Hi,

I hope you are all doing well in this new year.

I have designed my first PCB using KiCad, and it has been produced by JLCPCB.

I use FSR sensors to measure pressure. The analog signals are then digitized using an LTC2497 16-channel 16-bit ADC chip. The PCB is connected to a microcontroller using an I2C interface.

The schematic is attached.

I am using an Arduino Uno, but I cannot connect to the LTC2497.

I used this I2C scanner script from Adafruit to detect the I2C addresses, and it found an I2C device at 0x01, but I cannot communicate with it. The I2C address of the chip should be 0x14 (the three address pins are connected to ground, see datasheet p.17).

The PCB LED works, and the voltage between the ground and power pins is 3.1 V.

The first time I used the PCB, the voltage at the SCL and SDA pins was around 3.1 V. Now the voltage stays low, around 0.2 V for these two pins. The LTC2597 chip seems to be working because it is quite hot. I do not have an oscilloscope to debug the I2C connection further.

If you have any insights about this, I would appreciate it very much.

 

Hi,

I hope you are all doing well in this new year.

I have designed my first PCB using KiCad, and it has been produced by JLCPCB.

I use FSR sensors to measure pressure. The analog signals are then digitized using an LTC2497 16-channel 16-bit ADC chip. The PCB is connected to a microcontroller using an I2C interface.

The schematic is attached.

I am using an Arduino Uno, but I cannot connect to the LTC2497.

I used this I2C scanner script from Adafruit to detect the I2C addresses, and it found an I2C device at 0x01, but I cannot communicate with it. The I2C address of the chip should be 0x14 (the three address pins are connected to ground, see datasheet p.17).

The PCB LED works, and the voltage between the ground and power pins is 3.1 V.

The first time I used the PCB, the voltage at the SCL and SDA pins was around 3.1 V. Now the voltage stays low, around 0.2 V for these two pins. The LTC2597 chip seems to be working because it is quite hot. I do not have an oscilloscope to debug the I2C connection further.

If you have any insights about this, I would appreciate it very much.

If the LTC2497 is hot then that means that something is wrong or that the IC is blown.

 

Since the I2C scanner does not find a valid address, then something is wrong with the setup or hardware itself. Recheck the pcb traces and if you have a scope, verify the I2C traffic.
I agree with Mrchips that something is blown since you are seeing 0.2v on the I2C lines.

 

Thanks for your replies.
I have ordered five PCBs from JLCPCB and the same issue has occurred with a second one. I am wondering if there is an error in the schematic. I'm hesitant to test the remaining PCBs as I don't want to risk damaging them.
I don't have a scope to test this.

 

The Uno has an output voltage of 5V and driving the PCB with only 3V exceeds the LTC input restriction of max Vcc+0.3V.
So the device ic most probably fried. Drive the PCB with 5V and i should work.

 

I agree with Mrchips that something is blown since you are seeing 0.2v on the I2C lines.

Seems like a jump to a bad conclusion.

Do the I2C SCL and SDA lines have pullup resistors?

 

I also tried to drive it with 5V, but the LTC became extremely hot and still didn't work.
Yes, the I2C lines have 2k pull-up resistors.

 

That chip should draw near zero quiescent currect across its operating voltage range. Check all connections before doing anything else.

 

I checked the connections and found a resistance of 1200 ohms between the SDA and SCL pins. This suggests a short between these two tracks. The problem is on all five PCBs I ordered. Please review the attached KiCad PCB file (or its screenshot) for a closer look.

Before production, I ran the design rule check (DRC) test and received an error that read " Front solder mask aperture bridges items with different nets " for LTC pads. After researching online, I decided to ignore the error and choose the Kicad option " Allow bridged solder mask apertures between pads within footprints ". I thought this was safe to do, as it is common for QFN packages. However, I am now concerned that this may have caused solder bridges leading to shorts between the tracks.

Do you believe this could be the cause of the issue?

 

Pullup resistors could be the cause of that. I would expect something like that in a normal board.

 

When I tested other boards with pull-up resistors, there was no conductivity between SDA and SCL.