Why DC/DC with external mosfet have limited current rating?

Question

Why does DC/DC converters with external mosfet have current limit?

From my understanding, for components with an integrated MOSFET, the limit comes from the components (less performance, the Rdson is often higher), if a too high current flows through the component it risks to heat too much and exceed the junction temperature.

But if I use a DC / DC "controller" which controls an external MOSFET, no current flows through the component itself. It would then be enough to choose a mosfet with a high current Id

I take as an example the TLE6389, the typical application is given in the datasheet

My hypotheses: Does the current limitation come from the current measurement function (the internal logic works up to a current value) or does the limit come from the ability to control a more powerful mosfet (the GDRV pin which controls the gate of the mosfet provides an output between 6 and 8 V)?

Answer 1

The reason behind the current limit may be a business reason, rather than an engineering reason.

The design may have been commissioned by a specific customer to a specific brief (accompanied by a large and lucrative order) - probably in the automotive industry judging by the product validation specification. The parameters in the brief then became an essential part of the product specification. Given that Infineon is a German company, the original customer is likely to have been a German automotive manufacturer. Beyond that, your guess is as good as mine.

In the same way, the first microprocessor Intel ever made, arose from a commission by a customer company, for the desk calculator market.

How the company implements that brief is less important to the customer : in this case, the use of an external Isense resistor was presumably acceptable, and at least in theory, allows any other customer to meet other current limit specifications by changing the sense resistor. The internal logic merely senses the voltage across that resistor : it has no way of knowing what the value of that resistor is.

@DKNguyen is correct that the device has gate drive limitations - both in limited current, and fairly slow rise/fall times - and these are adequate for the original brief with the original recommended MOSFET. It is good that these are clearly stated in the datasheet, but that datasheet is primarily written to hand hold the customer in implementing the original brief.

Using the device for other purposes (e.g. a 10A supply) will likely be possible as long as you design around all the other stated limitations (e.g. choose a FET with acceptable gate capacitance, and a suitable inductor).

But I would not expect Infineon to endorse or support such a design - YOU ARE ON YOUR OWN especially if your design is safety critical - unless you get a specific waiver from Infineon after reviewing your design with one of their FAEs. (Contacting them would be a good idea for any serious use of the product ... they may at least confirm (or correct) this hypothesis. But I doubt they'll rewiew your design for less than a hundred thousand units.) This may or may not not matter for your purposes but should be kept in mind if using it outside the datasheet brief.

Answer 2

The stated limit of 2.3A is not a limit of what the controller can withstand, it is a limit that is enforced by the controller to protect the rest of the circuit. The 2.3A value is based on the reference circuit.

You can change this limit. See section 7.10.2 in the data sheet. It says, "The peak current which the buck converter is able to provide is determined by the peak current limit threshold voltage VLIM and the sense resistor RSENSE." It then gives formula to calculate a new RSENSE for a new limit value.

If you want to find the expected peak current for a given design, you might want to simulate the circuit. Or use a handy calculator like the buck calculator at http://schmidt-walter-schaltnetzteile.de/smps_e/smps_e.html If you find you need a higher current limit based on the sim then adjust RSENSE.

Answer 3

Gate drive current capability of a controller IC is not unlimited so you cannot just toss an infinitely big MOSFET onto the output. There may be other limiting factors but this is the only one I know of.

EDIT: It's been pointed out that specifying a current instead of a gate charge limit would be a roundabout way of doing things since "there is no reliable way to correlate gate charge to current capacity".

It is true that there is no reliable correlation. So where did 2.3A current rating on the front page then? Sure, there is no reliable correlation for engineering purposes, but you can ball park things based on existing technology, if your objective is marketing purposes.

Consider that similar current ratings are given for MOSFETs on the front-page, but you can't actually achieve them in practice (at least I never can), while \$R_{dson}\$, the useful parameter is buried deeper in the datasheet.

Instead of a gate charge limit (which would require knowledge of too many unknowns about the operating conditions of the end user MOSFET) , these DC-DC controllers tend to bury a gate drive current, not a gate charge limit deeper in the datasheet.

Answer 4

Devices that control external MOSFETs can have a wide range of applications, each with it's own output voltage and current requirements.

There may be a requirement that if the load short circuits, the output current may not exceed some stated maximum which can protect the PCB itself (from excessive current damaging the power path), the MOSFET (which will have some maximum allowable current) and the inductor (to protect it from excessive \$I^2R\$ copper losses.

We can set current limits externally on some parts that have an internal current limit (which is usually for thermal protection although not always).

Setting a current limit that is no higher than the requirement of the particular application also permits the use of smaller inductors than might otherwise be necessary (because the limit will prevent saturation).

The maximum current limit in your application is necessary as the body diode of the MOSFET will still conduct even if the MOSFET is turned off.

Answer 5

why DC/DC converters with external mosfet have current limit ?

It's sensible when designing any type of power converter to restrict current to prevent overload conditions creating a possible fire risk.

if a too high current flows through the component it risks to heat too much and exceed the junction temperature.

Many decent regulators will monitor their own junction temperature and shut down - current limiting is used for over-current scenarios when the temperature of components are not necessarily that warm but, the circuit could be dumping power into a load that creates a fire risk.

There are fusing limits that need to be prevented. There are also limits to the current applied when using inductors (such as \$L_1\$ in your picture). You don't want it to magnetically saturate.

So, the chip in question allows for using an \$R_{SENSE}\$ resistor and this needn't be used of course but, not using it is not usually a good idea.

Answer 6

First, current limit is a built-in protection for your external FET in case the FET experience dangerous level of current during some rare transient event. For example, if your FET is rated 20A and your normal operational current through the sense resistor is 10A, looks like you can use the 20A FET. But who will protect you if the sense resistor current suddenly go up to 30A under some transient conditions or fault condition? There is no fuse in a DC/DC converter board to protect you.

Secondly, voltage across sense resistor is amplified by a fixed gain amplifier internally. If the voltage is very large, it could saturate the internal amplifier. Sure you can always choose a smaller resistor to allow less voltage drop across the sense resistor, but also current limit has an accuracy requirement. The manufacture don't want to risk going out of spec by not specifying what kinds of sense resistor and current limit you can apply.

Third, the maximum load a DC/DC controller can handle is closely related to the stability requirement of the controller. The controller become less stable(perform less well for a step load change or supply change) when load current is very high. That being said, you can predict the maximum current going through the sense resistor by using your worst case vin/vout ratio. What is your reason for having current higher than the current limit? If that is because that your load go up, then I would say a FET change is not enough since the controller may have stability issue at higher than specification load.

Answer 7

Actually I know this device pretty well and would like to help you: Autmotive application does not oppose a limit on how and where you can use this device, it just determines the way it is qualified and reliability is checked (have a look at AEC-Q100).

The first obvious reason for why the controller can't handle arbitary load currents has already been mentioned. It is the gate driver's limited current/voltage driving capability that will prevent you from using MOSFETs with lower RDSON because these will come with higher gate charge...

The second reason has not been mentioned so far but it is a very important one. This controller uses the so called "peak current mode" control scheme. It is the main reason why there is a sense resistor which is feeding back the switch current into the controller. You can read up on the principles of "peak current mode control" in some appnotes or in Erickson's "Fundamentals of Power Electronics". The benefit of this method is that you end up with a first order control to output transfer function which is easier to compensate. Secondly you need the current measurement anyhow in case you want to protect your device and switch against short circuit conditions. This control method will however restrict the possible inductor values you can choose vs. the sense resistor and the set output voltage. There is a artifical compensation slope inside the controller (it is also depicted in the datasheet) that needs to be "matched" to the current slopes of your measured switch / inductor current.

In section 7.10.1 you can see this restriction on the chooseable inductor. When you do the math for your desired operating conditions you will see that when you choose the inductor according to the given formula, your ripple will run out of "headroom" to the upper current limit.