as FETs, can look a lot
like resistors, and the addition of a dedicated part may not be
necessary.

Why do you prefer high-side to low-side?

 

[Michael]

Keep in mind that many other components, such as FETs, can look a lot
like resistors, and the addition of a dedicated part may not be
necessary.

Hmm, this is a valid point. The power board will just be stepping down
a large DC voltage to various smaller DC voltages - so there very well
might be a good component to grab a current reading from. I'll have to
look into that further when the design is farther along.

Why do you prefer high-side to low-side?

So that I can use a shared ground.

-Michael

 

[Rich Grise]

On Apr 27, 10:18 am, Terran Melconian


So that I can use a shared ground.

That seems a little counterintuitive. If you used low-side sensing,
then all of the sense resistors would have ground in common with the
supplies. With high-side, you need more sophisticated circuitry
because the sense resistors will all be floating at their supply
voltage.

Hope This Helps!
Rich

 

[Anthony Fremont]

That seems a little counterintuitive. If you used low-side sensing,
then all of the sense resistors would have ground in common with the
supplies. With high-side, you need more sophisticated circuitry
because the sense resistors will all be floating at their supply
voltage.

IANEE, but not really counter-intuitive IMO. He wants the grounded side of
his stuff to all be at the same potential, ground. If all his stuff
ran on a kajillion volts, he probably wouldn't want the grounds to all be
thousands of volts apart from each other. ;-)

 

[kell]

Hi - I am working on a power board that needs to monitor currents on a
number of supply lines. Currents should all be fairly small - all less
than an amp. Largest supply voltage will be 11.1V.

What methods exist for sensing currents this small? The only one which
I am familiar with is where a small shunt resistor (preferably a high
side shunt) is used and then the voltage across that resistor is
amplified, often with a current sense amplifier.

Is that it?

Thanks,

-Michael

High side current sensing requires more complicated circuitry than
sensing at ground, but if that's what you want here are a couple of
ideas. I'm praying the ASCII comes out ok, because I'm using groople
and they seem to have eliminated the fixed font feature. Thank
Winfield Hill for the first one and Fred Bloggs for the second.


+ -----+--- Rsense ----+---->
.. | |
.. Rc |
.. +------------, |
.. | __ | |
.. c\| / +|--' |
.. NPN |---< | |
.. e/| \_-|-----'
.. |
.. Rload
.. |
.. - -----+-------------------->


..
.. .---------------------. R1
.. | | 1.24V - (Vled-Vbe)*-----
.. | | R2+R1
.. ----+-----[Rs]------+-------> I= ----------------------
.. | | | Rs* R2/(R1+R2)
.. | | |
.. | | |
.. | [R1] |
.. | LM185 | |
.. --- | |
.. / \ -------------+ ---
.. --- | \ / led
.. | | ---
.. [R4] [R2] |
.. | | |
.. | | |
.. +---------------' |
.. | |
.. | |
.. | |
.. >| |
.. |-------------------+
.. /| |
.. | |
.. +---> to comp [R3]
.. | |
.. [Rc] |
.. | |
.. --- ---
..
..

 

[Michael]

IANEE, but not really counter-intuitive IMO. He wants the grounded side of
his stuff to all be at the same potential, ground. If all his stuff
ran on a kajillion volts, he probably wouldn't want the grounds to all be
thousands of volts apart from each other. ;-)

Exactly. Even though everything is fairly low voltage, I still have a
number of supply lines on the same board. If I were to sense ground
currents I'd have to have a separate ground for each voltage - while
now I just have one big ground plane.

-Michael

 

[Ken Smith]

[... high side current sense ....]
I prefer doing this:
Modified ASCII art

+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

This has the advantage that the minus supply fo the op amp can be
something other than ground.

Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.

 

[Rich Grise]

Exactly. Even though everything is fairly low voltage, I still have a
number of supply lines on the same board. If I were to sense ground
currents I'd have to have a separate ground for each voltage - while
now I just have one big ground plane.

Oh, well that's different!

Never mind. ;-)

Cheers!
Rich

 

[kell]

[... high side current sense ....]
I prefer doing this:
Modified ASCII art

+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

This has the advantage that the minus supply fo the op amp can be
something other than ground.

Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.

+ -------+--- Rsense -------+---->
. | |
. Re |
. +-\/\/-+--------, R
. | ! __ | |
. e\| C / -|--' |
. PNP |----+-< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

And use an op-amp with input range that includes the top rail; a j-fet
op amp like TL082 would do, or there's the LMC6482, which has a lower
offset voltage.

 

[MooseFET]

kell <[email protected]> wrote:

[... high side current sense ....]
I prefer doing this:
Modified ASCII art

+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

This has the advantage that the minus supply fo the op amp can be
something other than ground.

Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.

And use an op-amp with input range that includes the top rail; a j-fet
op amp like TL082 would do, or there's the LMC6482, which has a lower
offset voltage.

Or there's the LM301. You can over compensate it to make it stable if
needed. You can also clamp on pin 8 to prevent the thing from trying
to make too much current in the PNP when you have an overload.

With R-R op-amps, you just limit the current in the PNP by limiting
the minus supply voltage.

 

[kell]

On Apr 28, 1:20 pm, [email protected] (Ken Smith) wrote:

[... high side current sense ....]
I prefer doing this:
Modified ASCII art
+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->
This has the advantage that the minus supply fo the op amp can be
something other than ground.
Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.
+ -------+--- Rsense -------+---->
. | |
. Re |
. +-\/\/-+--------, R
. | ! __ | |
. e\| C / -|--' |
. PNP |----+-< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

And use an op-amp with input range that includes the top rail; a j-fet
op amp like TL082 would do, or there's the LMC6482, which has a lower
offset voltage.

Or there's the LM301. You can over compensate it to make it stable if
needed. You can also clamp on pin 8 to prevent the thing from trying
to make too much current in the PNP when you have an overload.

With R-R op-amps, you just limit the current in the PNP by limiting
the minus supply voltage

I'm thinking, just use a base resistor.

 

[kell]

On Apr 28, 1:20 pm, [email protected] (Ken Smith) wrote:

[... high side current sense ....]
I prefer doing this:
Modified ASCII art
+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->
This has the advantage that the minus supply fo the op amp can be
something other than ground.
Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.
+ -------+--- Rsense -------+---->
. | |
. Re |
. +-\/\/-+--------, R
. | ! __ | |
. e\| C / -|--' |
. PNP |----+-< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

And use an op-amp with input range that includes the top rail; a j-fet
op amp like TL082 would do, or there's the LMC6482, which has a lower
offset voltage.

Or there's the LM301. You can over compensate it to make it stable if
needed. You can also clamp on pin 8 to prevent the thing from trying
to make too much current in the PNP when you have an overload.

With R-R op-amps, you just limit the current in the PNP by limiting
the minus supply voltage.- Hide quoted text -

- Show quoted text -

I just glanced at the data sheet of the 301.
It doesn't look like the input goes to the rail, at least not the
inverting input. It stops two or three volts shy.

 

[Fred Bloggs]

Hi - I am working on a power board that needs to monitor currents on a
number of supply lines. Currents should all be fairly small - all less
than an amp. Largest supply voltage will be 11.1V.

What methods exist for sensing currents this small? The only one which
I am familiar with is where a small shunt resistor (preferably a high
side shunt) is used and then the voltage across that resistor is
amplified, often with a current sense amplifier.

Is that it?

Yep- that's basically it, convert the current to a voltage with
something called a resistor. The actual circuitry to do this will be a
function of your intended use. For example are you monitoring for a trip
or foldback threshold , or are you monitoring for a measurement to be
displayed, or are you monitoring for some other purpose, and if so, then
how much range do you need i.e. 10:1, 100:1, 1000:1 etc. As usual we
will be standing by to give you the best possible answer, as well as
including a compendium of 10,000 ascii art drawings of typical circuits
if that's what you desire. We aim to please you in every way possible
and eagerly await your evaluation of our performance, maybe even
declaring one or the other of us a "winner," the ultimate purpose of our
existence...

 

[MooseFET]

[... high side current sense ....]
I prefer doing this: [....]
Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.
+ -------+--- Rsense -------+---->
. | |
. Re |
. +-\/\/-+--------, R
. | ! __ | |
. e\| C / -|--' |
. PNP |----+-< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->

[.....]
Or there's the LM301. You can over compensate it to make it stable if
needed. You can also clamp on pin 8 to prevent the thing from trying
to make too much current in the PNP when you have an overload.

With R-R op-amps, you just limit the current in the PNP by limiting
the minus supply voltage

I'm thinking, just use a base resistor.

The base resistor doesn't do a good job of setting a fixed limit. The
HFE of a transistor varies a lot from unit to unit.

 

[MooseFET]

On Apr 28, 1:20 pm, [email protected] (Ken Smith) wrote:

[... high side current sense ....]
I prefer doing this:
Modified ASCII art
+ -----+--- Rsense ----+---->
. | |
. Re |
. +------------, |
. | __ | |
. e\| / -|--' |
. PNP |---< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->
This has the advantage that the minus supply fo the op amp can be
something other than ground.
Also you may need to add a capacitor to this or the other version to
prevent oscillation. Adding an extra resistor in the input path to the
op-amp also protects it against short circuits in the load.
+ -------+--- Rsense -------+---->
. | |
. Re |
. +-\/\/-+--------, R
. | ! __ | |
. e\| C / -|--' |
. PNP |----+-< | |
. c/| \_+|-----'
. |
. Rload
. |
. - -----+-------------------->
--
--
[email protected] forging knowledge
And use an op-amp with input range that includes the top rail; a j-fet
op amp like TL082 would do, or there's the LMC6482, which has a lower
offset voltage.

Or there's the LM301. You can over compensate it to make it stable if
needed. You can also clamp on pin 8 to prevent the thing from trying
to make too much current in the PNP when you have an overload.

With R-R op-amps, you just limit the current in the PNP by limiting
the minus supply voltage.- Hide quoted text -

- Show quoted text -

I just glanced at the data sheet of the 301.
It doesn't look like the input goes to the rail, at least not the
inverting input. It stops two or three volts shy.

Look more closely.

 

[Alex]

Hi - I am working on a power board that needs to monitor currents on a
number of supply lines. Currents should all be fairly small - all less
than an amp. Largest supply voltage will be 11.1V.

Hi Michael
I have used some highend sensor-ic's from Maxim try search for high
end sensors at maxim.. they can do what you need.
alex