LMR33640: Output Voltage is Affected by the Load

Hello,

It looks that one of my problems was a loose connection of the input capacitors.

I still have a problem with the maximum current that is less than 100mA.

Regards,

Marius Raducanu

Hi Marius,

Can you provide some waveforms such as the SW node, VIN at the package and VOUT? Is the breadboarded schematic have the same values shown on page 15 of the application note? What is your input voltage? Did you check to see that you are not hitting the current limit of the power supply?

Ben  

Hi Ben,

Thanks for your answer. I will add the measurements.

The capacitors that I am using in the test are 100uF. I have tried inductors in the range 2.2-10uH - used a big inductor core to be sure it does not limit the current. The output is set by the feedback resistors to -3.3V. The power supply (input) is set to 12V/1A.

The circuit works with a ~100mA load (33Ohm). The converter is not heating up.

If the load is increased to 18Ohm (~183mA), the power supply switches to current mode (1A, 4V) and the converter get damaged instantly. After this, even with the load removed, the power supply current is 1A (limited) and the converter smoked. 

Regards,

Marius Raducanu

Hi Ben,

The measurements are below.

SW:

VIN (12V):

-VOUT (-3.3V):

There measurements are for a 50Ohm load (~66mA). For a load > 100mA, the LMR33640 is getting damaged.

Regards,

Marius Raducanu

HI Marius,

Thanks for the update. I am not sure what is going on since the converter is working when the load is around ~100mA. Are you testing only the LMR33640? Based on the power conversion, where is the rest of the power going to? At Vout = -3.3V with 183mA load, the power is a little over 0.6W but the input power goes up to 12W before going down to 4W. Have you evaluated this setup with our EVM?

Ben 

Hi Ben,

Powered (12V) when the converter test circuit is connected with a 183mA load, the current goes instantly to 1A. The converter is damaged when powered and I suppose it is shorted internally making the current reach the limit of the Power Supply - all the extra power goes in the shorted converter (releasing smoke).

I have not tested the evaluation board of this converter. I wanted just to validate the circuit used in its documentation (5V/3A) before using it in a PCB design. It looks that in my tests, the maximum current is 100mA.  

Regards,

Marius Raducanu 

Hi Marius,

Just wondering, is your thermal dap also going to -3.3V? Can you provide a picture of your PCB?

Ben

Hi Ben,

Yes, we have a wire connected to the thermal pad and then connected to the -3.3V. We are using a bread board, and the converter is connected with an 8SOIC adapter (no heat sink).

I was aware of the need to release the extra heat but in this case the converter is damaged instantly (< 1s), at power up. With a load <100mA, the converter is cool when touched with a finger. In the worst case, I expect the converter, with such a low load (~180mA), to heat up before blowing up. The final PCB design will have the thermal pad soldered on a cooper plane (heat sink). The inductor is not in the picture but connected - a big inductor is used. 

Regards,

Marius

Hi Ben,

I have ordered the evaluation kit for the LMR33640. This is a positive buck converter but in the circuit with problems, it is used as an inverter (negative output) - according to snva856b.

The evaluation board is for the positive buck converter topology and for testing, I will try to modify it to work as an inverter.

Regards,

Marius Raducanu

Hi Marius,

Just to confirm, does your board have input capacitors from VIN to GND and VIN to -VOUT including the high frequency capacitor? See below:

How much output capacitance does the output have?

Ben

Hi Ben,

VIN to GND - 100uF

VIN to -3.3V - 100uF

-VOUT to GND - 100uF

Regards,

Marius

Hi Marius,

My thought is try and see if you see the same issue with the EVAL board. The other suggestion I have is to try and reduce the parasitics of your setup by bringing all the components closer to the IC itself.  

Ben

Hi Ben,

I will test the converter on the evaluation board - I will receive it Friday. The problem is that the evaluation board is for the converter topology with a positive voltage output (buck). We tested this using a bread board and the max current was ~100mA without a drop in the output voltage. At bigger loads, the output voltage dropped function of the load but have not damaged the converter.

I will try to convert the evaluation board to an inverter topology.

Regards,

Marius

Hi Marius,

It should not be too difficult to convert the evaluation board to IBB configuration. The "GND" of the board will be the -VOUT voltage and the actual VOUT on the board will be the new "GND". Below is another E2E thread showing how to configure the evaluation board into an IBB output:

https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1006191/lmr33640evm-can-not-output-negative-9v

Ben

Hi Ben,

I tested the LMR33640 +3.3V and -3.3V outputs circuits using the TI evaluation board (LMR33640ADDAEVM) and everything worked.

The conclusion is that our tests, using a bread board instead of a PCB, created problems when a load was connected.

Regards,

Marius Raducanu

Hi Marius,

Thanks for the update! I am glad everything has worked out for you. Good luck to you!

Ben

Hi Ben,

Thanks for your advice.

I will still try to find what was wrong with the bread board circuit comparing the signals and noise on the two setups.

Regards,

Marius Raducanu 

Hi Marius,

Please keep posted. Thanks,

Ben