BQ24105: Heating issues

Hi Ion,

This is not normal.  I am unable to install the EAGLE software so I cannot review your layout. Can you print each layer in pdf?  Is the IC powerpad connected through vias to an unbroken ground plane that is much larger than the pad itself?  If not then the IC may not have proper heat sinking.

Regards,

Jeff

Hi Jeff,

Yes, the powerpad is connected to the bottom layer. The board size is 26*21mm.

Uploaded the top and bottom layer. The board is double layer.

Thanks.

bq24105_bottom.pdfbq24105_top.pdf

Hi Ion,

Thanks.  The grounding looks okay.  Is your board 2oz copper? 

The return path from the input capacitors to PGND pins is not ideal due to the digital pin traces on the bottom layer, as shown below.  Can you send take measurements of V(IN), V(BAT) to GND at the chip and also I(IN) and I(BAT) so we can confirm efficiency?  Also, can you take a scope shot of the SW node, showing the square wave?

Regards,

Jeff

Hi Jeff,

The board is 1oz copper.

Tomorrow I will have the measurement tools and will send you the results.

Also, I will try to assembly one more board with another IC and check if the first IC wasn't overheated during soldering.

Thanks.

Hi Jeff,

Here are the measurements:

Iset1: ~0.666A

V[in]: 15.00V

I[in]: 0.568A

V[bat]: 11.12V

I[bat]: 0.649A

Power supply used: Aim-TTi EL302R Linear lab power supply.

Attached are 2 screenshots made with Rigol DS1054Z oscilloscope on L1 pin 1 (from schematic).

Are these results expected?

Thanks.

Hi Ion,

The overshoot and undershoot are not expected.  Is your scope probe ground connection large/long or are you using a ground spring like shown below? 

If not spring and long ground connection, you may be picking up switching noise.  If not, then the undershoot is enough to damage the IC. 

If you replace the IC with a new one, do you get the same results?  What is the top marking of your IC?

Regards,

Jeff

Hi Jeff,

Unfortunately I cannot find the spring for the probe.

The measurements in the previous post were done with some wires on both - positive and ground side (bad measurement).

I redid the measurement again (screenshot)  with the ground hooked to a very small wire with the crocodile clip. The positive side was touching the inductor pad directly.

I've assembled another PCB and have the same behavior.

The top marking of the IC is CIF / TI 84J / PENV .

Is it a problem if I have a 16V Zener instead of 18V (as per datasheet)? On the EVM, the D1 is not present.

Hi Ion,

I would not expect a lower voltage zener to cause an issue. 

Regards,

Jeff

Hi Ion,

Do you have a BQ24105EVM onto which you can install one of the overheating ICs?  This would confirm whether the problem is IC related.

Regards,

Jeff

Hi Jeff,

Unfortunately, I don't have an EVM. 

I've made another measurement and charged a 3S battery with 1A charge current and got 75% efficiency, which is very low. The IC temperature was ~65 degrees C, which is still well beyond the maximum temperature, but I still think there is a problem with my IC/board/setup.

Ion,

Do you have a current probe?  Can you measure the inductor current?  What is the inductor's current rating?  Is it also getting hot?

Regards,

Jeff