Trouble recommend high temperature resistance (above -40 to 105 °C) PIN TO PIN replacement

Hi user3756530,

We do not have a p2p replacement for SI5328B-C-GM.

The closest functional replacement is LMK5B12204. Note that the junction temperature maximum for LMK5B12204 is only 135°C, but TI offers more thermal modeling and better characterization than what is available on SI5328B datasheet, so it is easier to estimate typical and worst-case performance. With two outputs using HCSL (the output format with highest current draw) and 2.5V VDDO rails, I calculate about 1.18W power dissipation. Even on a 4-layer board model with ΘJA = 24°C/W, and 105°C ambient temperature, the junction temperature is 105 + 24 * 1.18 = 133.3°C, which is within recommended operating conditions. In practice, a customer board will have many more layers, ΘJA will be smaller, and the device temperature will be much lower than 135°C - the datasheet gives an example for a 10-layer board with ΘJA=9.1°C/W, in which case device temperature would be 105 + 9.1 * 1.18 = 115.7°C which is well under the recommended operating temperature limits. If the customer selects a different output format or powers VDDO using 1.8V supply instead of 2.5V supply, the power consumption and device temperature can be further reduced. In summary, even though LMK5B12204 does not have any explicit ambient temperature rating listed in the datasheet, I don't think it will be a problem to use LMK5B12204 at 105°C.

Regards,

Hi user3756530,

Again, I stress that the datasheet does not actually list an ambient temperature rating. There is only a recommended junction temperature range (-40°C to 135°C). We have to put an ambient temperature range on the product web page to help sort it in the table search, and 85°C is a common maximum rating for industrial applications. But for the reasons outlined in my prior post, the operating temperature range listed on the website is not a good reflection of the actual valid ambient temperature range for the device.

To demonstrate the artificial nature of the ambient temperature range on the web page, consider the following example. The LMK5B12204 has two analog PLLs and four output banks, and the output banks can be operated at 3.3V. If the customer tries to use both PLLs and all four outputs at 3.3V on a 4-layer board, the worst-case power dissipation is close to 2.7W, ΘJA in the datasheet for a 4-layer PCB is 23.3°C/W, and at 85°C ambient the junction temperature is 85 + 23.3 * 2.7 = 146.8°C, which exceeds the recommended maximum junction temperature by about 12°C. It's not that the part isn't suitable for 85°C operation, just that some use cases require better thermal design: more board layers, lower voltage to the VDDO rails, or lower-power output formats could reduce the junction temperature increase to acceptable levels.

The competitor device has only two outputs. If the customer is using only two outputs of LMK5B12204 at 2.5V or 1.8V VDDO, on a board with more than four layers, to generate two outputs, they could disable one of the APLLs and two of the output banks. Note that the customer use case for the LMK5B12204 would use less than half the power of the worst-case example above (1.18 / 2.7 = 43%), and their PCB likely has better thermal performance. In the previous post, I calculated expected junction temperature for a "worst-case" 4-layer board, and a more typical 10-layer board, and explained why I expect the LMK5B12204 to function properly even at 105°C ambient. Please do not disqualify the solution solely on the basis of an arbitrary temperature range that is not taken from the datasheet, used by TI.com only for sorting devices in product search tables. As long as they stay below the recommended maximum junction temperature, the LMK5B12204 can be suitable for their application.

Regards,