TPS55288: TPS55288 No communication and wrong output.

Hi Ryan,

Could you share your application Vin range, Vout, max load current information?

I have checked the attached schematic, below are some findings that would affect the device normal working:

1. EN/UVLO pin is directly connected to VIN. EN/UVLO pin absolute maximum voltage rating is 20V, make sure the VIN voltage is no large than 20V, otherwise the device would be damaged.

2. It's must and needed to have enough input capacitor from VIN to GND and enough output capacitor from Vout to GND. For a high switching frequency dc/dc converter, the input capacitor and output capacitor would make a important effect on the device normal working. I would suggest you follow the TPS55288EVM-045 schematic design and put enough input caps and output caps.

3. compensation: The compensation network needs to be designed carefully based on the Vin, Vout, Iout, external inductor and output capacitor. Without a proper compensation network, the output voltage would oscillate and even shutdown.

4. ILIM: It's suggested to follow the datasheet inductor average current limit equation to select suitable resistor value.

5. Layout: From our experience, 50-70% failure cases are related to layout. It's needed to follow TI datasheet and TPS55288 layout guideline app note on ti.com.

Hi Zack,

Thank you for responding.

Our main concern is not being able to talk to the chip over I2C. With our current setup, I believe it should at least be able to do that.

Does this chip require programming before it can properly output desired voltages? 

Our input voltage has been primarily 5V. On the one board we have tried upping the input voltage to near 20V but it did not resolve anything. Every board still behaves the same.

The expected output is 20V with a load up to 3A.

1. We saw in the data sheet that the EN/UVLO pin max rating was 20V and have not exceeded it.

2. We have 4 x 10uF caps on the input as the datasheet calls for. We can try to add more capacitance as the EVM shows it has 89uF on the input. We are missing the output capacitor. The output cap shouldn't impact the device from turning on.

3. We will look into the compensation network.

4. It appears that the resistor value is suitable for the inductor.

5. We will look into the layout to see if any issues stand out. 

Hi Ryan,

You can firstly check the TPS55288 VCC pin voltage. Once the Vin and EN exists, the VCC voltage should be regulated at a certain voltage. For example, once Vin>5.2V, VCC=5.2V. 

If VCC exists, then please check the SCL, SDA is correctly send to TPS55288. Check the I2C slave address 74H. To let the TPS55288 starts switching, you need to set MODE register bit 7 (OE) value = 1.

If the VCC is correct and I2C signal is correct, then please probe SW1 and try to catch if any switching waveform could be found.

Hello Zack,

I set Vin to 5.3V, with Vin being tied to EN, EN also equaled 5.3V. VCC measured to be 5.16V. 

I recorded the I2C output which shows that 0x74 is correctly being sent as a write, but the TPS55288 will not send the acknowledge signal. On the last clock cycle, the TPS55288 should hold SDA low. Due to the TPS55288 not acknowledging, my I2C program won't continue to send the data to program the mode register. 

Hi Ryan,

Have you tried replaced the IC with a new one and still have the same result?

Another test if changing the TPS55288 slave address to 75H by modifying the MODE resistance and try again.

Could you also share the layout?

I have not replaced the IC yet, but since all of my boards are behaving the same way I don't think replacing one IC on one board will fix this issue. We would like to get an EVM board to figure out how this chip is supposed to operate, but everything is sold out right now.

We changed the Mode resistor to 15kohm(should be 14.3k per the datasheet) to set the i2c address to 75H, still not receiving the ACK from the TPS55288. It could be the due to the wrong resistor value, but given that it won't communicate with the correct 0ohm mode resistor at the 74 H address, I was not surprised that this didn't work either.

As far as the layout, I tried to prepare all the layers to be somewhat useful to show you but there is too much detail missing to make any sense of it. This board is very densely populated.

I don't know if we received a bad batch of TPS55288s, but nothing else seems to be standing out.

Hi Ryan,

As you mentioned, if you have confirmed that the IC is I2C version, not TPS552882 (non-I2C version), it's quite strange to see the behaviors.

Have you tried to read the internal registers values? It seems there is some problem to write value to IC but I also want to confirm if it's Ok to read some value.

Could you contact TI FAE so that FAE may apply for one free EVM for you for debugging.

Zack, the current I2C code I have written dump reads all of the registers, but it is getting hung on the address due to the TPS55288 not acknowledging. I will make a post on some issues I am having with another chip, TPS65982, and will include my register dump there. It uses the same code I am using to attempt to read from the TPS55288, the only difference is the address and SDA and SCL lines.

I appreciate your help, I have reached out to TI FAE to see if they can send me an EVM.

Hi Ryan,

OK. Let me know anytime if there is any update.

Hi Zack,

I believe I was approved for the sample program. Unfortunately it seems the EVM boards are not something that can be requested as a sample. At least when I search for them, only the individual chip appears in the search and not an EVM. On top of that, all the chips are out of stock as well.

Hi Ryan,

The TI FAE should be able to provide you one EVM. Have you contacted them?

The first time I attempted to contact them I was told to apply for the Sample Program. I just reached out again and I think FAE will reach out to me soon to see if they can provide me with the EVM boards. I will let you know how it goes.

Hi Ryan,

OK. Let me know if there is any further update.

Hi Zack,

I attempted again to get in contact with FAE, I thought there was some progress being made because the live chat person I was speaking to last week sounded hopeful. It seems like the live chat person who helped me create the case is not the same person who is dealing with the case currently. They did not know why I was reaching out to FAE even though I had explained that during the live chat. It doesn't seem this new person has access to the chat logs, or didn't look? Regardless, I have placed an order for the TPS55288EVM-045 last week so hopefully I can make some progress soon.

Hi Ryan,

OK. Looking forward to hear to good news.

I received the EVM and was able to program it over I2C and get an output. I still don't see anything that stands out on our board that would prevent this chip from communicating. It is interesting that this chip does not hold on to the programmed settings when turned off. I didn't see anything in the data sheet about it having volatile memory.

Hi Ryan,

Thanks for the update.

If the chip is shutdown, all the internal registers are rest to default values.