1-wire over CAT7 - CRC errors

[DanF]

Hi,

Now I'm done with all my CAT7 runs to each room for tree devices door/window contacts (using the spare pairs) I've moved onto 1-wire.

I've seen it mentioned that minimum AWG 22 is recommended for 1-wire, also somewhere it suggested that unshielded cat cable was preferred, but I'm using the CAT7 cable (AWG 23) I have on-site.  I am avoiding parasitic wiring.

I have my sensors in clusters of 2-3 and each (Loxone) temperature sensor has a 3m lead.  I'm using CAT7 to hop from cluster to cluster creating a "bus with short stubs" topology.  The only caveat is that, in most places, the stubs form a small star with 5-way's connected to bus-in, bus-out, and the 2-3 sensor leads.

So far I'm up to about 40m and have connected 8 sensors. I am hoping though, to get a full 20 sensors on the bus and not be limited by length, even if I end up with around 150m.

At the moment I'm using the blue pair for Data/GND and the green core for VDD. I haven't (yet) done anything with the green-white core or the shielding, but I may experiment with this later today.

Loxone discovers and gets temperature readings from all sensors easily, but I'm seeing a fair number of CRC errors from some of the sensors.  It seems that the furthest sensors give the most errors, but some of the closer ones also give random CRC errors too.

Any thoughts?  Has anyone got a tried-and-tested approach to best using CAT7 cores and shielding to get the best length/reliability?  Or has anyone found that unshielded CAT5/6 is actually better? 

Thanks!

[Duncan]

1-wire isnt a balanced (twisted pair) bus, so try connecting the shield to the GND and see if things improve

shielded cat7 will work better than unshielded for 1-wire and it has a lower capacitance and therefore less 'weight' or virtual distance

there are 2 likely possibilities for your errors:

1) wiring topology, particularly if you have long stubs and/or physical connectors where the stubs join, which will make things much worse - can you solder the joints?

2) its common to get some duff sensors - get some spares to swap out the ones with crc errors (or move them to a different place to see if the same sensor still produces errors)

[Duncan]

some homework for you ;-)

https://www.maximintegrated.com/en/design/technical-documents/tutorials/1/148.html

[Daniel Feist]

1-wire isnt a balanced (twisted pair) bus, so try connecting the shield to the GND and see if things improve

shielded cat7 will work better than unshielded for 1-wire and it has a lower capacitance and therefore less 'weight' or virtual distance

I'll try two things:

- Twisting VDD with GND.

- Connecting the shield to 1-wire GND at the Loxone end.

1) wiring topology, particularly if you have long stubs and/or physical connectors where the stubs join, which will make things much worse - can you solder the joints?

Stubs are no more than the 3m sensor lead, but I am currently using connectors, yes.   This is what I have currently.  This repeats itself at different locations along the bus.  Any solderless suggestions for doing this tidily while minimizing noise/capacitance?  Would teeing off the bus 3 times instead of one help in any way (it would be more work when sensors are close)?

Thanks,

 

2) its common to get some duff sensors - get some spares to swap out the ones with crc errors (or move them to a different place to see if the same sensor still produces errors)

On Sunday, 28 February 2021 at 13:18:21 UTC DanF wrote:

Hi,

Now I'm done with all my CAT7 runs to each room for tree devices door/window contacts (using the spare pairs) I've moved onto 1-wire.

I've seen it mentioned that minimum AWG 22 is recommended for 1-wire, also somewhere it suggested that unshielded cat cable was preferred, but I'm using the CAT7 cable (AWG 23) I have on-site.  I am avoiding parasitic wiring.

I have my sensors in clusters of 2-3 and each (Loxone) temperature sensor has a 3m lead.  I'm using CAT7 to hop from cluster to cluster creating a "bus with short stubs" topology.  The only caveat is that, in most places, the stubs form a small star with 5-way's connected to bus-in, bus-out, and the 2-3 sensor leads.

So far I'm up to about 40m and have connected 8 sensors. I am hoping though, to get a full 20 sensors on the bus and not be limited by length, even if I end up with around 150m.

At the moment I'm using the blue pair for Data/GND and the green core for VDD. I haven't (yet) done anything with the green-white core or the shielding, but I may experiment with this later today.

Loxone discovers and gets temperature readings from all sensors easily, but I'm seeing a fair number of CRC errors from some of the sensors.  It seems that the furthest sensors give the most errors, but some of the closer ones also give random CRC errors too.

Any thoughts?  Has anyone got a tried-and-tested approach to best using CAT7 cores and shielding to get the best length/reliability?  Or has anyone found that unshielded CAT5/6 is actually better? 

Thanks!

--
You received this message because you are subscribed to a topic in the Google Groups "Loxone English" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/loxone-english/PqPn2sSWE68/unsubscribe.
To unsubscribe from this group and all its topics, send an email to loxone-englis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/loxone-english/cf1a3459-548f-41e9-bb35-d334de5c791fn%40googlegroups.com.

[Duncan]

i dont think there are solderless solutions that are going to be any better or worse unfortunately - given that you have a lot of 3m stubs its worth trying the stub resistors as shown in the pdf, nothing to lose but a few pence!

[DanF]

I grounded the sheilding and the errors went away!   At the weekend I'll add in the rest of the run and see what happens.   Total is about 120m with 20 devices.