TPS55332-Q1: The need for diodes on the input side

Hi Hirotsugu,

Thank you for using E2E forums, regarding your question these two diodes are for protection against the current flowing in the opposite direction into the supply, they are recommended if you check the EVM schematics here: https://www.ti.com/lit/ug/slvu319/slvu319.pdf?ts=1658846901248&ref_url=https%253A%252F%252Fwww.ti.com%252Ftool%252FTPS55332EVM you can see some diodes at the input as well, so I would keep the diodes in those recommended positions.

BR,

Haroon

Hi Haroon-san,

Will the IC break if there is no diode?

I think that D1 is the reverse voltage protection and D2 is needed for low Vin application, is it different?

Best Regards,

Hi Hirotsugu-san,

D1 is reverse voltage protection yes, and D2 is the output Diode which is necessary external output diode which conducts when the power switch is turned off. This provides the path for the inductor current to the output capacitor.

BR,

Haroon

Hi Haroon-san,

I understand the purpose of D1.
I'm sorry, I made a mistake in the number. Please tell me the role of D3.

Best Regards,

Hi Hirotsugu-san,

About the role of D3 I think it is used for protection against current flowing from Vin into the inductor and then Vout, according to datasheet it is recommended for this diode to be included, so I would use this diode in the design.

BR,

Haroon

Hi Haroon-san,

Please tell me where in the data sheet D3 is recommended.

Hi Hirotsugu-san,

On page 17, figure 11. typical application schematic, it is in the circuit design, also in the EVM of this device we see this diode present as well. I am not sure on the purpose of that diode, but I can ask some more experienced engineers and designers and get back to you.

BR,

Haroon

Hi Haroon-san,

I am using it to boost from 12V to 40V, but several units have failed.

I don't use a diode equivalent to D3, so I would like to know if the presence or absence of a diode causes a failure.

Best Regards,

Hi Hirotsugu-san,

I see, could I have the design specs and the design schematics (if possible maybe the layout as well) to better help you with your issue. 

BR,

Haroon

Hi Haroon-san,

Could you tell me the voltage of the internal Vreg?

I think that if Vin>Internal Vreg, there is no reverse flow.
Or is VIN = VREG = VOUT when D3 is present?

Best Regards,

Hi Haroon-san,

Since it is a boost converter, I think that VIN < VREG in general.
Do you mean the D3 diode is must part?

Best Regards,

Hi Harry-san,

I checked with EVM and confirmed a voltage rise on the VIN terminal.

I have three questions.

1. Please tell me the circuit diagram of the Internal VREG block.

2. I presume that the cause of the failed board is that there is no reverse current protection diode (D3), but I cannot confirm the rise of the VIN terminal on the working board.

Has the internal configuration of the IC changed between lots?

Is there some kind of timing for the rise of the VIN terminal?

3. Is it possible to operate without connecting the VREG pin to VOUT?

I don't think it's necessary to connect VREG and VOUT because I'm using 12Vin 48Vout.

Best Regards,

Hi Hirotsugu-san,

> Please tell me the circuit diagram of the Internal VREG block.
I am sorry, I cannot find any information about this detail

> Has the internal configuration of the IC changed between lots?
I cannot imagine that this is the case

> I presume that the cause of the failed board is that there is no reverse current protection diode (D3), but I cannot confirm the rise of the VIN terminal on the working board.

In case of the EVM there is this diode installed, so VIN can rise.
On your working board, without this diode, the VIN pin is tied to the input voltage of the board, so it cannot go higher.

> Is it possible to operate without connecting the VREG pin to VOUT?
> I don't think it's necessary to connect VREG and VOUT because I'm using 12Vin 48Vout.

You need to connect the VREG pin to the output voltage.
I have checked all (confidential) designs that I have access to.
There is not a single board without this connection. Also the datasheet does not mention that this would be optional.

Long story short:
Please follow the concept that you see on the EVM, which means
- install the diode towards the VIN pin   and
- install a 10uF capacitor directly from the VIN pin to GND.

Best regards,
Harry

Hi Harry-san,

We tested it on a working board.

Our board has an R-C filter between Vsupply and VIN.

We changed the resistor to a diode, but we didn't see any rise at the VIN pin.

Why does it behave differently than EVM?

Hi Hirotsugu-san,

I don't know why your board behaves differently.
Maybe the VIN pin of the TPS55332 already got damaged during former tests without the diode installed.

Best regards,
Harry

Hi Harry-san,

Are there any changes to the internal circuitry after release?
Is the internal circuit and internal Vreg supply source as shown in the attached diagram?

Best Regards,

Hi Hirotsugu-san,

I do not have exact information about the internals of this part, so that I can have a 100% proof.

But looking at all data that I have available I would say that this concept that you show is correct.

The physical implementation though, will probably not be one simple diode directly from VREG to VIN which is "bypassing" the internal regulator. 

Best regards,

Harry

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Hi Harry-san,

I tested with the EVM with the VREG pin open.
It works even if the VREG pin is open.
What is the function of the VREG pin other than bias to VIN?

Best Regards,

Hi Hirotsugu-san,

As far as I understand the VREG pin is the output of the internal regulator and at the same time the unregulated supply for the part.
To avoid that the internal regulator is used you can (in addition to VIN) feed some supply voltage into VREG so that the internal regulator is not used and does no longer dissipate heat.

If you do not feed any voltage into the VREG pin, please connect a ceramic capacitor (maybe 10uF) from that pin to GND.

All information in this correspondence and in any related correspondence is provided “as is” and “with all faults”, and is subject to TI’s Important Notice (http://www.ti.com/corp/docs/legal/important-notice.shtml).

Best regards,
Harry

Hi Harry-san,

When the VREG pin is open, it is not output from VREG.

And Isupply decreases when the VREG pin is opened.

I'm assuming it's because the internal circuitry is powered directly from VIN instead of via a boosted VREG.

Comparing the EVM with our working board, I suspect that the Internal Vreg structure has been altered.

Could you please confirm if there are any design changes?

Best Regards,

Hi Hirotsugu-san,

As I had mntioned earlier: I don't know why your board behaves differently.
Maybe the TPS55332 on your board already got damaged during former tests.

You could use a new part or do a cross change to find that out. 

Best regards,
Harry