LM3488Q Sepic Startup Problems

I have to add that the SEPIC is not working. The input voltage goes down to 3,2V during current limitation of the external power supply (I think this is the threshold to Vin_min of the LM3488Q).

The output provides 3,2V during a short time and goes then down to 0V. Normally there should be 12V.

Hi Thomas,

Can you please share your WEBENCH schematic and the layout?  if you don't want to share your design in this forum, you can email me at howard.chen@ti.com.

One note, since the design is created by WEBENCH, it's critical that the exact BOM is used as suggested by WEBENCH.

The WB file report is in the original post.

Hi Howard,

the PDF is already in the opening post. Hope this file format is suitable for you.

I only used the suggested parts for my circuit. When there was a part not available i picked another one with the same or slightly better characteristics.

Thank you for your help!

Hi Thomas,

Sorry I over looked the posted WEBENCH design report. I've reviewed the schematic and it looks ok for both the schematic and BOM. If you don't mind, could you please share your PCB layout? Can you also capture some waveform of Vin, DR, and Vout, Comp voltage and FA/SD pin. Can you try to short the Rbp resistor?

No problem.

Here is the PCB file:

I hope you can recognise which footprint belongs to which name. As it is only a prototype it was no factor to layout that the name prints are in good position.

Waveforms are quite difficult as the circuit input voltage breaks down to 3.2V and the LM3488Q quit working immediately. So every signal is either 3.2V or GND. I will try to catch some Signals during power on moment when I am in laboratory.

Thank you for your help!

I already shorted the Rbp resistor as it determines a point where the LM3488Q changes voltage supply (not sure what in detail). But with no Rbp installed (and shorted) the same problem.

I am so sorry but my USB device in the oscilloscope is not working. So i had to take literally screenshots.

Hope you can use this pictures.

Vin:

DR:

Vout:

Comp:

FA/SD:

As you can see, after connetcting to power supply, most signals go to right around 3V. Comp to 2V and FA/SD to less than 1V.

Hope this waveforms are helpfull for you!

Timebase is 1ms/Div and vertical 2V/Div. The input voltage from the power supply is 12.1V.

I found out, that the Mosfet short circuits inductance Lin to GND and does not stop this short circuit.

So the Mofet is getting hot as the current goes right through it all the time. I tested this with current limitation of external power supply @1A. The input voltage at this point is 3.5V.

Has anybody a idea why the LM3488Q does not the Gate-PWM startup?

Shutdown function at FA/SD-pin with 5V signal is working correctly. If LM3488Q is shut down, there is no short circuit or extreme high input current. Seems like the problem is that either the LM3488Q stops working during high level at gate pin oder low level at gate pin. If its during high level the circuit creates a short circuit and when its during low level the device is not working.

I am not sure but maybe it has something to do with EMC problems. During shutdown test I attached an open ended 100R restistor to FA/SD pin. Only with it one-pined attached the device startet working.

After removing it it was not working anymore. Again attached it was still not working. Sounds very strange to me.

After more testing I got the strange situation that the mosfet stays in off-state.

Seems like the LM3488Q quits work either in mosfet-on-state (short circuit) oder mosfet-out-state (no output voltage).

I am really looking forward to your answer!

edit: yes its definitely a EMC problem. I can change the state of LM3488 (permanent mosfet-on or permanent mosfet-off) by simply touching the mosfet or LM3488. Has anybody a idea what is here the problem?

Hi Thomas,

Looking at your layout, there is too much inductance and it needs to be fixed.  There has to be some ground plane to help with thermal and as well as to minimize the high current loops. 

I can send you the source of the LM3481(In Altium) as a guidance. You can reach me at:  howard.chen@ti.com

Thank you! I have sent you an e-mail.