Measure smallish current in circuit

Thread Starter

MikeKulls

Joined Apr 4, 2016
51
I'm trying to measure a current that could vary from 50mA to 1A. I'm thinking I would just use a resistor on the negative side and an opamp to boost the voltage. But when I search for shunt all I get is big units for 50A+. If I search for small current measurement I get stuff for micro amps. Is there a shunt available for what I want? Does it have a special name? Any better way to measure that sort of current?
 

ebp

Joined Feb 8, 2018
2,332
The resistance you require will depend on how much voltage drop you can tolerate at 1 ampere. Actual "shunts" are often 50, 100 or 200 mV at full-scale, but you aren't constrained to those values. The LVR series of resistors from Vishay is a good all-round choice for this sort of thing. They are reasonably readily available, good accuracy, reasonably compact and not too expensive. You are unlikely to need a power rating of more than one watt for such small current e.g. 100 mV at 1 A is only 100 mW. If you are using surface mount, there are many choices for low-resistance resistors that will do the job well for a few tens of cents at most.

Once you have chosen your full-scale voltage "burden", you must consider the voltage you will get at the lowest current you want to measure and how op-amp errors will affect accuracy. For example, if you chose 100 mV at 1 A (100 milliohm resistor), then at 50 mA you will get 5 mV. Many low-cost op amps have input offset voltage in the range of 1 or 2 millivolts, which means the error could be as much as 40%. You would likely need to choose an amp with much lower input offset voltage - there are lots, but they are more expensive.

If you want to put the shunt in the "high side" (between the supply voltage and the load) rather than the "low side" (between the load and "ground" or circuit common) then you will need a "differential" or "instrumentation" amplifier. If you put the shunt in the low side, you need to chose one that is "rail to rail" for both input and output - will operate with inputs down to the voltage at the negative supply pin and able to produce output down that low (I'm assuming you are measuring current in a circuit with the supply voltage positive with respect to the circuit common).
 

danadak

Joined Mar 10, 2018
4,057
One additional approach is to use a processor with a differential input
A/D to get rid of any ground drop/common mode. Depends on accuracy
and resolution you want. SAR for 12 bits or less, DelSig to 20 bits.

Depending on part the DelSig can exhibit CM 100 mV beyond the rails.

See attached.

Regards, Dana.
 

Attachments

Last edited:

Thread Starter

MikeKulls

Joined Apr 4, 2016
51
Thanks everyone, accuracy isn't a big deal at the lower end of the scale. It's really just an adjustable constant current supply I want to make. The main thing is just to ensure the voltage can't be adjusted over about 750mA.

Decell, I've used the allegro devices before on a bigger scale, I would consider using them again but thought it might be overkill in this case.

I think 100mV drop would be perfectly acceptable. I'm already dropping 0.6V with a diode for reverse polarity protection.
 

ericgibbs

Joined Jan 29, 2010
12,958
hi Mike,
I guess the next point you need will be a OPA.?
So consider 600mV as a maximum input, what voltage do want out.?
E
 
The main thing is just to ensure the voltage can't be adjusted over about 750mA.
Well, that makes little sense. Voltage isn't measured in mA.

Constant voltage power supplies can be classified as CV/CL, CV/CC, CV/Fuse or CV/FBCL

The last two, I basically made up. CV is constant Voltage, CL = Current Limit, CC is constant current. I added fuse and FBCL which is Fold Back Current Limit. The last one is a wierd one.

Current limit is different than constant current. Constant current is usually more precise. A constant current source has a high output impeadance as well.

Sometimes you might classify the "power supply" as a "Voltage source" or "Current source". The difference is these both source and sink current. You can also have "dynamic loads" which may do constant current, power, resistance and power. These cab operate in all quadrants such as V+/I+; V-I-; V+/I- and V-I+.

Many supplies have "sense leads". An open sense lead may allow the voltage to get about a diode drop on each lead.

A "Source Measure Unit" is an interesting beast with capabilities to source femtoamps of current at up to say 1000 V.

That said, you need to know exactly what your trying to build and the accuracy required.

In general, a resistor is inserted into the high side of the supply and voltage sensing is done after the resistor. High currents probably require 4-terminal resistors or hall effect sensors, but in general, a non-inductive resistor.

Sensing the high side requires overcoming some challenges so it's probably best to use an IC that's has the functionality of a "High side current sense amplifier"

Now, I built a weird device that I'll call a 4-terminal current to voltage converter with +-10V biasing. I did have a two-terminal mode and a Voc (Voltage open circuit mode). the max voltage drop was a few mV. The voltages I needed were between -1.5 to 1.5 volts. Currents from 100 mA and lower in 4 decade steps full-scale at +-10V output. The output had to be analog as well as digital.

Ac performance was on the money, but the DC current measuring capability had about a 40 pA offset. Voltage Biasing, a few mV.

I used an OP amp in a current to voltage converter configuration with a current booster in the feedback loop.

It was kind of an analog SMU, but not really. Incidently, the way an SMU measures Voltage is you tell it to source or sink a few pA and measure the voltage, In order to measure current, you tell it to source 0V and measure the current. So, it comes out in the wash.
 

-live wire-

Joined Dec 22, 2017
912
Thanks everyone, accuracy isn't a big deal at the lower end of the scale. It's really just an adjustable constant current supply I want to make. The main thing is just to ensure the voltage can't be adjusted over about 750mA.

Decell, I've used the allegro devices before on a bigger scale, I would consider using them again but thought it might be overkill in this case.

I think 100mV drop would be perfectly acceptable. I'm already dropping 0.6V with a diode for reverse polarity protection.
Why not use an n-chanel mosfet? It will only turn on for the correct polarity. It will decrease power losses substantially, and allow a higher value shunt. Here is a good video that explains it how to use a mosfet like that. But I think he uses the inferior P-chanel ones.

If you can tolerate a drop of 1V, which it sounds like you can, why not use a resistor like this?
https://www.mouser.com/ProductDetai...GAEpiMZZMtlubZbdhIBIEidPY%2bx/LaatFPzklPW9sM=
It is very precise, and can handle much greater power so it will not heat up and cause problems. It is 1 ohm, meaning 1 volt = 1 amp. So even at 50mA, that's .05V. Even if there is +-1-2mV, that is only +-4% at most. It seems pretty well suited here.
 

DECELL

Joined Apr 23, 2018
96
There are devices specifically aimed at current meaurement:
Current sense amplifiers are available to suit most needs high side or low side from Toshiba, Melixis, Maxim, Zetex, TI, AD, ST, ON Infenon etc, etc. You can even get zero drift types. Current sensing for consumer electronics is a big deal- every, set top box, phone, tablet and laptop has one, you car will have a few as well!

A shunt is not the right thing here. What you need is a current sense resistor kelvin connected to the Current sense amplifier. Note that you do not need a Kelvin 4 terminal type, just connect it up that way. The volt drop can be in the milvolt range. Loads of ideas in http://www.analog.com/media/en/technical-documentation/application-notes/an105fa.pdf and made to measure http://www.ti.com/lit/ds/symlink/tps25926.pdf
 
Top