Post Go back to editing

# LTC3872 current limit

I am trying to get an output voltage of ~34.8 V with the current limited to ~20 mA. I did all the calculations from the datasheet and decided to use the Rsense resistor method because it fluctuates less with temperature than RDS (on) does. It doesn't say how to calculate the values for the resistor and capacitor at the Ith pin or Run/SS pin, so I used values from a sample circuit. The current is not limited though. At a light load, it allows it to draw >4.5 A and my output voltage is <5.0 V. How do I make the current limit work?

Parents
• You have connected the SW of LTC3872 node to the wrong point ;-)

kr

Markus

• Hello again,

Ok, had not my lightest moment with my first answer. deeper reading of the ds helps here as well. Your schematic is correct so far as you use a discrete shunt instead of RDS,on.

But your complete setup limits the deliverable output power capability not your output current.

And your circuitry has no margin to cover slight changes in Vin or Iout.

Intention?

You should increase your load resistors towards 2k or more to establish Iout<20mA. Then the circuit will reach the Vout (after a long time, as output power is limited)

kr

Markus

• Hello,

try this setup:

No changes in the ITH network yet, and your initial 270uH seem too big for me. IPRG=VIN and then the circuit can deliver the required output power.

I reduced startup time as I pulsed the Iout between 100uA (for startup) and 20mA (after Vout has settled).

kr

Markus

• Hi Markus,

Thanks for responding!

I have a few more questions and concerns. The ds says that one of the features is an Adjustable Current Limit and then it says to select your MOSFET (or in my case, Rsense) based on the resistance value to check the maximum output current as per this formula:

I am going to have a variable load, so I want to set a current limit independent of the load resistors. I only included the load resistors in the circuit because they had load resistors in the sample circuit. It says the device should have excellent load regulation so the output shouldn't depend on the load, should it? It also doesn't have that anywhere in the ds or any of the equations.

With IPRG grounded, at my duty cycle, VSENSE should be ~ 76 mV. If I tie IPRG to Vin, it will be ~220  mV which will increase my output current by quite a lot.

Regarding the inductor, the ds says to choose a value of X between 0.2 and 0.4. Then I used my max output current to solve these two equations:

This gives me a range of L values between ~170uH and ~350uH, so with my 270uH, I get X = ~0.26.

With these inductor and current values, the device is pulling Iin(max) of about 129 mA at 6V, which gives P = IV = ~770 mW. Then the output is ~34.8 V at 20 mA which is P = ~718 mW, so it should have enough power, depending on its efficiency.

Please let me know if that makes sense and where I should go from here.

Cheers,

Jamie

• Hello,

first of all, the LTC3872 is a boost converter which is working in CV mode (constant voltage) Vout.

This means as long as your circuit is able to deliver the desired output power its Vout of 34,8V will be generated resp. stable, no matter if load current is 1uA, 1mA or 10mA.This is meant with 'good load regulation'

Adjustable Current limit was implemented to enable usage of different sized FET with different RDS,on.

As you have decided to use a discrete sense resistor you can use this adjustable current limit to tailor your sense resistor for size/power losses/accuracy.

The possible output power is defined by your sense resistor in combination with the IPRG setting. As stated in the formula.

The upper limit will be given by the chosen FET and the inductor size, ie the LTC3872 should not deliver/enable more current than the FET or inductor might carry. The configured upper limit should be larger than your maximum output power for your application, so the circuit can deliver always a stable output voltage.

Your last assumption is not correct as you neglected the switching characteristic/triangle shape of input current. This is where ripple current deltaIL comesinto play. 129mA is the peak current, this is not a dc current. As you have the LTSpice schematic you can look at all signals easily.

So if you have a max output power of ~718mW i would at least tailor your circuit for 1W, better more. This allows faster/better regulation margin in case of load transients.

And on top of that my recommendation would be to

1. use an integrated boost regulator like our LT8330 or LT8570, these have the power switch integrated and allow for an easier design. They're are capable of your desired output power.

https://www.analog.com/en/technical-articles/how-to-select-a-boost-regulator-controller-ic-and-use-ltspice.html

kind regards

Markus

• Hello,

first of all, the LTC3872 is a boost converter which is working in CV mode (constant voltage) Vout.

This means as long as your circuit is able to deliver the desired output power its Vout of 34,8V will be generated resp. stable, no matter if load current is 1uA, 1mA or 10mA.This is meant with 'good load regulation'

Adjustable Current limit was implemented to enable usage of different sized FET with different RDS,on.

As you have decided to use a discrete sense resistor you can use this adjustable current limit to tailor your sense resistor for size/power losses/accuracy.

The possible output power is defined by your sense resistor in combination with the IPRG setting. As stated in the formula.

The upper limit will be given by the chosen FET and the inductor size, ie the LTC3872 should not deliver/enable more current than the FET or inductor might carry. The configured upper limit should be larger than your maximum output power for your application, so the circuit can deliver always a stable output voltage.

Your last assumption is not correct as you neglected the switching characteristic/triangle shape of input current. This is where ripple current deltaIL comesinto play. 129mA is the peak current, this is not a dc current. As you have the LTSpice schematic you can look at all signals easily.

So if you have a max output power of ~718mW i would at least tailor your circuit for 1W, better more. This allows faster/better regulation margin in case of load transients.

And on top of that my recommendation would be to

1. use an integrated boost regulator like our LT8330 or LT8570, these have the power switch integrated and allow for an easier design. They're are capable of your desired output power.

https://www.analog.com/en/technical-articles/how-to-select-a-boost-regulator-controller-ic-and-use-ltspice.html

kind regards

Markus

Children
• Hi Markus,

Adjustable Current limit was implemented to enable usage of different sized FET with different RDS,on.

So, which current is limited by the Adjustable Current Limit? I would like to limit the output current while keeping the output voltage constant. But it seems like maybe this limit is an input current limit? And then if the load draws more current the output voltage will drop and the load will be allowed to pull as much current as it wants.

The possible output power is defined by your sense resistor in combination with the IPRG setting. As stated in the formula.

Do you mean this formula?

Isn't "Io (max)" the max output current, not power?

Your last assumption is not correct as you neglected the switching characteristic/triangle shape of input current. This is where ripple current deltaIL comesinto play. 129mA is the peak current, this is not a dc current. As you have the LTSpice schematic you can look at all signals easily.

Isn't Iin (peak) the peak current? I thought (Iin (max) + dI/2 = Iin (peak) )

And on top of that my recommendation would be to

1. use an integrated boost regulator like our LT8330 or LT8570, these have the power switch integrated and allow for an easier design. They're are capable of your desired output power.

Can any of the boost controllers or regulators actually limit the output current? I need something that supplies a steady voltage for any load between 0-20 mA. Then if the load tries to pull more than 20 mA, I still need the voltage to remain steady but the current to be limited to 20 mA.
I was using a boost converter and LDO (to limit current and provide line and load regulation) in a similar circuit but I saw that some of these ICs say they have adjustable/programmable current limits, ripple reduction, and internal LDOs so I thought it would be great if I could use one IC instead of two.

Cheers,

Jamie

• Hello,

starting withan answer to your first and last question/statement: A standard boost regulator has no output current limit the way you're looking for. Normally the phrase 'current limit' refers to the current mode topology where either transistor or inductor current is limited/monitored for proper circuit operation. For sure this limit 'translates' to the available output power, but It's not a specified output current limit as you requested. You should have a look into our LED driver portfolio, these parts have two current limits in their control loop and feature a constant voltage/constant current operation.

The block diagram of the LT3519 for example does show these two limits:

Blue circle shows the 'classic' current limit of the power transistor. red circle shows the additional current limit for constant current regulation.