Multiple shunts for current sensor

[PeterFW]

Hello!
My favorite device for current measuring is the TI INA226 and it works really well for me.
But the measuring range/resolution is a bit limited by the maximum shunt voltage.

I have been toying with the idea for a little module to measure over a wider range but every solution i can come up with has its drawbacks.

It seems the only feaseable solution would be to have a bunch of INA226 each with their own shunt and swith those in and out of the voltage path.

Would i be correct in that assumption?

Greetings,
Peter

[Alex Trofimov]

Look for ACS711 series. They have analog output, so you should use an ADC or use micro's ADC.
The most sensitive model is capable of measuring +/-12.5 A, with slope of 110 mV/A.
They utilize Hall's effect, so full galvanic isolation is here.
Pretty funky little beasties, I plan to use them in one of my projects.

[ovnr]

So basically several different shunts to do auto-ranging? Consider a decent ADC/AFE with PGA instead. Also, what are you trying to accomplish? The more general the question, the more useless the answer.

As for the ACS711/any other hall effect sensor: They're not suitable for anything remotely like precision work due to drift, noise, etc. If a 16-bit ADC isn't enough, a hall effect sensor is not going to improve your life.

[PeterFW]

Look for ACS711 series. They have analog output

I have two of the allregro hall effect sensors here, they are good if you just need a rough estimation and no voltage reading.
They do not cut the mustard

[PeterFW]

Hello and thanks for your reply!

So basically several different shunts to do auto-ranging? Consider a decent ADC/AFE with PGA instead. Also, what are you trying to accomplish? The more general the question, the more useless the answer.

Autoranging and manual range seting, yes.
The INA226 are really easy to use, have a nice digital interface and measure the bus voltage.

What i want to do is to measure the input vs. output of a DCDC converter, log that to an sd card with maybe 20hz resolution and generate a curve. For that i need the current and voltage.

I do not need a huge current range but 3 current ranges for different tests.

Greetings,
Peter

[rx8pilot]

I use the INA226, but have never really pushed a big range through it. Is 16bits not enough?

What minimum current?
What maximum current?
20hz is nice and slow so you can use the averaging built in as noise reduction.

[PeterFW]

I use the INA226, but have never really pushed a big range through it. Is 16bits not enough?

What minimum current?
What maximum current?

About 100µA to 30A, but... during a single test run i either am in the 100µA to 100mA range or 1 to 5A or 1 to 30A.
So during each run i do not have to change the range, so i could just etch 3 different boards and swap the "sensors" out.

Seems to be the best choice, but i wanted to ask here if someone has a different idea when using this chip.

20hz is nice and slow so you can use the averaging built in as noise reduction.

That was the idea, i love this chip since it does everything on its own and i just have to ask it for the data over i2c, such a lovely little thing

[rx8pilot]

So, is the setup to build a one-off custom tool for you to use in your own bench? If so, I would definitely build 3 PCB's each with a different sense resistor.

You could, potentially figure out a way to swap the sense resistors, but at 30A is would be easier to just have them soldered down so you can count on the kelvin connection. Have you laid out a PCB with kelvin connection before?

[PeterFW]

So, is the setup to build a one-off custom tool for you to use in your own bench? If so, I would definitely build 3 PCB's each with a different sense resistor.

Just a custom, one of for my own bench.
I will go with the seperate PCBs then, thank you!

Have you laid out a PCB with kelvin connection before?

Once, this is the layout i had been using so far. I attached a picture of the important bit of the layout.

Greetings,
Peter

[rx8pilot]

I would at least consider giving yourself some filtering options. You don't have to populate the spaces if you don't want to, but I found some benefit in filtering on the PCB.

The data sheet has an example, but at least a pad for a capacitor or two.

[PeterFW]

The data sheet has an example, but at least a pad for a capacitor or two.

My first layout had them in but i have no suitable resistors at hand.

Had not thought about just not populating/bridgeing them.
I will put them in again, thanks!

[rx8pilot]

If you never use them, nothing is lost. If you ever need them, everything is gained.

When I do protos and one-offs, I try to put contingency parts in the layout where possible. All un-used pins get fanned out in case I  discover a use for them later.



Sent from my horrible mobile....