Dual Counter False Counts due to noise

[Mark42]

I thought I would pass on some recent experimental information about using the dual counter +5V as the voltage source for the channel counting.

I an using a TEMP08 (with 200mA power supply) on a single daisy chain 1Wire network about 100' long constructed with CAT5 cable. The network has 2 1WIO cards and a single Dual Counter card located at the end of the 100' daisy chain. The TEMP08 is tied into a HomeSeer Hometroller 2. The dual counter was connected to a breadboard application circuit via a 4N25 opto-isolator; the counter card +5V was tied to one side of the 4N25 detector and Channel B input to the other side.

The dual counter card appeared to be 'seeing' excess counts associated with household electrical noise events like the furnace circulating pump stating and stopping or the oil burner starting and stopping. I initially assumed this to be caused by the application circuit I was working on and made a number of modifications to its +5V and +12V power supplies; these had no effect on the false counts.

After much incremental testing I performed the following test; I tied the counter card +5V directly to the Channel B input completely eliminating my application circuit; since the counting occurs on high to low transition, the channel should never count. However, when the furnace start/stopped the false counts continued to be generated (also whenever I use the electric pencil sharpener).

I then externally powered the dual counter with a battery pack and performed the same test as above; the false counting stopped. Since the application circuit has a +5V power supply, I am now externally powering the counter card by connecting the application circuit GND to the dual counter ground and the +5V to the dual counter +5V. I was initially concerned that the addition of another ground would compromise the 1Wire communications but this does not appear to be the case.

Bottom line, the dual counter is no longer reporting false counts during household electrical noise events.

Best Regards,

Mark

 

[sda]

Doesn't this point to the TEMP08 or its power adapter having voltage fluctuations causing the +5V supply to dip to the point of triggering the counter?

Did the counter increase by 1 or many?

Did you try a different power adapter? Putting the adapter on a UPS?

 

[Mark42]

<<Did the counter increase by 1 or many?>>

The counter would typically increase by 1-3 counts when the furnace would start and stop (a frequent occurrence at the moment in NJ). The pencil sharpener I used to force the false counting would cause 2-3 counts every time the motor started/stopped.

<<Did you try a different power adapter? Putting the adapter on a UPS?>>

The adapter is on UPS power. I did not try another adapter. The UPS powering the TEMP08 and the noise sources (furnace and pencil sharpener) are on the same phase but all on different circuits and in physically different areas of the house.

<<Doesn't this point to the TEMP08 or its power adapter having voltage fluctuations causing the +5V supply to dip to the point of triggering the counter?>>

The dual counter card only has 2 connections, GND and DQ. I believe the +5V available on the counter card is somehow derived from the DQ communications power (parasitic?). If you look at the +5V on the card with an oscilloscope, it is more or less +5V but you can see it changing, presumably from the 1Wire communications.

The problem may be the TEMP08, the power adapter, the length of the 1Wire, the routing of the CAT5 cable, the counter card, or a combination of all of the above. But, at least in this case, the path of least resistance was to externally power the counter card.

Mark

 

[sda]

The dual counter card only has 2 connections, GND and DQ. I believe the +5V available on the counter card is somehow derived from the DQ communications power (parasitic?).

The Hobby Boards dual counter board
http://www.hobby-boards.com/catalog/product_info.php?cPath=23&products_id=42
has 5 connections, GND, DQ, +5, A, B; and it has a battery to power it even if +5V is missing.

The TEMP08 has +5V on pin 2 of its RJ12 connector.
http://www.midondesign.com/Documents/1-Wire%20RJ-12%20Wiring.PDF

If you don't send +5V to the board, the board will run on battery power. When the battery runs out, it'll be only on parasite power. On parasite power, the DS2423 will have minimal functionality, it'll identify itself and that's about it.

 

[Mark42]

<<has 5 connections, GND, DQ, +5, A, B; and it has a battery to power it even if +5V is missing.>>

You are, of course, correct. When I said 'only two' (GND and DQ), it was in reference to the three 1WIO 1Wire module screw terminals; GND, DQ, and +12V.

<<The TEMP08 has +5V on pin 2 of its RJ12 connector.>>

Ahh, so it does. I only have 1WIO modules in the 1Wire network besides the Dual Counter card; the 1WIO modules only use +12V, I never considered the +5V. Fortunately, when I converted from screw connections to RJ connectors the +5V came along for free.

I hooked up and measured the +5V from the TEMP08 and then performed the prior tests. Unfortunately the counter is still responding to noise events. When I revert back to powering the counter card locally, the noise susceptibility goes away.

I believe I will stick with my current solution.

Thanks for the heads up about the +5V.

Mark