Cbus Network Issues.

I'm hoping one of the cbus gurus here can help me out, I have a cbus install that's going on 10 years old now and over the last month or so i have been having some gremlins in the system.

I have been having issues scanning the network with some units failing to scan although its never the same unit each time I scan, I have been able to get around this by adding extra burdens, I'm up to 3 now (1 software and 2 hardware) Not ideal I know but it works for the time being.

My CTouch MKII randomly reboots.

After a power failure or disconnecting the network cables from the units with built in power supplys the network takes a while to settle down E.g if you try to use a switch straight away it just flashes quickly and the load wont switch on, however after 5 min or so this goes away.

Lights randomly turn off and on, sometimes in the middle of the night which is always fun.
I have had cbus toolkit running for a few weeks logging the traffic but haven't been able to capture anything, of course it wont do it now I'm watching.

I have noticed in the last day whist fault finding the touch screen and the EDLT back lights pulse slightly sometimes (this behavior comes and goes) leading me to suspect a faulty PSU but checking the output they appear to be functioning correctly individually, each supply is outputting between 30 and 32vdc and are balanced within half a volt between cbus +ve and earth and cbus -ve and earth.

My next step is to start segregating parts of the network to try to isolate a possible faulty unit but i am hoping to avoid having to do this just yet.

Any advise would be greatly appreciated as I'm at a loss on what to try next.

 

Firstly, this will never work. It's just hiding the fault. Go back to one (or even none if it's a big network).

First thing to check is power supplies. Check the voltage at various places in the network with everything running. Checking an unload power supply is pointless. Run the diagnostic utility and get it to report all unit voltages. Look for anomalies. Power supplies can also go noisy. Easy to test with an oscilloscope but that's probably not an option. Try disconnecting each power supply at a time, ensuring you still have enough current available for the network. Run the diagnostic utility and watch the log. Look for unusual behavior like a unit transmitting regularly. Run a load test with the diagnostic utility and look for unit errors. Any unit can go bad and upset the bus. Cable faults can also cause these behaviors. Moisture ingress or even old age of a marginal made connection can be the cause.

Ultimately you will have to segregate the network. Unfortunately most installers don't think about this which can make it challenging. It's just a bit of well planned detective work.

 

Thanks Ashley for the quick reply, I have been running the diagnostic utility for the last few days and voltages seem to be pretty consistent with a min of 24.5 and a maximum 29.8 unfortunately I don't have the power budget to spare at the moment to have any unplugged for testing as with just 1 of the 4 unplugged the voltage dips to 13v.
I will to try and source a spare power supply today to try some more tests this week.

I have have been running a network reliability test over the last 24hours but nothing pointing to a single unit to being faulty due to the poor network condition most units are coming back as marginal condition.

I did put a scope on the power supplies, one seems fine but the other 3 all appear to have some noise present but not knowing what is acceptable amount if any its really not a helpful test at the moment, 1 is worse than all the others so I might try that one first.

 

Hi
Everything Ashley said plus check voltage between cbus + and ground(earth) then cbus - and ground(earth) it should be about half of the voltage between + and - if greatly uneven one leg is going to ground via water or damaged cable also I have found by measuring the voltage at the units with a power supply and no cbus cable connected the no load voltage would normally be around 38 volts if it is over 40 that power supply is most likely faulty. The best way for testing cbus network is with your PC with one burden and when it scans clean all is ok.
When I am looking for network issues I start with outside devices first as they will be the ones with water entry issues.

 

Other points to consider
input units in bathroom, kitchens check the terminations are clean and tight.
If you have seen insects (ants in particular) like to reside inside electronics.
Patch leads between output unit sometimes the RJ45 plug isn't making great contact, take them out and clean them.
With regard to your voltages, though you are within spec. Check to see if there a considerable volt drop between two adjacent units, indicating a poor connection. I can only presume from what you have said that your power supplies are running close to there maximum load, (system voltage in that range is the clue).
Hence power supply issue would be high on my list, if the system has run faultlessly for 10 years. Though do check for deterioration of terminations

 

Hi All,

Just jumping on this thread as I've got a clients system which is approx 20 years old. Not overly large 10 or so dimmers and relays and 30 or so switches and PIRs. Having a devil of a time getting the network to be stable unless I disconnect the field wiring. (Unfortunately it was wired by someone clearly trying to create some kind of maze, one network but jumps all over the place.

Didn't quite a lot of time checking clocks but testing this I'm wondering if 1 or more of the power supplies has gone noisy. I've replaced one in the past and it's possible that one of the dimmer or relay ones is now creating noise.

What wave would it be looking for on a fully functional power supply?

Lastly for now I want to check the clock pulse whilst I have the scope on site, is it a square wave or saw tooth in looking for?

Cheers all

 

I was talking to cbus tech support last week and they gave me some pointers on checking for a faulty powersupply you might find useful.


Here are the steps to test those inbuilt power supply with C-Bus Clock and measure C-Bus voltage using multi meter.

Steps to test the inbuilt Power Supply

Separate C-Bus Clock and C-Bus Power Supply from those output units

1. Scan network

2. ensure C-Bus Clock enabled on the PC interface

3. ensure no C-Bus Clock enabled on all output units

4. close network


Use C-Bus Clock to Test

1. Test one output unit at a time, disconnect an output unit from the network

2. C-Bus indicator on the output unit is OFF once disconnected from the network

3. C-Bus indicator on the output unit is ON once connect the PCI to the output unit

4. keep watching the C-Bus indicator on the output unit and remove the PC interface, if the C-Bus indicator goes Off straightway, then the inbuilt Power Supply is OK, if it flashes or stays ON for a while and OFF, this inbuilt Power Supply is faulty, replace the output unit.



Measure C-Bus voltage and check balance

1. test one output unit at a time, disconnect an output unit from the network

2. measure C-bus voltage: + to -, it should get 34.5Vdc. NO good if it is too high, replace the output unit

3. check the balance: measure + to E and - to E, NO good if the deference > 1 V, replace the output unit

 

Thanks for this Matt that is very helpful. I had been concentrating on the clocks rather than the power supplies so this gives another avenue of investigation. Cheers

 

With regard to the wave form see the attachments
You should have a square wave with 5v peak to peak
On a live network you will see other waveforms as communications take place on the network.
Note these screen grabs are of an O/P unit not loaded

 

For some reason part of my post got deleted.
The First picture above the O/P unit has a burden connected the other one has no burden, hence the slow decay of the wave form and would make communication very difficult.

 

Thanks Wonkey, I'm onsite tomorrow will grab a scope and see what I can find

 

Thanks Wonkey. Am onsite now have tried the clock tests and checked voltages that Matt detailed. Everything appears to be ok. Getting voltages of mostly 35.2 as the highest. Is this too high? About to see if I can get a scope on the network

 

Have you used Reliability test and traffic analyser within the diagnostic utility, Ashley mentioned this and it can help in identifying issues.
Ashley, Graham and I all made suggestions earlier in this thread, can you provide some feedback, we all know and understand that it can be time consuming, but unfortunately many issues are not discovered using a laptop.
One tip when you start pulling stuff apart document what you've done, or you may end up going round in circles.
Also did you get hold of another power
supply.
Post your result so the community can help.
Colin

 

Afternoon from the UK Colin,

I have tried to use the reliability test and traffic analyser but the network is so unstable that it struggles to even connect. As it happens I'm onsite right now but haven't been since Monday.

All the suggested power and clock tests proved inconclusive and I'm now just looking at the network on a scope. Shall upload images of the waveform shorlty

 

For some reason I can't upload another small video but at the end of this clip it shows the network "dropping" I would suggest.

Right what I've been upto here:

Tested each output unit as per Matt's detail from Aus Tech support. All units were showing between 34.2 and 35.6 all unloaded just them connected with the pci nothing else.

Checked the clock type operation l, again as per Matt's email and all indicators came on as you would expect for a correctly working clock.

Attached the scope and with only the dimmers connected am getting a stable wave.

Therefore deduced that it's likely to be field wiring and field devices.

Once I can get it to run with field devices I shall run the reliability program

 

I presume that the picture of the waveform is from the complete network. As I zoom in it the clock wave seems OK, but there is some ripple(noise) between each pulse.
How many burdens did you have connected.
Some simple questions.
Is the installation in a home, factory or office,etc.
Do check C-Bus + and - to earth for the complete network,.
You mentioned a maze of a network, given that info I guess there are many joints in the cable, check to see if the cable loops back on itself, easily done disconnect the wires at a joint and one end should have voltage the others none.
When you mention getting the network to be stable what issues are you experiencing.

Has something changed in the building that could induce noise onto the C-Bus network, look for inverter type devices with they cabling running close to C-Bus cables, try turning off other breakers and see if you can run the diagnostic utility.

 

What happened when you removed the PCI? did the cbus indicator go out immediately or stay on for a few seconds? if so may indicate a faulty PSU. I had 2 such units that did that, I have only been able to replace one at the moment being that there is not much stock available at the moment, I'm hoping to send the one I removed away for repairs so I can replace the 2nd failing unit.

I have this fault finding PDF from cbus tech support that might also be of help.

I was also able to put a scope on my network this afternoon not sure if its good or bad though

 

I presume that the picture of the waveform is from the complete network.
*** It's not the whole network, just the output units and a few of the switches***
As I zoom in it the clock wave seems OK, but there is some ripple(noise) between each pulse. *** exactly ***
How many burdens did you have connected. *** 1, a physical rj45 one ***
Some simple questions.
Is the installation in a home, factory or office,etc. *** Home ***
Do check C-Bus + and - to earth for the complete network,. *** Did that for each output unit. All of them were +/- 16. Never more than ½ volt difference between the two ***
You mentioned a maze of a network, given that info I guess there are many joints in the cable, check to see if the cable loops back on itself, easily done disconnect the wires at a joint and one end should have voltage the others none. *** I actually thought I had a reasonable handle on the network layout, but Thursday evening, when I read last there, I'm still finding odd connections which could lead to a loop. Typically that's almost at the furthest point of the network ***
When you mention getting the network to be stable what issues are you experiencing. *** If I connect up too many switches, the network becomes unresponsive, switches won't change state, or the switch changes state but the output doesn't trigger AND scanning the network falls over almost immediately with unresponsive units ***

Has something changed in the building that could induce noise onto the C-Bus network, look for inverter type devices with they cabling running close to C-Bus cables, try turning off other breakers and see if you can run the diagnostic utility.[/QUOTE] *** nothing that I've been told about it can see has changed. There is evidence of internal water damage but it appears to be remote from c-bus and it's wiring. Haven't seen any evidence of water ingress in any of the switches etc. My gut feeling is that one or a number of units have begun to fail and are putting noise onto the network. I tried to download a video of the show showing the network on the scope, showing it collapsing for a short period when I connect up more of the network. I'm back there Tuesday and I'm going to carry on tracking down the network layout and testing each switch against the scope as it's connected

 

What happened when you removed the PCI? did the cbus indicator go out immediately or stay on for a few seconds? if so may indicate a faulty PSU. I had 2 such units that did that, I have only been able to replace one at the moment being that there is not much stock available at the moment, I'm hoping to send the one I removed away for repairs so I can replace the 2nd failing unit. *** Each time I disconnected the PCI with the clock on, the output unit CBUS light dropped out immediately. Plus all the voltage checks on individual powered output units were showing his colleagues as per your previous details. Shame as it sounded like the perfect answer to my conundrum***

I have this fault finding PDF from cbus tech support that might also be of help. *** Thanks Matt I'll take a look and use next time onsite, Tuesday ***

I was also able to put a scope on my network this afternoon not sure if its good or bad though [/QUOTE] *** using a scope too which is useful***

 

Gordo: From all the info you have provided as Gordo mentioned PSU issues look like a serious possibility or as you mention failing units.
From my own experience Key inputs that are failing normally recover for a short period of time after they have been down powered,.
Spend the time and get the wiring correct and then remove a number of units off the network (say 10 units) but leave the wiring connected. If the system runs OK, reconnect the units and repeat the process but with some other units disconnected again ensure the cabling is still connected.
Should everything work again you can be reasonably sure its a PSU issue.
Check individual PSU as mentioned before and if possible use the scope on each one and if they are an output unit with and without clock enable and compare results.
Final idea for now change the unit that is running the clock .
Matt I noticed that the clock pulses are not equal positive and negative and there is some ripple on the peaks. This could be the picture or scope or a poor clock pulse being generated.
Check it again disconnected from the network and also try a clock from another unit. At some point in C-Bus history (long time ago have forgot when) the clock pulse was improved so if you have newer unit that can generate a clock use that, if the install is old.
Just out of interest none of these installs iare using the old White box PSU or PCI?