Problems interfacing with muliple DS2130 digital thermometers
Zenexer
Posts: 20
I have four DS2130's connected to a Propeller microcontroller. It's a very basic setup: they all have their GND, Vdd, RST, and CLK pins connected to each other, with RST and CLK going to Pin 1 and Pin 0 respectively. Each DQ (data) pin gets its own pin on the Propeller.
Everything works fine if I only have one DS2130 connected. However, the more add, the less reliable they all become. One IC works 100% of the time, two ICs about 10% of the time, three ICs about 1% of the time, and four ICs... never. My definition of "reliable" is generous: all of the ICs must report a temperature greater than freezing. When an IC is removed from the socket, -.5 gets transmitted (two's complement, nine bits, .5 increments; 0 0000 0000 = -.5). When an IC is in the socket but not functioning because too many are present, it always reports exactly 0 (1 0000 0000 in binary). I'm providing each IC with 5V, which should be more than enough; as far as I can tell, the voltage never drops 4.5V with everything plugged in. There are no completely defective connections or ICs, as they all work individually, and all slots work when used exclusively. I've tried using the 5V from the USB cable and an AC adapter supplying 9V DC (actual output measured at 9.20V), the latter with a voltage regulating circuit to drop it to 5.0V. Both behaved identically. Even when none of the ICs are reporting correct temperatures, commands such as configuration writes can still be performed without any problem. Fiddling with the wires a bit can affect whether or not a reading is successful, but there isn't one specific wire that seems to make more of a difference than any other.
Any ideas? I'll post a complete schematic and/or picture tomorrow if necessary. The only other standalone components are several capacitors, a crystal, a PropPlug, a lot of wires, a few bread boards, and two voltage regulators (3.3V for the Propeller and 5V for everything else). The longest total stretch of any DQ/RST/CLK connection is an estimated 5 inches; shortening that length does not seem to affect reliability.
Everything works fine if I only have one DS2130 connected. However, the more add, the less reliable they all become. One IC works 100% of the time, two ICs about 10% of the time, three ICs about 1% of the time, and four ICs... never. My definition of "reliable" is generous: all of the ICs must report a temperature greater than freezing. When an IC is removed from the socket, -.5 gets transmitted (two's complement, nine bits, .5 increments; 0 0000 0000 = -.5). When an IC is in the socket but not functioning because too many are present, it always reports exactly 0 (1 0000 0000 in binary). I'm providing each IC with 5V, which should be more than enough; as far as I can tell, the voltage never drops 4.5V with everything plugged in. There are no completely defective connections or ICs, as they all work individually, and all slots work when used exclusively. I've tried using the 5V from the USB cable and an AC adapter supplying 9V DC (actual output measured at 9.20V), the latter with a voltage regulating circuit to drop it to 5.0V. Both behaved identically. Even when none of the ICs are reporting correct temperatures, commands such as configuration writes can still be performed without any problem. Fiddling with the wires a bit can affect whether or not a reading is successful, but there isn't one specific wire that seems to make more of a difference than any other.
Any ideas? I'll post a complete schematic and/or picture tomorrow if necessary. The only other standalone components are several capacitors, a crystal, a PropPlug, a lot of wires, a few bread boards, and two voltage regulators (3.3V for the Propeller and 5V for everything else). The longest total stretch of any DQ/RST/CLK connection is an estimated 5 inches; shortening that length does not seem to affect reliability.
Comments
Also my search engine does not come up with a thermometer with that part number. Could you link to the datasheet?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen
Post Edited (Franklin) : 1/24/2010 7:30:36 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen