ECG hyperDMX - the spec

j1sys

There are no rules, and those are the rules.
Community project designer
Joined
May 19, 2010
Messages
242
Location
Knoxville TN USA
Greetings -

Just a few notes about the new hyperDMX option soon to be available on the ECG-DMXRen8 and ECG-DR4.

hyperDMX is a combination of options that will allow up to 4 universes of data to be sent down a single RS-485 DMX line. It can be used for any use by any DMX controller designer but its principle use is for high pixel count LED protocol bridges that use DMX universes to drive pixel strings.

hyperDMX adds two options to the standard DMX slave software:

  • The option of using alternate start codes (ASC) to multiplex up to 4 universes over a single line. This is allowed under the DMX standard and will play well with other controllers that only accept the default start code of zero. The trade-off is that in order to send two universes you must allow for approximately 50ms updates, to send three universes 75ms, to send four universes 100ms.[br][br]
  • The option of increasing the transmission rate to a non-standard multiple of the standard 250Kbps. Depending on the distance between transmitter and receiver a simple RS-485 point-to-point connection is more than capable of working at much higher speeds and modern microcontrollers are capable of handling the faster rates. Being non-standard this option will not play well with other controllers on the same line. We feel that since its purpose is to transport multiple universes to a single controller which then drives up to 2048 channels (approx 680 pixels) that this deviation can be acceptably managed.

The approximate update rates that can be performed using variations of these options are as follows:

Warp 1*[br]250Kbps Warp 2[br]500Kbps Warp 3[br]750Kbps Warp 4[br]1Mbps
1 Univ25ms13ms9ms7ms
2 Univ50ms26ms18ms14ms
3 Univ75ms39ms27ms21ms
4 Univ100ms52ms36ms28ms

*Warp 1 is the only speed that is standard friendly

We fully expect that the ECG-DR4 will be able to support any of these rates on all four slaves simultaneously. So for an ECG-DR4 you could drive four hyperDMX capable pixel controllers each with four full universes at a 28ms update rate.

Due to design limitations and total bandwidth limitations we expect that the ECG-DMXRen8 will only be able to drive an aggregate total of 16 universes also. So with 8 slaves it would only be able to do a mix that totals 16 universes. One example would be: 4 universes on Slave #1, 4 universes on Slave #2, 2 universes on Slave #3, 2 universes on Slave #4, and 1 universe on Slaves #5-8. This would allow standard DMX devices on Slave #5-8 and hyperDMX compatible devices on Slave #1-4. Any other mix that does not exceed sixteen total universes should be acceptable. We will not limit you to 16 universes we just believe that you may start experiencing overruns and dropped packets if you exceed that number.

Of course neither the ECG nor the DMX controller will probably be able to detect the transmission speed that is set and therefore must be independently configured to operate at the desired speed. This may lead to some additional confusion during hookup and testing.

Well, enough for now, back to work on the software that makes this all work ....

-Ed
 

mschell

Full time elf
Joined
May 17, 2010
Messages
155
Location
Southeast US
Nice stuff. It's amazing what some smart folks come up with...

Of course, this means that I'll have to come up with lots more $ to fund the pixels and controllers, etc...but at least I already bought the DMXRen8! :D
 

Cypress1138

New elf
Joined
Oct 7, 2010
Messages
1
Its funny but I have been thinking that with RGB pixels and their associated high (insane?) channel counts that DMX may not be the protocol to use. But then once again people vastly smarter than me are way ahead and are already working on a solution! This will be very interesting to follow!

I agree with Wombat and bow to the Master Po
 

j1sys

There are no rules, and those are the rules.
Community project designer
Joined
May 19, 2010
Messages
242
Location
Knoxville TN USA
E1.31 direct to the pixel level is the great way to go but not ready (from me) right now and this will help people quadruple their capacity on TP3212 units RIGHT NOW if needed.

It will always be part of the communication/wiring arsenal and there are place's where it may be used for years to come.

There is not always one 'right' answer to fit all solutions.

Someone else, either here or on chat, mentioned DVI as the high pixel protocol of choice for the pro world. All protocols should fit the solution, not the other way around.

-Ed
 

j1sys

There are no rules, and those are the rules.
Community project designer
Joined
May 19, 2010
Messages
242
Location
Knoxville TN USA
IT'S ALIVE!!!

Just got hyperDMX transmitting 16 universes (4 per slave) out a DR4 (into the bit-bucket) at 25ms updates from Vixen.

Next to retro-fit to the DMXRen8 and test it's limits.

There is now a Warp5 for 1.5Mbps to give a little more headroom for 25ms updates with 4 full universes.

So v1.3 with hyperDMX is coming soon to an ECG near you!!!

-Ed
 

budude

Way behind schedule - again...
Joined
May 13, 2010
Messages
127
Location
San Jose, CA
Geez - - I'm almost afraid to get on this forum anymore - I feel like anything I have now is from the Stone Age! :) Next year should be a lot of fun...
 

JTR

New elf
Joined
Jun 2, 2011
Messages
7
Location
Geelong
Just goes to show that I came into this game a little too late. All my great ideas have been stolen from me even before I knew I had them! :D

Just remember though. I had WARP-13 back in 1999.

I have been playing around with multiple universes and 1M baud rates and doing all sorts of math with my mock-ups. Not too seriously though as to be honest E1.31 is the way things are going to go. Still just like the LIN bus still has its place under CAN, DMX/RDM will for a long time yet still exist both natively and as a sub layer to E1.31.

-Jim
 
Top