LM5118EVAL: 800mV Output

Hi Garrett,

As you observed UVLO less than 1.23V, it means "a current limit fault exists for more than 256 clock cycles, the regulator will enter a hiccup mode of current limiting and the UVLO pin will be pulled low by an internal switch".

And I would suggest you to add some bulk cap for both input and output.

Best Regards,

Feng Ji

This is all unloaded, so I'm not sure where the current limit fault would be coming from. I did try a 50mA load in case there was a minimum load and still saw the same behavior. As previously stated, I was able to get the UVLO pin above the 1.23V threshold and didn't see any change. I did this by bringing the input voltage up to 34V which is still within the input voltage range I am designing for and saw the same 800mV output. This makes be think that the UVLO pin isn't getting pulled low, or it would stay low even at a 34V input, right? My UVLO resistors are 88.7k for R1 and 6.8k for R3. This should make the UVLO pin above 1.23V until the input is below about 17.2V. I also tried removing R3 so that the UVLO pin is just pulled up to the input voltage and saw no change on the output.

We are designing for a small solution size, so I don't want to test with bulk capacitance that won't be in the final design.

Thanks

Garrett

HI Garrett,

can you probe the UVLO and SS pin.

Do that have a static signal or do they ramp up and down. 

If they are ramping up and down - esp. the SS pin then it indicates that there had been an error detected.

This is most times an over current detection - e.g. happening due to an to high inrush current.

So you can then also check the voltage on the Current sense resistor.

Best regards,

 Stefan

The voltage on UVLO, SS, and the current sense resistor are all ramping up and down. I measured the inrush current and found it to be 20A, so I thought maybe a current fault on start up. I then added a 3 ohm resistor in series with Vin which brought the inrush current down to 4A, but I still am seeing the same problems with these voltages ramping up and down.

Thanks

Garrett

Hi Garrett,

check the voltage on the current sense resistor. It should increase with each switching cycle.

If it is already to high at the beginning you may have a short or damaged inductor.

Now you can increase the softstart cap to increase the output voltage slower so that the current limit is not hit.

You can also check by math:

Vout * Cout = Imax_out + tss

Imax_out is the max output current the PowerSupply is designed for, if you have already a constant load at the output you need to subract the current for that, because this is the current needed to charge the output cap.

Best regards,

 Stefan

I put the original high side FET back and now this is what I see on the HO pin. It just stays at 1V after it ramps up initially.

I don't see anything voltage across the current sense resistor.

I am a little confused by the equation you gave. If I do 25V output with 30uF of output capacitance then the left side of the equation is 0.00075. By PowerSupply do you mean my benchtop supply of the max current that the buck boost I am designing will output? Either way, when you subtract that (5A or 1A) to solve for tss you get a negative time. Can you clarify what the Imax_out is supposed to be?

Thanks

Garrett

Hi Garrett,

just to ensure the voltage level to measure at the current sense is below: 300mV depending on the operation mode:

To the equation:

there was a type above this should be:

Vout * Cout = (Imax_out - ILoad,Start)  * tss

Vout: output voltage of the power stage build with the LM5118

Cout: output capacitor at the output of the power stage build with the LM5118

Iout_max: max output current of the power stage build with the LM5118 

ILoad,Start: load current (from the load connected to the power stage) during startup

Best regards,

 Stefan

Is the Imax_out the current through the output capacitors? I'm still a little confused by the terminology. 

With testing unloaded (ILoad,Start = 0) and a 2.5ms soft start, I calculated Imax_out to be 300mA. I increased the soft start time to 12.5ms just to be sure, but I didn't see any change

Hi Garrett,

sorry for the confusion.

The Imax_out is given through over current protection of the power stage build with the LM5118

See also the Datasheet section:

7.3.5 Current Limit

8.2.2.4 R13 = RSENSE

So measuring the voltage at the current sense might help to see if the issue is coming from that point.

Best regards,

 Stefan

I don't see any voltage across the current sense resistor. This is what I see on startup:

Blue is voltage across CS resistor and purple is Vin.

I got my I1peak to be 1.82A and my I2peak to be 3.48A from equation 15 in the data sheet from section 8.2.2.3 Inductor Selection. 

I then calculated the 50mOhm resistance (rounded to a standard value) in 8.2.2.4, which is what I have on the board. This is also the recommended value by WEBENCH and the Quick Start calculator.

Thanks

Garrett

HI Garrett,

does the converter start to operate at all - do you see toggling signals on the MOSFET gate? Do you see the SW node going up and down?

If not check the system for the required signals:

EN

UVLO

VIN

Other soldering issues.

Best regards,

 Stefan

Here are some pictures of some of those signals at startup. The blue is the specified signal and the yellow is Vin.

EN:

UVLO: Hard to get start up picture because of smaller time scale needed.

HO:

LO:

SW1:

SW2:

It looks like LO tries to start switching but then stops immediately. The UVLO pin looks like it is operating in a fault condition, but I'm not sure what it would be other than over current. Is it possible I need a different current sense resistor value?

Soldering looks good. I tried reflowing everything just in case, but didn't see any changes.

Thanks

Garrett

Hi Garrett,

some points to your measurement results:

- how can UVLO ramp up and down if VIN is stable and there is just a Resistor divider there?

So something is wrong here - it might be that this is the SS pin - please double check

Best regards,

 Stefan

That's the main thing that I have been confused by. This is what the SS pin (blue) looks like with Vin (yellow). 

Thanks

Garrett

Hi Garrett,

OK, this looks like a SS pin signal where the controller continuously restarts.

Still leaves the strange behavior on the Enable input.

Can you try to drive this signal with a power supply to a level of ~ 3 V and check what happens.

Note: would be good if you limit a the current to a few 10mA.

Please check if the current limit gets activated at the source of the Enable

Best Regards,

 Stefan

Hi Garrett,

just checked again. The UVLO gets pulled low if the overcurrent is triggered.

So you please check again the value of the current sense resistor.

Best regards,

 Stefan

Hi Stefan,

One of my coworkers has been working on a different solution and he is getting good results, so we are going to move forward with his design. Thanks for all of your help with debugging my problems.

Garrett