Active Precision Current Limiting circuit needed

[hm_fa_da]

Hi all,

I'm trying to design a precision active current limiter circuit to limit output current down to 100mA.
I saw many different circuits in net, some made by BJT transistors which are not precision enough.
I don't have much space on my PCB, so priority of the design is being simple and low size.
Do you have any suggestion for compact ICs to limit the desired current value by a setting resistor only ... ?

I found this circuit useful:


As you see it has only 2 components, so low size and compact enough, but to achieve to 100mA limitation, 12.5ohm Resistor should be used which makes high voltage drop ... (1.25V drop).

The other way is to use op-amps driving a mosfet ... but those circuits also have many components which is not good for my case.
The best solution would be some simple circuit like the picture i sent (but with lower V drop) or a compact controllable IC for current limitation ...

I'll be thankful for your suggestions.

 

[FvM]

Maximum input voltage is an important parameter. There are adjustable voltage regulators with smaller reference voltage than LM317, but only for small input voltage.

I don't see a handy solution using less than an OP, a MOSFET, a voltage reference and several passive components.

- - - Updated - - -

Analog has a number of original Linear adjustable current sources https://www.analog.com/en/parametricsearch/11402

 

[wwfeldman]

Hi all,

I'm trying to design a precision active current limiter circuit to limit output current down to 100mA.
I saw many different circuits in net, some made by BJT transistors which are not precision enough.
I don't have much space on my PCB, so priority of the design is being simple and low size.

How precise is precise?
Is the cutoff the important parameter?
Is the slope of the cutoff important?
How sharp a cutoff do you need?

Depending on how critical 100 mA is, don't use a pot, use two resistors
one to set the "coarse" limit and one to set the fine.

 

[hm_fa_da]

Maximum input voltage is an important parameter. There are adjustable voltage regulators with smaller reference voltage than LM317, but only for small input voltage.

I don't see a handy solution using less than an OP, a MOSFET, a voltage reference and several passive components.

- - - Updated - - -

Analog has a number of original Linear adjustable current sources https://www.analog.com/en/parametricsearch/11402

Thanks,

Maximum input voltage is 7.5V.
I currently found LT3092 , very suitable as respect of size and technical parameters but it's little expensive for mass production purposes.
I'm still searching more to find a better choice ...

- - - Updated - - -

How precise is precise?
Is the cutoff the important parameter?
Is the slope of the cutoff important?
How sharp a cutoff do you need?

Depending on how critical 100 mA is, don't use a pot, use two resistors
one to set the "coarse" limit and one to set the fine.

As precise as general active op-amp current limiter circuits, not like only BJT with a sensing resistor to make the 500-700mV bias voltage !
And of course, i will use fixed resistor, i don't need it be adjustable or changeable later.

 

[asdf44]

Load switches come to mind although their accuracy is usually pretty poor. But there are tons of adjustable current limit IC’s for USB and 12V automotive applications. FPF2700 is one example to get started.

Otherwise, since you discarded a common standard solution (Tl431 is another) I can’t think of others that don’t add several parts.

 

[dick_freebird]

"Current limit" is usually a fault mode and considered not
especially critical accuracy-wise; its purpose is reliability
when designed into a part.

Maybe you want a precision current source output, period.
But if this is an "applique" to limit a non-limited voltage mode
part, that's probably not applicable. Heh.

You would have to say what the output attributes (minus
the limiter) are. Your approach would be quite different
between low voltage logic, a medium-voltage amplifier and
a high voltage power switch stage. Available "auxiliary"
supplies outside the output-driver's rails? Switching speed
in the overload and the normal operating cases? Allowed
voltage drop when not-limiting? Actual -required- accuracy
(not wish list or heroic fantasy) across what range of temp,
supply, signal, "make"? Actuation delay, tolerable overshoot
magnitude and duration?

 

[KlausST]

Hi,

As precise as general active op-amp current limiter circuits, not like only BJT with a sensing resistor to make the 500-700mV bias voltage !

If an engineer asks for precision ... he usually expects values with units.
I don't know which Opamp you refer to, and which circuit you refer to.
For sure there are Opamp circuits to generate accurate (indeed I think you ask for accuracy rather than precision) current with less than 0.01% error, most of them maybe are within 2% error, but still there are a lot with more than 10% error.

Indeed almost any voltage regulator or adjustable voltage reference can be (mis-)used as current limiter.
You need at least a shunt and a Mosfet.
But the question is: can it fulfill your requirements on reaction time .. which may cause overshoot (time, value)

When your regulation system is out of regulation (too low input voltage, too high load resistance...) the output saturates (and maybe the input stage, too). There are (old) Opams, they need several 10s of milliseconds to come out of saturation.

******

Accuracy: = absolute accuracy. Accuracy error = deviation from the expected value. Let's say you expect 100mA but the circuit generates 98mA, then there is a 2mA or 2% accuracy error.

Precision: = repeatability. Let's say your circuit produces currents in the range of 97.7mA to 98.3mA with different situations like different temperature, different load resistance, different time .... then the precision error is +/-0.3mA or +/-0.3% (although the accuracy error is 2%)

Klaus

 

[crutschow]

What is the load resistance range?
What is the minimum supply voltage?

 

[crutschow]

How about the AL5809?
It's inexpensive, requires no extra parts, and can be purchase with a 100mA rating.
The data sheet show in operating down to about 2V across it.

Below is a discrete circuit using a TLV431 programmable reference that also works down to about 2V.

 

[hm_fa_da]

Thank you all for replies,

I need a constant Max 7.5V output with max 100mA Current limiter.
I already have 7.5V constant voltage regulator.

About what you guys suggested, FPF2700 is not good because the current limit starts from 0.4A and has tolerance of 20% ...

AL5809 Can be good choice because of low price and compact size (5% Accuracy), LT3092 is also good (1% Accuracy) but expensive ...
5% is max acceptable tolerance for accuracy.

But there is a question here, according to datasheets AL5809 Starts working from 2.5V and LT3092 from 1.2V.
Does it mean that i would have 2.5V (for AL5809) or 1.2V (for LT3092) minimum voltage drop before the current is clamped to limitation(100mA) i.e at 30mA Load ?


If so i have to find a way to increase the input voltage to get 7.5V output (after limiter ... ) ?

 

[crutschow]

according to datasheets AL5809 Starts working from 2.5V and LT3092 from 1.2V.
Does it mean that i would have 2.5V (for AL5809) or 1.2V (for LT3092) minimum voltage drop before the current is clamped to limitation(100mA) i.e at 30mA Load ?

Yes, that is true.

 

[asdf44]

To be specific I think it means that it isn't guaranteed to deliver the specified current until it has the required voltage.

At lower voltages it will almost certainly limit to lower currents.

AL5809 is a good part. I hadn't seen that class of 2 terminal current limit with 5% accuracy before (many are much higher).

 

[crutschow]

asdf44 : At lower voltages it will almost certainly limit to lower currents.

Yes.
Below is a graph from the AL5809 data sheet showing current vs. voltage across it.
It reaches the 100mA limit at a little below 2V.