LDO current limiter

[--Oz--]

The lm317 is an adjustable linear voltage regulator and typically connected like pic 1.
It can also be connected to provide constant current and works very well and only two components, pic 2.

The LM317 is a 1.5A to-220 part, I want a 1A C.C. circuit, so using a 1.25 ohm resistor works, but it requires 3.3V total dropout voltage across both parts.

I am looking for a dropout voltage of 2.5V max (hopefully lower) at 1A,
something like +/-25mA C.C. regulation,
Vin minimum 5V (hopefully more),
Through hole for dead bug wiring (no pcb),
low cost, low component count, TIA!

[fourfathom]

How precise and stable do you need?  A simple two-transistor, two-resistor current limiter can have a sub-1V dropout, but the limiting will be soft and the tempco and precision may be found wanting.

[SiliconWizard]

https://www.onsemi.com/products/power-management/linear-regulators-ldo/ncp692

[--Oz--]

How precise and stable do you need?  A simple two-transistor, two-resistor current limiter can have a sub-1V dropout, but the limiting will be soft and the tempco and precision may be found wanting.

Good question. I am thinking of +/-25mA of the 1A C.C. mode.
Do you have a link to that circuit?

[--Oz--]

https://www.onsemi.com/products/power-management/linear-regulators-ldo/ncp692

Looking at the DS, it has current limiting, not really a C.C. mode. I looked for applications for the chip and there is none, just typical C.V. circuit.
And, I plan on this to be dissipating the heat, so a little qfn package is not ideal, plus needs a circuit board, I was thinking through hole and something like to-220 so i can burn some watts (at least 5W or more), I will add to my 1st post. Thanks for the reply.

[BigBoss]

I believe you're looking for something like that
It has its own current limiting input between two pins and any dropout does not occur at the output.
https://www.analog.com/en/products/lt3033.html#product-overview

[ledtester]

You can use any linear regulator as a constant current source -- see this article:

https://www.maximintegrated.com/en/design/technical-documents/app-notes/4/4404.html

So then the challenge is to find one with:

- low dropout voltage
- low output voltage
- enough current capacity
- high enough Vin-GND voltage

Here's a list of some in-stock devices at Mouser which might work for you:

https://mou.sr/3dILUCS

[Colt45]

1117?

An LDO with lower reference voltage would be better though, i suppose

[--Oz--]

You can use any linear regulator as a constant current source -- see this article:

https://www.maximintegrated.com/en/design/technical-documents/app-notes/4/4404.html

So then the challenge is to find one with:

- low dropout voltage
- low output voltage
- enough current capacity
- high enough Vin-GND voltage

Here's a list of some in-stock devices at Mouser which might work for you:

https://mou.sr/3dILUCS

If that is the case, that is great news. I am used to adjustable regulators (linear & switching) having a Vfb ref voltage referenced to ground, so this would seem it would not CC, hope I am wrong, haha. Where this lm317 regulates Vout pin to the adj-pin to 1.25v, most regulators I have used with is not common. This is where I think it makes it easy to do constant current, and the majority cannot, hope I am wrong.  And I would expect more DS showing this simple one resistor CC mode. Will test it out on Monday. Thanks!

[Zero999]

If sinking is acceptable and there isn't a tight budget, the LT3080 can do this.
Here's a circuit I designed for driving LEDs in another thread.

https://www.eevblog.com/forum/chat/led-lighting-(350-ma)-is-there-any-standard-for-of-leds-is-a-string/msg4356295/#msg4356295
https://www.analog.com/media/en/technical-documentation/data-sheets/3080fc.pdf

[magic]

If you just want an LM317 with lower dropout, there are LM1117 or LM1085/LD1085 for higher currents. Both are 1~1.5V dropout, depending on load conditions and die temperature.

Beware that they may require some output capacitance for stability, perhaps also in CC mode. I think you may get away with a capacitor in parallel with the current sense resistor rather than one going from OUT to ground, which would compromise CC dynamic performance.

[fourfathom]

How precise and stable do you need?  A simple two-transistor, two-resistor current limiter can have a sub-1V dropout, but the limiting will be soft and the tempco and precision may be found wanting.

Good question. I am thinking of +/-25mA of the 1A C.C. mode.
Do you have a link to that circuit?

I was thinking of a two-NPN circuit, but it either has a larger voltage burden or it draws more quiescent current that i would prefer.  So here's a similar topology using a PMOS pass transistor and a PNP sense transistor.  The current-limit depends on V_BE of the PNP transistor so that variation will be about 2mV per degree C (giving about 3.3mA per deg C), and the initial value will vary a little.  So it sounds like you need more stability.

But the voltage burden (dropout) is quite good, under 0.6V until the current hits that 1A limit.

But if you can find an IC solution with the specs you need that's probably a better choice.

BTW, I just chose the first reasonable FET I could find in the LTSpice library.

[SiliconWizard]

https://www.onsemi.com/products/power-management/linear-regulators-ldo/ncp692

Looking at the DS, it has current limiting, not really a C.C. mode. I looked for applications for the chip and there is none, just typical C.V. circuit.
And, I plan on this to be dissipating the heat, so a little qfn package is not ideal, plus needs a circuit board, I was thinking through hole and something like to-220 so i can burn some watts (at least 5W or more), I will add to my 1st post. Thanks for the reply.

I didn't suggest it for its internal current limiting (which is not configurable anyway), but for its very low drop-out voltage @1A.

[--Oz--]

If you just want an LM317 with lower dropout, there are LM1117 or LM1085/LD1085 for higher currents. Both are 1~1.5V dropout, depending on load conditions and die temperature.

Beware that they may require some output capacitance for stability, perhaps also in CC mode. I think you may get away with a capacitor in parallel with the current sense resistor rather than one going from OUT to ground, which would compromise CC dynamic performance.

You the man! haha, looking at the datasheet, LDO 3A to-220 and has a full 1 volt less dropout voltage than the LM317, has the CC mode mentioned in the DS, only 50 cents each on Ali, ticks all the boxes, this is a perfect part for my application. Thank you!

And thanks for everybody's replies!

[mariush]

Yeah, though make sure they're not counterfeit or renamed chips.

Another IC you may want to check is LM2941 ... should have around 0.5v dropout voltage up to 1A of current, for normal temperature ranges (may go up to 1v in extreme temperatures)
Also LM2940.

Though just like 1117 regulators, it can be picky about output capacitors and their ESR - a capacitor with ESR between 0.1 ohm ... 1 ohm is recommended.

[magic]

only 50 cents each on Ali

That's not a lot of money to risk on Chinese roulette, but this is what it is when you buy from Ali.

It could be perfectly good and unused surplus parts, or recycled chips from scrap boards, or a Chinese clone which is maybe good enough, or some different LDO with fake markings which is maybe not good enough (think LM1117 designed for 0.8A), or LM317 with fake markings, or a transistor with fake markings, or hell knows what else. You may get anything.

LD1085 is cheap enough that it doesn't cost a lot to buy it locally from a legit IC distributor. Delivery is faster too.

[magic]

Regarding capacitors, LD1085 datasheet suggests that the chip should be stable with infinite capacitance and 1.25Ω ESR.

This means that it may also prove stable with just a 1.25Ω resistor from OUT to ADJ, assuming that OUT-ADJ capacitance is good enough to stabilize it and there is no need for OUT-GND.

[--Oz--]

only 50 cents each on Ali

That's not a lot of money to risk on Chinese roulette, but this is what it is when you buy from Ali.

It could be perfectly good and unused surplus parts, or recycled chips from scrap boards, or a Chinese clone which is maybe good enough, or some different LDO with fake markings which is maybe not good enough (think LM1117 designed for 0.8A), or LM317 with fake markings, or a transistor with fake markings, or hell knows what else. You may get anything.

LD1085 is cheap enough that it doesn't cost a lot to buy it locally from a legit IC distributor. Delivery is faster too.

I have bought quite a lot of fets, IC's, leds, modules, etc from Ali (around 200 orders) and have only really got one fake fet that I can tell from my applications.
It was a IRFB4227 200V fet, but avalanched at 74V, I simply use a 10V ps connected to a inductor, then fet to ground, add a square wave to the gate and read the avalanche voltage, no questions on getting my money back.
" alt="" class="bbc_img" />

My most risky purchase was a server Xeon E5-2678v3 12core/24thread cpu, MB, 32GB ram for $277, its been great for almost three years now.

[--Oz--]

I found this interesting circuit to make this adjustable, buy has a few cons.
Too many parts and req negative supply.
https://electronics.stackexchange.com/questions/215418/smartest-way-to-use-current-limit-using-lm317

[--Oz--]

You can use any linear regulator as a constant current source -- see this article:

https://www.maximintegrated.com/en/design/technical-documents/app-notes/4/4404.html

So then the challenge is to find one with:

- low dropout voltage
- low output voltage
- enough current capacity
- high enough Vin-GND voltage

Here's a list of some in-stock devices at Mouser which might work for you:

https://mou.sr/3dILUCS

Out of curiosity of your link info, I found some LDO's and they all have the Vfb reference to ground, connected them up as normal to verify they were was working (they are brand new).
Then tried the info in your link, none of them worked, I think its because the way most regulators Vref is tied to ground, where the lm317 and a few others are not.

I tried these three
MIC49150WR
MIC29152WU
MIC29502BU

They simplify would not regulate, tried 100mA and 1A shunt resistors, Thoughts?

[flowib]

LM334 plus 2SC6144 may be what your looking for.

The 334 can be used to precicely control a NPN transistor. And the 334 has a rather low compliance voltage of 1V.

If you are looking for a constant voltage supply with LDO properties and current limiting, the MIC5156 is available in DIP8 and does 1.25-30V

[ledtester]

Out of curiosity of your link info, I found some LDO's and they all have the Vfb reference to ground, connected them up as normal to verify they were was working (they are brand new).
Then tried the info in your link, none of them worked, I think its because the way most regulators Vref is tied to ground, where the lm317 and a few others are not.

I tried these three
MIC49150WR
MIC29152WU
MIC29502BU

They simplify would not regulate, tried 100mA and 1A shunt resistors, Thoughts?

What exactly is your circuit? You might be running into this caveat mentioned in the Microchip paper:

There is an important requirement for this circuit: the minimum voltage required for proper operation (2.5V) must be provided between the IN and GND terminals. To satisfy that condition, choose an ROUT value that allows 2.5V to 5.5V between IN and GND. When driving a load of 100Ω maximum with VCC at 5V, for example, the device functions properly with ROUT above 60Ω. That value allows a maximum programmable current of 1.5V/60Ω = 25mA. Voltage across the device then equals the minimum allowed: 5V - (25mA × 100Ω) = 2.5V. This IC can source up to 500mA.

[Zero999]

LM334 plus 2SC6144 may be what your looking for.

The 334 can be used to precicely control a NPN transistor. And the 334 has a rather low compliance voltage of 1V.

If you are looking for a constant voltage supply with LDO properties and current limiting, the MIC5156 is available in DIP8 and does 1.25-30V

Adding a transistor increases the voltage headroom requirement. Bear in mind the LM334 has a positive temperature coefficient. It can be countered by adding a diode, but that increases the voltage headroom further.
https://www.ti.com/lit/ds/snvs746e/snvs746e.pdf

[magic]

Not a problem, because the 1.7V minimum applies to the supply voltage required for the circuit to work. OP's supply is 5V.

OP's problem was dropout voltage, which is laughably small for this circuit: just the PNP saturation voltage plus approximately 64mV across the current sense resistor.

A potential problem is the "approximately" part, because the exact voltage regulated by the chip increases 0.2mV/°C from 64mV at 25°C.

[--Oz--]

What exactly is your circuit? You might be running into this caveat mentioned in the Microchip paper:

There is an important requirement for this circuit: the minimum voltage required for proper operation (2.5V) must be provided between the IN and GND terminals. To satisfy that condition, choose an ROUT value that allows 2.5V to 5.5V between IN and GND. When driving a load of 100Ω maximum with VCC at 5V, for example, the device functions properly with ROUT above 60Ω. That value allows a maximum programmable current of 1.5V/60Ω = 25mA. Voltage across the device then equals the minimum allowed: 5V - (25mA × 100Ω) = 2.5V. This IC can source up to 500mA.

My circuit I tried all three of the MIC regulators I posted above, I first connected them as normal CV and tested to make sure they worked normal mode, they are all brand need, just checking for piece of mind. Then I connected like the picture below. I set the resistor to CC at 100mA first, then 1A (changing the shunt resistor).

I first tried Vin=5V and played with a variable 2W 10 turn pot (testing the 100mA), it just would not CC anywhere. Retested at 1A shunt resistor, I tried up to Vin=10v and no luck.
I then tried just shorting the CC output to ground and played with Vin, same, no CC, this method works with the lm317.

I reconfigured it to normal VC mode and working, again to have piece of mind.

Thanks again for the link.

The LD1085 should fit my application, but am still curious why 95% of linear regulators dont show CC mode in their datasheet (like the lm317 does), weird when it could sell more parts.