Issue with Propeller Demo Board...
JimInCA
Posts: 80
I have a HMC5843 3-Axis Magnetometer Breakout board from Sparkfun. I soldered a connector to the board so that I could plug it into the solder-less breadboard of my Propeller Demo Board for evaluation. The device requires 3.3V, gnd, and a two wire I2C interface, which I wired up appropriately. When I powered up the demo board, I had no problem reading or writing any of the HMC5842's registers. But the x, y, and z axis registers always read back as $0020 regardless of the movement of the device. I did notice that every once in a while, the device would work. I eventually tracked the problem down to the powering sequence of the Propeller Demo Board.
I found that when I turned the power supply switch off, that I still had 1.6V on the 3.3V input the the HMC5843. So I unplugged the breakout board and I still had 1.6V present on the 3.3V signal. I then realized that I still had the USB cable connected, so I unconnected it and the 1.6V went away.
It is my assumption that the 5V input to the USB bus is present on the FTDI chips RXD and/or TXD signals. This voltage is enough to energize the input protection diodes on the Propeller's P29 and/or P30. This voltage bleeds through the Propeller's Vdd pin to the Vdd power plane on the Demo Board.
I was wondering if anyone else has seen this "feature" with the FTDI/Propeller combination before? Also, people need to be aware of this when working with the Propeller Demo Board.
I believe that my problem with the HMC5843 has to do with the power sequencing of the device. It's obvious that the device didn't like the 1.6V being present prior to the 3.3V coming up. I'm assuming that this was enough to cause the magnetometer inner circuits to latch-up. The fix for my problem is to first power up the board and then plug in the USB cable. Either that or apply power to the HMC5843 after the board is powered up.
Jim...
I found that when I turned the power supply switch off, that I still had 1.6V on the 3.3V input the the HMC5843. So I unplugged the breakout board and I still had 1.6V present on the 3.3V signal. I then realized that I still had the USB cable connected, so I unconnected it and the 1.6V went away.
It is my assumption that the 5V input to the USB bus is present on the FTDI chips RXD and/or TXD signals. This voltage is enough to energize the input protection diodes on the Propeller's P29 and/or P30. This voltage bleeds through the Propeller's Vdd pin to the Vdd power plane on the Demo Board.
I was wondering if anyone else has seen this "feature" with the FTDI/Propeller combination before? Also, people need to be aware of this when working with the Propeller Demo Board.
I believe that my problem with the HMC5843 has to do with the power sequencing of the device. It's obvious that the device didn't like the 1.6V being present prior to the 3.3V coming up. I'm assuming that this was enough to cause the magnetometer inner circuits to latch-up. The fix for my problem is to first power up the board and then plug in the USB cable. Either that or apply power to the HMC5843 after the board is powered up.
Jim...
Comments
Your inference regarding the parasitic voltage pathway is correct. Usually, however, it's the opposite effect that people notice: the Demo Board powering a disconnected FTDI chip via P30 and causing multiple Propeller resets by power-cycling the DTR pin.
-Phil
Jim...