Attaching multiple temp sensors

I have ordered several DS18b20 temp sensors.
Could someone please instruct me on the correct connection procedure? i.e.: are all 8 sensors connected the same 3 wire temp input or does one use the 8 digital Aux inputs?
 
I did research and another gadget, seems similar than the Webcontrol, had instructions on adding one, having it recognized, then adding another - and so on. Would this be correct with this board as well?

Also, what would be the max length these sensors can be  placed from the board? If this should go on a separate Topic please let me know. Did read iobe thread on the subject, but at the end was confused on seperate power/ versus parasite power (I would imagine means the same)

Each one-wire device has an burned in address, so they all connect to the same bus. Extend the three temperature terminal (5V, Gnd, data) to the locations you need sensors.  WC does not support parasite power, that is where a sensor "steals" power from the data line when it is not in use.
 
WebControl detects sensors at power up. Once detected you need to assign each sensor to one of the Temp location. To simplify this connect one sensor at a time, power up the WC, assign  the sensor, turn WC off, connect another sensor, rinse and repeat.
 
Use a bus topology where stubs to each senor are under a couple of meters.
Maxim has a bunch of valuable app notes for 1-wire stuff
 
AN148 guidelines for reliable long line 1-wire networks
http://www.maximintegrated.com/en/app-notes/index.mvp/id/148
 
AN1796 Overview of 1-wire technology
http://www.maximintegrated.com/en/app-notes/index.mvp/id/1796
 
AN5026 Protecting 1-wire slave from overvoltage
http://www.maximintegrated.com/en/app-notes/index.mvp/id/5026
 
Network length can be dozens of meters. A lot depends on the noise environment.
 
/tom

TY very much, Tom!

I will do further reading per your links,
This electrical/ electronic stuff seems to be fun, with a steap learning curve
 
Rainer

Again showing my total ignorance, but does the below correctly depict 9albeit simplified) how the sensors can be attached in a chain WITHOUT external power source?
 
 

Thanks for those Tom!
 
There is good information there about signal noise that may explain some of the problems I am having. In particular slew rate control.
 
I saw a lot of this on my scoping of the  1Wire I/f but couldn't figure out why this was happening. I get some nasty ringing from the edges of switching on some of them but others are clean. Since I have no time delayed trigger on the scope I couldn't zoom into closely analyse later in the communications and it was almost impossible to pick apart who was doing what.
 
I have to conclude that the sensors, being Maxim built 1Wire interfaces, have slew rate control and do NOT produce as much ringing on the data line and that the WC8 1Wire interface is only an open drain/collector interface, without any slew rate control components, and produces the nasty ringing on the data line. This assumption furthers my line of attack on solving some comm issues with 1Wire on the WC8 interface.

Again showing my total ignorance, but does the below correctly depict 9albeit simplified) how the sensors can be attached in a chain WITHOUT external power source?

The WC8 bad supplies three lines to the 1wire devices (non-parasitic), +5v, data, and common. The power source is supplied by the WC8 on the 5v separate line.

Thanks - so simple chain attachment would be fine?

Yes daisy chain is the best by definition. You may want a termination resistor on the end (or every one) of the chain to suck up ringing by lowering line impedance and lessening sudden impedance changes at the ends.  Some people have reported using a star topology or even combinations also very successfully with and without termination resistors.. If cables are short almost almost any topology can be used. It's the long cables that are the worst for "echoes" off the end of the data line.
 
Cables are a half turn coil impedance. Add that to capacitance between the lines in the cable (or at circuitry ends) you can have a tuned tank circuit that can ring to sudden voltage changes (your data).

LarrylLix said:

I have to conclude that the sensors, being Maxim built 1Wire interfaces, have slew rate control and do NOT produce as much ringing on the data line and that the WC8 1Wire interface is only an open drain/collector interface, without any slew rate control components, and produces the nasty ringing on the data line. This assumption furthers my line of attack on solving some comm issues with 1Wire on the WC8 interface.

Click to expand...

Well said ;-) Did you experience a stuck 1-wire bus?  I have no idea why sensors get lost in operation....

It would be nice having some standard like JST SM connectors, also having Y-connectors...
http://www.pololu.com/product/2546

When electrical signal reached to the end of the wire, it will bounce back.  There is nothing 1-wire bus controller can do to get rid of it. That is like tossing a rock into a small pond, you will see the wave starting from where the rock went down and reached to the edge of the pond, then bounced back. 
 
Maxim IC published documents that being cited by Tom in #2, in those articles, it described how to reduce the bounce back by adding a little resistor, especially avoid the star configuration.  The idea is to let bounce back weaken so much, so that bounces will not cause bus change its TTL state.

The idea is not splash the rock in the water so hard so you don't make a wave that reflects at the edge of the water. Maxim shows a small capacitor across the data line to PS and a resistor in series with the data line. This should reduce the sharpness and the high frequency components of the signal level transition and reduce the reflection (standing wave).
 
I haven't experimented with this yet but I hope to soon after the Halloween season. This big in our family and it takes most of my time digging up a few of our relatives for a visit.

What would the Ohmage have to be fore the resistor?
I was trying to apply what I learned regarding Ohm's triangle, but neither does the Webcontrol manual give the Amperage of it's output, nor do the Temp sensors given indication of resistance.
I would imagine the end line resistor would have to be the strength of the Webcontrol output?
But, would I not have to know what the DS18B20 max amperage is?

I believe most use about 4.7kohm resistors for the end-of-the-line terminator.
 
I wouldn't bother with any of that yet. Your sensors should work without that and then later if you are running long cables or you find flakey reliability from the sensors then you can start attempting to clean-up signals.
 
Yes the WC8 interface, as well as every sensor on the 1Wire bus, needs to have enough power to drive everything on the bus.
 
To get a feel for this and some confidence just connect one DS18B20 to the terminals on the WC8 and get some PLC code goin for it. After you can get that going start introducing more sensors and extend wiring. If you have problems then start with termination resistors.
 
What humidity sensor do you have? Most of them are not 1Wire devices (in fact I am not aware of any) but rather analogue units and are not to be connected to the 1Wire bus.
 
BTW: In case this didn't become apparent: parasitic power means the sensor only has two wires, instead of three. Not compatible with the WC8 board because of the way it handles some of the signals.

TYVM for the information - the wires to each sensor should not exceed anymore than 10ft each, though once I am confident enough to try this inside the home (with around 6 sensors and a line probably 40ft long) I will have to be more diligent.
The Humidity sensor is the Honeywell 4000 which currently is working fine on the separate 3 wire outlet.
 
Electricity actually is fun to learn, but not easy. I will keep reading and thank any input!
 
Rainer