INNO2603K SR driver issues
I have design with a DC input of 150V and 4-outputs (-3.3V, +3.3V, -16V, +16V). The 2603K is powered via 6.6V. The transformer primary inductance is about 550uH, with 45 turns (outputs at 1 turn for 3.3V and 5 turns at 16V). The total rated output for all rails is 3W. All of the secondary SR FETs are connected to the SR pin output of the chip.
The system works well when only the two 3.3V outputs are operating with the SR FETs (16V outputs SR FET gate drives not connected), even when the 16V transformer secondary windings are connected. However, as soon as the 16V SR gate drives are connected to the SR pin, the INNO2603K chip heats to the point when it reaches thermal shutdown (occurs at 2.5W output). At this point, there is also an increase in audible noise from the transformer.
After recording much data, all of the SR FETs are turning on at the exact same time. There is circuitry to ensure that there is sufficient Gate-Source voltage for each secondary FET, since the secondary ground reference is -3.3V. Without the 16V FETs being gate driven, the body diodes conduct and the output voltages reach up to 200% of the output voltage (open-circuit voltage) compared to when the FETs are gate-driven. The chip does not overheat when the 16V SR gate drives are not connected, but I would like them to be connected for the design. After making many minor circuit changes, there has been little change to the results.
Is it possible that the chip cannot drive 4 SR FETs? Based on the waveforms, the duty cycle seems to be quite high at low power levels, and it appears that the primary FET occasionally conducts twice in one cycle (second time occurs after the secondary FETs turn off). I am not quite sure what the cause of the chip overheating is, but as the load increases, the chip increases in temperature to the point of shutdown (i.e. the chip is still warm at lower power levels but does not reach thermal shutdown). Any ideas to what the cause could be?
评论
We cannot provide the schematics and board layout. The InnoSwitch sits directly under the transformer. The schematic is very similar to the 5V and 12V output schematic in the datasheet for the EP. However, there are two more outputs (4 total).
The data we have is above in the first post. Transformer primary inductance is 550uH, and the total output power is 2.5W.
We have tested and analyzed different components of the system. Our conclusion is that the issue could be the transformer (possibly too much primary inductance for output specs and causing a high duty cycle) or the chip having problems driving 2 more FETs for SR rectification. There might be cross-conduction on the falling-edge of the SR pin output with the primary MOSFET. Based on the waveforms, it seems that the primary MOSFET is conducting twice in one cycle, possibly causing the cross-conduction. Please take a look at the waveforms attached to this post (note that voltage output specs may not be incorrect in a few of the waveforms, but that issue has been solved).
| Attachment | 大小 |
|---|---|
| PI Data to be sent.pdf (86.55 KB) | 86.55 KB |
Hi,
Unfortunately with out looking the details of schematic, PCB layout and transformer design, I can only give my comments based on the limited data that you have provided.
The transformer seems to have very high leakage inductance because of multiple outputs, I would recommned you to try sandwitched transformer construction inorder to reduce the leakage inductance. Praimary drain current is not looking right either.
Could you also please check with SR FET's replaced with Schottky diodes and see if it works.
If you are not comfortable sharing your schematics on forum the you can contact one of our field office in your geographical region so that they can help you in resolving the issue. You can find our filed office locations at following page.
https://ac-dc.power.com/sales/
Regards,
PI-NANO
Hi,
Can you send your power supply schematic, board layout and transformer design etc?
Can you also capture SR FET gate and drain to source voltages along with primary drain to source voltage?
Regards,
PI-NANO