CNI and PC interface on the same network

I have a PCI and CNI interface next to each other on the same network. UNIT Ids 0 and 1 (neither is a clock or a burden)

The PCI interface works fine all the time. (CNI in the network or not)

The CNI works but slowly and times time fails during duplicate unit checking {error: network is not communication} . (using Toolkit 1.1 or old Cbus software with COM port redirect) when PCI Present

If I remove the PCI from the network the CNI works perfectly every time Quick and no errors during "Duplicate unit checking"

To the cbus network they both look like PCI interfaces I assume?

They both have firmware 4.2.00

I assume there are no issues with 2 PCI interfaces on the same network ??

 

Charlie,

We have a couple of networks here that are running both CNI's and PCI,s on them. So far no problems with this has been noticed. I have changed my unit address on both the CNI and PCI so they are the same as yours and both units will scan the network with little noticable difference. I have done this test with V2 and Toolkit.

Where do you have you burden located on your network?

 

Minder is the Clock and the Burden. I have no choice !!!

I might disconnect minder and see what happens. making one of these the clock and the burden

 

Get rid of the Burden off of Minder. It has been know to cause unrreliable network coms when connected. Also only enable one other clock on your network other than Minder clock and see how you go.

 

Stopping Minder being the clock

Mike

You were 100% correct. When I unplugged the minder from the cbus network both the CNI and PCI worked fine together. PCI was Clock and Burden. I then plugged in the Minder (live) and all continued to work ok. I then disabled the PCI clock Minder immediatly took over (flashing LED). and sure enough the PCI and the CNI stopped talking together. Plug either the CNI or PCI on alone and all worked fine. (burden in the CNI,PCI).
Well this is great news and easy to fix, But there is no way to stop the Minder supplying the clock. From power up with Clocks enabled on the PCI Minder always wins. The only way to get the PCI to become the clock is to unplug the CBUS cable from minder (Turning the power on minder off and on does nothing as the module is CBUS powered !). This is painful. If there is any cbus glitch Minder immediatly takes over. To get around this problem I have currently wired the CBUS via a minder output relay. Minder then takes is self off the cbus network at power on !! I can not automate this as I can see no where in the minder the clock status, so I can not tell when minder is the clock. If I wanted to get realy tricky I could wire up a sensor on a minder input module that watched the minder Cbus SIM LED. If it started flashing then the minder could takes is self off the Cbus network letting PCI become a clock. !! What over engineering!!!! There must be a way to stop minder becomming a clock. things like.... cutting tracks.. Removing crystal (this might be important for other things.....) Please Please if some one knows how do stop minder becomming the clock please tell me. (even if it involves soldering, cutting voiding warrantys) (send Offline if requrired)

Thankyou

Charles

 

Kill the minder Clock Fairy

System worked fine for 11 days without minder being the clock. CNI stopped working today... and sure enough Minder had become the clock.
Turn my minder relay that breaks the Cbus connection to minder and minder looses the clock and all is OK. For how long.........

There has got to be a better solution. THERE MUST BE A WAY TO STOP MINDER BEING THE CLOCK

Please Please if some one knows how do stop minder becomming the clock please tell me. Even if it involves soldering, cutting tracks removing components. Sending the Minder in to clipsal for a MOD, voiding warrantys (send Offline if requrired).

Is there another way to fix this problem ??

Charles

 

I have the same problem with a CNI and there is no PCi installed. I do have a bridge and a PCI installed on the other network. Would this count as the same thing.

My burden is on the first relay and clock is ??. Gotta be somewhere but not sure off the top of my head.

 

Charles

Sorry, but there is no way to stop the Minder from donning her mean step-sister uniform and taking over the clock. The same tracks on the circuit board that do the clock generation also do the data communication so you can't get rid of the clock signal without getting rid of the data at the same time. The clock and data leave the micro on the same pins.

The only things I can suggest are that you don't use the burden on the minder, turn off all burdens and use a hardware burden or you try a different Minder unit.

I have a small network running with Minder, PCI and CNI all co-existing and I can scan the network through all 3 devices. The network is reliable. The Minder is the clock generator and I have 1 hardware buren connected to my network. Errors/retries on the network are basically non-existent.

The Minder front-end C-Bus circuitry is quite antiquated and neither the most robust or the kindest to other units on the network.

 

Hardware Burden the fix.

thanks newman,
I wanted to exhaust all avenues before I over enginered the problem.
I currently have minder breaking the c-bus connection to Minder on minder "AC Power Loss" returning.

Thanks for taking the effort to setup in the lab. Dont know why you can not replicate. As I took the units out to the bench and tried too and still had problem

I tried a CIS hardware burden (the one that comes with PCI interface) and wired across minder. Disabled minder burden and HEY PRESTO it works when minder is the clock !!

The burdens seem to like being next to the clock generator. Is there a difference between Software and Hardware burden (Electronic effect on cbus ??) WHy does the Minder Burden not have this effect. I assume it is the same circuit with jumper connecting to cbus, Or is it a software burden with a jumper enabling.


Thanks

Charles

 

The fact that putting the burden directly across the Minder seems to fix the problem reinforces the speculation that the Minder is the cause of the problem. Placing it here means that the burden will have the maximum effect because the impedance of the cable is not reducing its effectiveness.

The component values used in the Minder burden and the plug-in hardware burden are the same, the difference is that the capacitor used in the hardware burden has a lower ESR (Equivalent Series Resistance for the technically minded) than that of the capacitor used in the Minder burden. Basically the capacitor type used on Minder is older. This means that the current hardware burden is more capable of handling any sharp spikes if they occur on the bus. As I've said previously the Minder front end circuit is quite old and I am speculating that the generated clock waveform is not as clean as what comes from our other devices.

The Minder has no Software Selection for the burden, it is a hardware jumper only. The difference between the Software Burden and the plug-in Hardware Burden generally is that the load applied by the software burden is a fraction less. The Software Burden also has a few extra components to do the switching. The Hardware burden in old 5100PC's and the like applies the same load to the network as the software burden.

 

Newman,

The CNI is powered by a external power. If this was on UPS or Battery supply and was set to be the clock(and was the clock) and the entire network lost power would it still be the clock when power returned (assuming the output modules also had the clock option set)

(No minder on the network. )

Charles

 

Charles

Yes, I think that would work quite nicely, because the CNI takes no power from C-Bus it is powered separately. If the CNI is continuously powered and has it's clock source enabled when all the other units go off-line due to the power failure the CNI will internally become the clock generator, even if it wasn't before.

When power comes back the CNI is already running as the clock generator so all the other units, even the Minder, would just pick that up and run with it because it's already there. I tried this half a dozen times to check and every time power returned to the network the CNI remained the clock source, even with a Minder on the network.

Good thinking, Charles.