Smart String Coop

[castortiu]

I would think about 1100 nodes, if the coop is too early in 2011 then I would have to skip it and hope there is a second coop later.

I would like to use more but I’m in doubt how they will work in a big scale.

Hardware: 20ft mega tree = 240inches / 3.5inchesXnode = 68nodes per string * 16strings = 1088 nodes TOTAL

Data: 1100nodes * 3chXnode = 3300channels x 40FPS (frame per second 1000ms/25ms) = 132000frames x 10bitsXframe = 1320000 = 1.3Mb

So the data throughput needs to be at least 1.3M baud rate.

I would like to see how Vixen behaves at that rate since rendering the UI and keep a constant 1.3M baud can be a challenge without start to get UI/Audio/Data unsyncs.

Anyway I think Vixen is not the solution for Nodes since create sequence can be a challenge and we have xLights already that can provide the networking infrastructure, so I guess Vixen won’t be a candidate for SSs.

Cas.

[RJ]

Cas,

 I am running just short of that right now on my show using Vixen and it keeps up easily. Plus it is kicking out 512 channels of DMX at the same time.

Phil has twice that on his running from Vixen.

Of course since you guysl have better software for us anyway this is not an issue right. 

RJ

[batdive]

Cas,

Of course since you guys have better software for us anyway this is not an issue right.  

RJ

             


And the proverbial gauntlet has been thrown.



-JS

 

[jandelee]

In for about 150 nodes...'course it seems that will be in the noise...

John

[csf]

Cas and RJ Please correct me if I am wrong. But doesn't dmx only have to give out change commands, the dongle just keeps giving out the same signal until a change for that channel occurs.

So once a light is set to turn green, it no longer needs a command until it needs to change to blue. Then The green will be turned off and the blue turned on.

If this is correct, and pixel net building on top of dmx, stile keeps this feature then net work load will most of the time be allot less.

Now granted I could be wrong on all this, Matt did all the protocol coding on xlights, since it is something I have never done before, but I did research dmx protocols and I believe what I said above is correct.

Edit:

BTW lor now supports 16 networks giving  65280 channels, if they can manage I am sure we can too

[castortiu]

In simple terms this is what DMX does.

There is no language or commands for DMX, the protocol is a very simple asynchronous protocol and consist of sending raw frames to the hardware, there is a about 11bits per frame/channel with no error correction from which 8 bits are used as data.

This makes DMX a very simple protocol to be implemented by hardware without the need of an interpreter or a common language. So the baud rate must be constant, if there is a problem in the transmission the hardware might decode some frames incorrectly, but then 25ms or 50ms later comes another frame with probably the right data.

So there is a burst of 512 or 513 frames then a break then another burst of data (DMX512). So if you have hardware programmed in the channel 100, then the RX the only thing it does is wait for the “start” signal and discard 99 frames and take the frame 100.

DMX does not care about errors in the transmissions, so you might have interference and probably you won’t see the difference in your show also because is unidirectional is ideal for broadcast, if the protocol would implement commands or a language then a bidirectional line may be needed and extra CRC checks making the protocol way more complex.

My question to RJ was because usually DMX  go up to 250Kb, when now with pixelnet looks like we are talking at least about 1.3Mb, and I wonder how Vixen was doing to keep the QOS for 1.3Mb other than just expecting to have a fast CPU, if it just relies on just the CPU idle time then other application may starve the CPU for several ms and Vixen will get out of sync, but so far from RJ comments that’s not happening.

I don’t know how the “Enttec DMX USB Pro” plug-in talks with the dongle, so it may be sending only the changes and the dongle keeps sending DMX frames to the hardware until more changes comes from the plug-in, I have no idea what is involved between the plug-in and the dongle. If it is pure DMX then the process is like I explained above.

Cas.

P.S.: I think the subject is getting out of track, we can create another topic about it.

[rrowan]

Hi Cas

The Lynx Dongle and the Entec USB Pro do all of the hard work of timing and keep sending out data. Vixen or any sequence software just sends updates when a change happens. If you use the non- pro version of Entec dongle than your computer needs to keep sending out the data and you will see a lot of data errors do to the way Windows handles the interrupts to the com port.

I hope that explains it

Cheers

Rick R.

[mcalderaro]

500 nodes

I plan on using them for 10' mega tree.
16 strings with nodes every 4 inches, approx 30 nodes per string

[ron d]

Im in for a mega tree and more for sure. Im thinking 120 nodes or so per strand on the mega tree and do an up/down segment for each strand so I have 8 more outs on the hub for other options. I should be easily able to assign the channels to work easily when sequencing.(there will be no real easy sequencing)
Anyway wouldnt that make the most efficient use of nodes / hub limitation ?

[rrowan]

In the vixen/LSP plugin for the E131 gateway (EthConGateway - You are not allowed to view links. Register or Login), it has an option to change the output speed (I was running mine at 40ms) and how often to "refresh" the output on channels that don't have changed data (like in the LOR protocol) - mine was every 4th cycle.  So, it would appear that you can offload some of the work into the hardware if it supports that (like the Enttec Pro vs the open).

Well Dave
I am confuse, please explain to me how the above info has anything to do with SS plus you are bring up a older post that RJ has already answer. Is this another round of Dave pushing commercial products on the DLA community?

Please stop pushing your and your buddies commercial equipment. You can post it and the vendor section if you have the need to.

Thanks

Rick R.

[tonypgst]

Depending on my negotiating skills, I'm looking for between 0 and 2000 nodes.

[RJ]

I would think about 1100 nodes, if the coop is too early in 2011 then I would have to skip it and hope there is a second coop later.

I would like to use more but I’m in doubt how they will work in a big scale.

Hardware: 20ft mega tree = 240inches / 3.5inchesXnode = 68nodes per string * 16strings = 1088 nodes TOTAL

Data: 1100nodes * 3chXnode = 3300channels x 40FPS (frame per second 1000ms/25ms) = 132000frames x 10bitsXframe = 1320000 = 1.3Mb

So the data throughput needs to be at least 1.3M baud rate.

I would like to see how Vixen behaves at that rate since rendering the UI and keep a constant 1.3M baud can be a challenge without start to get UI/Audio/Data unsyncs.

Anyway I think Vixen is not the solution for Nodes since create sequence can be a challenge and we have xLights already that can provide the networking infrastructure, so I guess Vixen won’t be a candidate for SSs.

Cas.

In the vixen/LSP plugin for the E131 gateway (EthConGateway - You are not allowed to view links. Register or Login), it has an option to change the output speed (I was running mine at 40ms) and how often to "refresh" the output on channels that don't have changed data (like in the LOR protocol) - mine was every 4th cycle.  So, it would appear that you can offload some of the work into the hardware if it supports that (like the Enttec Pro vs the open).

I believe he was talking about the Lynx Smart Strings since he said PixelNet not E131 and since he posted it here in the Smart String thread and not the EthConGateway thread. It was also already answered.  Lets keep the thread on topic of Smart Strings. I believe all my users are familiar with the other ways to do it. If not 100 more post with links and videos will not help. I sure have no problem with them being free to choose, I make nothing from it so I have nothing to lose. 

RJ

[keitha43]

736 nodes for me. I finally figured out I had to post in the new members thread first to post here. I am figuring about 46 nodes per string approximately 13 feet for a 11.5 foot megatree. Although I may run a string up and then back down to cut the controllers in half if it will be easy enough to sequence. Looks like I will need a new dongle also because mine is a different brand.

[wwwgator]

Count me in for at least 680 nodes- maybe up to 1000, depends on when the coop starts.
Thanks!

[rogerwh]

Money tight but I would like to try at least 128 nodes.