Is MPM3695-10 READ_TEMPERATURE_1 8-bit plus sign or 9-bit plus sign?

The MPM3695-10 datasheet isn’t clear about the size of the READ_TEMPERATURE_1 PMBus command. Is it 8-bit plus sign or 9-bit plus sign?

The command description indicates 8-bit plus sign and a range of -255 to +255.

READ_TEMPERATURE_1 (8Dh)
The READ_TEMPERATURE_1 command returns the internal sensed temperature. This value is also
used internally for over-temperature fault and warning detection. This data has a range between
-255°C and +255°C.

READ_TEMPERATURE_1 is a 2-byte, twos complement integer. The sign bit is bit[9]. Table 5 shows
the relationship between direct and real-world values.

But then it refers to Table 5 which shows 9-bits plus sign and a range of -511 to +511.

Table 5: Direct and Real-World Values for READ_TEMPERATURE_1

| Sign | Direct Value | Real World Value (°C) |
| 0    | 0 0000 0000  | 0                     |
| 0    | 0 0000 0001  | 1                     |
| 0    | 1 1111 1111  | +511                  |
| 1    | 0 0000 0001  | -511                  |
| 1    | 1 1111 1111  |  -1                   |

Which is correct?

Welcome to the forum! Apparently the table was generated with Excel and it defaults to 9 bits plus sign so +511 to -511. But the MPM3695-10 and other versions use just 8 bits plus sign so are valid only for -255 to +255. Realistically its a formality since the device likely wont operate well at over +150’C or below -50’C.

Thanks for the clarification James. I don’t know if you are making an update to the datasheet, but if I understand correctly it should say:

READ_TEMPERATURE_1 (8Dh)
The READ_TEMPERATURE_1 command returns the internal sensed temperature. This value is also
used internally for over-temperature fault and warning detection. This data has a range between
-255°C and +254°C.

READ_TEMPERATURE_1 is a 2-byte, twos complement integer. The sign bit is bit[8]. Table 5 shows
the relationship between direct and real-world values.

With the bitfield table having “Sign” in the column for Bit 8 (the ninth bit).

And then Table 5 should be as follows:

| Sign | Direct Value | Real World Value (°C) |
| 0    | 0000 0000    | 0                     |
| 0    | 0000 0001    | 1                     |
| 0    | 1111 1111    | +254                  |
| 1    | 0000 0001    | -254                  |
| 1    | 1111 1111    | -1                    |

(Also the bitfield table indicates bit access as R/W, but all bits for READ_TEMPERATURE_1 should be read-only).

Thanks,
Gregg.

So James, littel confusion, for 3695-10 , the bit[8] is sign bit? or the bit[9] is the sign bit?