TPS546C23: TPS546C23RVFR

Hi Vigita,

Are you trying to program a one phase device? If it's a multiphase, are you trying to write that to primary or secondary? If you are trying to program secondary device that wont be possible as I had mentioned to you on a previous thread:  TPS546C23: TPS546C23RVFR , you can't write to those exact commands (02, 20, 21, 24, 27, 29, 2B, 41, 45, 60, 61, 62, 63, 64, 65, D4, D5, D6, D7) regardless of the method or board you use, the device is just build to always nack for those commands if device is slave/secondary, that is just how it was set to be and that is normal behavior. Reason why we built it that way is because only primary device can set those values.   

Can you ellaborate on the following,

1, Can you please explain what do you mean by primary or secondary device?

2. Also can you explain 'one phase device'

I am programming only one device, using BPM programmer. How does the device decides it is secondary? can you please ellaborate on this? When you say 'only primary device can set those values', what do you mean by that. I am a bit confused about this primary and secondary device concept.

I am giving 5V to pin 29. 13 14 15 19 20 26 38 41 are grounded. I am then using sda and scl pins for communications. Whatever pins that have a pindriver connections is internally pulled up to 3.3V too.

Can you look at our schematics and see if something here is causing out problem.

The current issue is we are not able to program these registers (02, 20, 21, 24, 27, 29, 2B, 41, 45, 60, 61, 62, 63, 64, 65, D4, D5, D6, D7) and I think one of them will set device output. Our end customer needs it. Please assist us on fixing this issue.

Did you check 80h(bit 1) value as I mentioned above?

and can you write to other commands besides those?

Yes, I have checked 80h, I read 3h, meaning is_slave and sync_fit is set right?

Yes, I am able to program and verify other NVM registers.

Yes based on 80h = 0x03h = b'0000 0011, device is reading as slave and that is why you can't write to those commands.

do you have a better schematic? It will take some time to review since I have to trace all 41 pins because they aren't labelled properly, and then check each one, plus is a bit blurry. 

Something in this format would be more readable:

As of now, I can share only the image I shared before. Kindly review it and let me know, how to make my device a loop master. Thanks!

Vigita,

I tried to evaluate but that image is not readable, I even tried increasing quality. Please request full readable high quality schematic from your customer in the format I mentioned above. 

Also check if there's any other devices on the PMBus bus. If so, disconnect all but our device, power cycle and check 80h again to see if it's still reading as slave. 

6735.Schematics.pdf 

J9 is the socket on which the device wiill rest.

Please hep us resolve this issue. Help us make the device loop master so that we can program all the registers.

Vigita,

I'm trying to help but I need the right format. Can you request the other format I mentioned to your customer? This is missing a lot of components.

For reference:

1) What's the state/connection of VSHARE and ISHARE? Are they connected to something, is there a trace? 

2) What's the state/connection of DIFFO and COMP? COMP can't be going directly to ground and DIFFO must have a network of R and C.

3) Is there's any other devices on the PMBus bus besides that one TPS546C23?

The image shared is the image of the socket module design. On top of the socket the device will reside. The Vshare and Ishare has no connection. Diffo is also not connected to anything. I apologize - Mild correction. Pin 36 and 37 arec connrcted to GND through a 2.2uF cap.

Does CNTL pin has any significance?

I see. 

It is a little bit difficult to add those R and C to the module now for DIFFO, FB and CMP, as the module is built already. Is there anyother way we can make it loop master. Like, if the device is expecting certain voltage level on the FB pin, to make it get detected as loop slave, we can feed the FB pin with needed voltage. That wil be easy for us. Let us know if that can be done. 

sorry there was a mistake in the previous question.

It is a little bit difficult to add those R and C to the module now for DIFFO, FB and CMP, as the module is built already. Is there anyother way we can make it loop master. Like, if the device is expecting certain voltage level on the FB pin, to make it get detected as loop master, we can feed the FB pin with needed voltage. That wil be easy for us. Let us know if that can be done.