I know there are zillions of posts about current measuring, but I'm seeking advice on making my own shunt resistor.
If I create a shunt resistor by using 14 x 68 Ohm metal film 0.25W 1% resistors I will get an equivalent resistor of 4.85 Ohm 3,5W with a tolerance >= 1% which should be stable temperature wise and through which I should be able to measure currents up to 700mA at 5V (safer if we say up to 500mA).
Bumping to 25 x 120 Ohm I get 4.8 Ohm and up to 6.5W max or 1.3A at 5V (1A to stay safe).
Unless I need to measure higher currents, will I get decent readings by sampling voltage on such resistor array?
The principle is sound, but in practice you'd use a much lower value shunt resistance
and amplify the voltage across it before measurement.
So 0.1 ohm shunt gives 100mV per amp, boost that by a gain of 50 to pass
to an analog input. A precision rail-to-rail opamp in differential amplifier configuration
would be used in a Kelvin connection to the shunt (Kelvin connection = 4 wires, 2
for the current, 2 for the sensing). Note that a 0.1 ohm resistor only dissipates 0.1W
at 1A.
There are various chips that are specialised to do this BTW, such as the INA226 high-side
current-shunt chip.
However the shunt of 4.85ohms with 500mA will cause a drop of 4.86 x .5 = 2.43 volts. which is significant if you are in a 5, 12 or 15V circuit.
The aim is to have as small a drop as possible, hence MarkT suggestion of small shunt value and amplify the dropped voltage with an instrumentation or differential amplifier.
Right, I actually did came to the same conclusion myself after trying the circuit on jCircuit: to get 1A flowing through a 5V circuit I had to lower the circuit resitance down to 5 Ohm, which was practically a voltage splitter when a 4.9 Ohm "shunt" resistor is placed in series with it.
I now better understand why a valid shunt resistor is in the order of the mOhm or less: I was just trying to avoid the need for the opamp, but it seems a requirement.
I had a look at the INA226 and the INA220 (which is both high and low side) and they are both TSOP: any suggestion for TH equivalent?
In any case this ICs seems to require an external shunt (am I wrong?) which increases the BOM cost while I was looking to a way to reduce it....
There are break-out boards obtainable for these current sense chips complete with shunts,
checkout various vendor's "sensor" categories.
Opamps are easy to get in DIP packages of course. Just make sure its rail-to-rail
input and output.
A large-sized surface mount shunt resistor is both cheap and hand-solderable, note. You
can use a short length of thin gauge copper wire as a shunt (charts of ohms-per-metre
are everywhere on the internet for various copper wire sizes).
rlogiacco:
I had a look at the INA226 and the INA220 (which is both high and low side) and they are both TSOP: any suggestion for TH equivalent?
Depending on what degree of "precision" you require, which I feel is but modest if you are using an ATMega328 to read it, I continue to suggest the LM386 as a rather elegant low-side sense amplifier which with minimal external components, scales directly to the Arduino input (though to be fair, with a limited range).
Errrrrrrr.... Are those INA ICs just OpAmps? I mean, I know you need an OpAmp, but I thought those INA IC were somewhat specialized in current sensing already providing Kelvin connections and such.... that's what I got by having a quick look at their application notes (didn't read the datasheet, I admit)
