TMP116: EEPROM/I2C General Call - Soft Reset/TMP117 Soft Reset

Hello Edward,

The Soft reset bit for TMP117 does not exist in the TMP116 device configuration register. I don't believe the soft reset bit in the TMP117 can replace the I2C general call to soft reset the bus. I can check though and confirm.

Thanks. The real question on the TMP117 is if the soft reset via the config register can ensure that data is correctly written to EEPROM when the tmp117 is soft rebooted. This would get around issue of having to issue the general call to commit the data to EEPROM. In my situation I was hoping to use 16 TMP116s/TMP117s (yes, with a mux). But in order to use the EEPROM in the TMP116/TMP117 you need to use the general call reset to write the data. However, issuing a general call reset would reset ALL 16 sensors AND ANY OTHER i2c device on the bus that responds to the same general call reset. In my case that's another half dozen ICs. In an RTOS (or nearly any other situation) this effectively changes the state of the other sensors and other i2c devices on the bus and all software process are now effected. The only choice is to essentially reboot the whole OS to get things back in to a state of sanity. And of course, if you were to write to all 16 tmp116s, that would mean 16 reboots. I am hoping that the tmp117's soft reset fixes this.

An alternative solution might be to unlock the eeprom, write the changes to the registers, and rather than using a general call to commit the data and reset the tmp116/tmp117, to hard reset it. In other words, I could use a GPIO to control the power to the TMP116/TMP117s, then reset the sensors via toggling the power rather than using the i2c general call reset. Will that work?

Thanks

Hello Edward,

I don't think that the TMP117 Soft reset bit in the configuration register would have the same effect or at least no mention of the same is there in the documentation. I will need to check to make sure I am correct in my interpretation of the data.

Hello Edward,

I checked the programming of the device on a TMP117 and it seems that the power cycle works fine for the TMP117. The sequence of events that I did was as per the datasheet, except that instead of the I2C General Call, I used the GPIO to power cycle with 15 ms wait for the device to power down after the GPIO was made 0, before powering it up.

Additionally, I made sure that the device was configured not to perform any conversions when the EEPROM was being programmed.

Amit

1) I have verified that using a GPIO to power down/up to reset works on the TMP116 as well. I did something similar to what you did, in my case I powered down for 50ms. This works, but is a little ridiculous as a work around.

2) I assume when you say you "made sure that the device was configured not to perform any conversions when the EEPROM was being programmed", you mean you used the shutdown mode to accomplish this? This does not appear to be a requirement from the documentation. Do you feel it is necessary? If so, then how would you write to eeprom a configuration where you need continuous conversion to persist in eeprom?

3) I do not have access to a tmp117, but you do. Can you also try and see if the software reset via the config register also works? I can then right my code for the tmp117 and use it when it becomes available. This would be much cleaner.

4) Looking carefully at the flow chart in figure 24 for the TMP116 and Figure 19 for the TMP117, notice the wording around the dashed box - "programming process". This is not described in the text, but a possible interpretation of the box and text is that programming is complete prior to the i2c general call reset. This would mean that the general call reset would be just a good practice to make sure what you read back after a reset is the values you programmed earlier and they are re-loaded correctly. This would be consistent with why the power down/up works for both the tmp116 and tmp117. But this is a interpretation, not something I can rely on for a product we ship to the market place. Can you verify that this is the fact?

Thanks