TMP117EVM: worked for a while, now says Hardware Not Connected

0x48 (default) is the only choice.  

Hi Richard,

Are you using TMP117EVM_GUI 1.3.0 from dev.ti.com? There should be four address options.

thanks,

ren

OK, I resolved that problem by installing version 1.3.  But changing the I2C makes no difference

Hi Richard,

Have you made any changes to the hardware? Your error message suggests your USB device is detected and firmware is working correctly. There must be a problem between the MSP430F5528 chip and the TMP117.

The header J1 should not be populated with a shunt if R3 is populated with a zero ohm resistor. Installing a shunt on the right side of J1 will short the TMP117's power supply. 

If the TMP117 has been separated at the perforation, there would be no I2C response.

If the resistors R15/R16 were altered on the break-able section, they would alter the I2C address and potentially result in no I2C response.

thanks,

ren

I left the jumper in place between VDUT and VOUT as received and am powering it via USB.  I did break off the sensor end and re-connect with a ribbon cable.  It was working like this before it quit, but I will re-check the ribbon connections.

Richard,

You wouldn't happen to have two TI EVMs plugged in at the same time, right? Seen in Options > Serial Port

thanks,

ren

Everything looks OK electrically, I checked all the connections from the device pins through the cable to the main board. There is only one temperature device. Changing I2C address makes no difference.

Would cable length cause problems?  Added resistance is about 0.5 ohms.

Hi Richard,

With I2C, the problem with cabling is going to be capacitance not resistance. The EVM is equipped with 5.1k ohm pull-ups which should be okay but could be a bit stronger. You could replace them with something as low as about 2kohm. Before you do that, please put an oscilloscope on the SCL/SDA if you can. You should see an I2C transaction when the GUI connects.

thanks,

ren

I have no way to do this here; possibly tomorrow.