LM5118

Ignacio,

The maximum current can be larger than that on the evaluation board.  The larger FET question will be determined by the gate charge, but I have driven MOSFETs up to 8 amps without a problem.  Keep the switching frequency as low as pratical to reduce the internal heating of the LM5118 while driving larger MOSFETs, and use VCCX if possible  to reduce internal overhead of the the linear VCC regulator.

The voltage set point can be dynmacally adjusted, and this part has been used in battery charger applications.  Just be sure the compensation is appropriate for the worst case output voltage.

I'm not sure what you mean by the third question.  The current is monitored and the current limit will be at the setpoint determined by the Rsense resistor.  The user must choose the MOSFETs appropriate for this current since Rsense is usually determined by the output current.  If the part goes into current limit, the buck output drive could be monitored for a decrease in duty cycle but then the output voltage will go out of regulation also.

Regards,

Ron Crews

Here is customer's clarification on question #3:

I need to know what current is being supplied for the charging function, if I don't have to add a circuit to monitor it, that is a plus.

I don't find such a function to support the customer's need for question #3, but I posted the question in case you can suggest an alternative way to monitor the current for a charging application.

Ignacio,

You can't use the current sense current to monitor the charging current since the current sense current monitors the inductor current.  If the LM5118 is operating in buck mode, then the average inductor current is the same as the output current.  However, in Vin is less than Vout, then the average inductor current is larger than the output current and depends on the duty cycle.  You would need to put a resistor in series with the output current and monitor that current via the resistor.

Ron Crews

Do you have a spice model available for the LM5118? The customer is asking for one.

Ignacio,

I do not have a spice model for the LM5118.

Ron

My customer will using this LM5118 to charge a 12V battery.  He wants to know when he sets Vin=0 and EN=Low (i.e. shutdown mode) how much current the LM5118 will draw back through the connected output pins/circuit if the Vout has the 12V battery connected to it?  I am assuming he is somehow disconnecting the resistor divider to eliminate that path to ground.

Can you recommend the correct way to spec the FET Gate to Source voltage?  The part in the reference design is 20 V, but the datasheet doesn't indicate how to determine the gate to source voltage for the FET.

Ignacio,

The gate to source voltage just needs to be larger than the driving peak voltage plus any spikes.  If you drive an external without using VCCX on the LM5118, than the gate drive voltage is just the regulated VCC voltage minus a small drop through the gate drivers.  The VCC voltage is between 7 and 8 volts so giving it a little margin.  A MOSFET specified at 20 volts maximum gate to source voltage will have a huge margin which is good.  Just be sure the turn on threshold is well below the gate to source drive voltage.  For example, suppose you have a 5 volt output which is tied to VCCX.  Then you gate drive voltage is about 5 volts.  In this case, you must choose a MOSFET which will fully turn of with 5 volts of drive.

Regards,

Ron Crews

I am the customer having these questions.  I have a evaluation board and I have configured it for my application:

13.8 V output

Feedback resistors changed:  R8 = 267 K, R9 = 26.1 K (raised the values to limit the current draw).

I am testing how much current is consume by the charged battery when the input goes away and the circuit is disabled.  What I found is the the circuit draws 1.6 mA when "Off". I am trying to identify the path.  Our reverse voltage would be about 12.5 V.  D1 appears to have a typical reverse current around 2 mA, but there is not an obvious path to ground.  Leakage through Q2 from there would be no more than 10 uA assuming the the LO pin is at zero volts.

With VIN grounded, I measure 0.4 V at the anode of D1.  With VIN floating, I measure 1.8 V at the anode of D1.  I don't see a significant change in current whether the input is grounded or floating.  It appears their is a current leak some where in the device.  Can you determine where the current flow is and if there is anything I can do to reduce it?

Michael,

There is a connection from Vout to the LM5118.  Without any bias on the IC, there could be a path to ground through the Vout pin.  since you output is in the 12 volt range, try connecting the output to VCCX.  This will keep the part biased up and should prevent this type of problem.  However, the current draw may be more than you can tolerate since the part will be biased and operating.  Another idea is to place a resistor in the line between the output and the Vout pin on the LM5118 to limit current.  Just make sure the drop across the resistor is much less than Vout, otherwise the emulated current mode control circuit will not work.

Before you do any of this, just disconnect the actual Vout form the Vout input to the LM5118.  If the current goes away this is where the problem is located.  Then try what I suggested.

Also, If you changed the voltage divider resistors, then you have changed the error amplifier compensation.  The upper resistor is part of the compensation network.  Check for oscillations when running to make sure you still have a stable power supply.

Ron Crews

I did not see any stability issues, but I have not done extensive testing.  I noticed on the reference board schematic a capacitor that was not installed (C19) across the upper resistor of the feedback network (R8).  Was that capacitor intended to address issues like this?  If so, is there a "ballpark" value I can add?  Not having a simulation model makes it hard to tune, since it is hard to prototype the entire circuit with a real life load.

Michael,

You should not need that capacitor.  It was put there early on as a "just in case" component.  It is part of a type III compensation which is not needed normally in a buck or buck boost current mode topology.

Regards,

Ron Crews