LMR16010: Part will not regulate

Ron,

I don't believe I can narrow down the issue from the information you've given. It could be something not soldered properly, a misplaced part, an incorrect component, ESD breaking the two boards, or lots of other possibilities.

You will have to probe around and find out what part of the circuit is not working and why. Compare pin voltages and waveforms on the LMR16010 between the good board and the 2 others.

In the meantime please share the schematic, layout, and system specifications.

Thanks,

-Sam

The issue is, the device does not go into regulation when the load changes. The device, although is greatly heat sunk, gets real hot and acts like it never comes out of sleep.

The attached are traces of the output from the device as the controller pulls current to enable various loads on the PI device vs the power input, 25 volts.

The second pic shows pin 8 in yellow against the 5V output in red.

I added capacitance to no avail.  The seemingly working board the device is constantly switching on pin 8 the other 2 don’t switch but randomly.

The board was professionally assembled with pick and place equip and there are no opens or shorts.  Heavy 2oz cu design with short traces.  The IC appears to be bad and unable to operate as designed.  The load at 5 volts130mA to 498mA.

I need some insight as to why this device will not go into CCM as the working model does.  Maybe it is the bad one.  Ics have been replaced and no difference made.

Best Regards,TI LMR16010 Design.pdf

Ron Bindl

Ron,

Your schematic has feedback resistors in the hundreds of kOhms and you've added a 22nF cap on FB. We recommend FB resistors in the 10s of kOhms which matters especially in this case with the cap. This cap should not be there. Please remove this cap and test the circuit with a fresh IC (and fresh board if possible).

Layout:

• Connect the GND pin of the IC to the thermal pad GND plane under the IC.
• Connect SW through a thick trace
• Do not use thermal reliefs to connect pads to planes when there is a lot of current flowing.

-Sam

Ron,

Check the switch node. Make sure this is in stable operation. You may need a feed forward cap across the high-side feedback resistor.

Your waveform shows SW is not switching. Is the good board switching?

Try swapping ICs from the good board to the bad board. And try swapping ICs to the EVM (order a few if you haven't).

-Sam

Sam,

I can't believe this does not work.  Tried everything you suggested and still works exactly the same.  The part regulates in sleep mode ...assuming so ...and never comes out.  The one that does work and what I mean by "works" is it regulates with the PWM always running and does not get smoken hot as these stuck in what appears to be sleep mode do.  The FB resistors are under 100K and removed the 22nF.... replaced the chip ....  Makes no difference!  There is no external control pin to operate this in CCM so it appears the Sleep-mode is activated to maintain high efficiency operation but there is nothing efficient about the16010.   IT should simply work!  It is a simple switcher after all!  The load I'm playing with is only 130mA, 50 ohm load on top of the normal draw of 30mA.  The Pi draws more but can't boot up because of the bad transient response and the chip is REALLY hot!

Any further advice?

Thank you,

Ron

Hi Ron,

Can you try to swap the IC from a good board to a not regulating board and see if the faults follow the IC?

Also is the output have 1000uF capacitor in there? Have you try run the test without the 1000uF in there?

Thanks

-Arief

Hi Arief,

I am at the test phase of the system and have the one good board or should I say the one that runs in CCM, installed in my system and can't go back for time constraints.  Besides I can't risk damaging that board.  I have to move forward with what I have.  I tried removing the output capacitors but it is only worse.  The problem is the device will not go into PWM mode ... ever!  It appears to go from Sleep 20mA to DCM at 130mA but not to CCM with the PI load.  The one that work never drops down to DCM or Sleep as I remember so it too isn't working properly but at this point it stays cool and don't care.

Thanks for the help.

Ron

Hi Ron,

I try to change the EVM inductor to 22uH and the RT resistor to 49.9kohm and I can see the part regulate up to 48V with 47uF output capacitor. I loaded the output up to 250 mA then I can see the CCM mode where the frequency is 500kHz

I see on your schematic that you are using this diode

Is this true? It seems the reverse blocking voltage for this diode is 30V

If that is true, would you mind try the following as something to test:

1) Update the schottky diode with a higher blocking voltage

2) Just use 3 x 22uF output capacitor

3) Double check the saturation current of the inductor and make sure its not saturating.

3) Try to load it up to 250mA-300mA and see if it goes into CCM mode.

Thanks

-Arief

Arief,

I tried all of the above and the Diode is actually a 150V schottky at 1A device.  The inductor saturated at 1.2 Amps.  Capacitance doesn't help if lowered in general.  The part just gets smoken hot!  and the greounds are all connected properly through the planes.

Ron

For grins I wired in a LM2592HV SIMPLE SWITCHER® Power Converter 150-kHz 2-A Step-Down Voltage
Regulator in a TO220-5 PKG and changed the inductor to a 47uH etc. and it will not regulate either. With a 30mA load I get 5.0V ...no problem and add 50 ohm resistive load (100mA) and the output drops to 1V. Does not regulate!!!!! How is that possible!!!

This is making NO sense.

Ron

So we have to think out of the box. Sorry if these seem to obvious. Is there a current limit on the 36 V supply to the regulator?

Have you checked things with a scope, or just a DMM? Is your output 1 V DC or 5 V at a duty cycle?

The power supply is current limited but at a 2 amp point at 36 Volts and I double checked that.  The 5 volt output at low current 30mA drops to 1 volt flat when adding 50 ohm load directly the the output caps!  I only use a scope so I can see what is happening.  There appears to be NO switching going on.  Also, the Tab is not tied to ground in this test.  The ground is solid and the input V is also solid 36v.

There appears to be no switching and looks just like a linear regulator only worse.  When I leave the 50 ohm load across the output the regulator gets warm.

Hi,

Do you have a layout for your prototype? Also if you like i can just ship you an EVM to test the component selection.

Thanks
-Arief

Hi Arief,

Attached is a .pdf of the design layout.  If you wish. I'd like a EVM to look at.

Send to:

       Rapid-K Energy Solutions, LLC

       2103 W Parkside Ln.

       Suite 105

       Phoenix, AZ 85027

2133.TI LMR16010 Design.pdf

Hi,

I submit the request to send an EVM

Thanks
-Arief

This design in as solid as any. Great ground planes, short traces, plenty of thermal relief, etc. I have designed more power supplies than I can count and have never had one not work until now. So, this is a unique problem and it appears to be a design flaw in the TI part. There is no switching going on in this part from low current to anything above 100ma. Heat is the only result!