ECG-DR4, ECG-M32MX, ECG-PIXAD8 Production & Shipping Notes

[j1sys]

Greetings All -

It has been a busy last three months around here working on pushing some production, product updates, and new products through the pipeline.

On a personal note, mixed in with all that was a cervical spine surgery, hospital stay, and recovery.

I wanted to update you on the status of several of our products and pre-announce a few coming rapidly down the pike.

The ECG-DR4s are finally tested and ready to ship. We will be sending email invoices to everyone in the queue today and tomorrow and start shipping them on Monday 4/11/11. We do not have the ARE in place so our Australian friends will need to get direct shipments.

The M32MX pre-production run is a great success. We have 5 of the boards sitting here tested and humming along and are placing a large production order today. We expect them to be ready in about 3 weeks. These will be used to send, gratis, as an upgrade to all ECG-DMXRen8 owners, as stock for new ECG-DMXRen8 orders, and as the master controller for several new products in the pipeline.

With some of that said we would like to open the curtain and let you know a little about some of the things that are coming very soon. If the new items cause some of our current ECG-DR4 'tentative' orders to no longer be needed we understand and there is no problem if people want to change their minds for this or any other reason on 'tentative' orders they have placed via email so far this year.

We expect to receive a prototype run of ECG-PIXAD8 PCB early next week and will build them up and test them and post photos as rapidly as possible. The ECG-PIXAD8 is a 3oz copper PCB signal and power adapter board that sits under an ECG-M32MX to provide fused power distribution connections to 8 pixel strings. The strings are split into two groups of 4 with a fuse on each output as well as the input of each group. The traces and parts are spec'd to handle 8 amps per string and 32 amps total per group. We will probably provide 7.5amp and 30amp fuses or possibly offer the more conservative group of 5amp and 20amp fuses. It uses standard automobile ATO fuses so field configuration and replacement should be easy.

We plan on initially supporting 6803, 2801, and 1804 pixels and will support either ArtNet II or E1.31 inputs. We will provide support for other pixel chips as demand warrants and time allows. The system will include our usual html configuration screens.

So as soon as we have confidence in our prototype we will order more PCBs. We are considering offering this as a kit and/or assembled and tested. It uses all through hole parts and should be a quick 1 hour build. The PCB is 5.65in x 3.85in with the ECG-M32MX sitting on top. All of the connectors for power and distribution are along one of the long edges.

We are considering a slightly smaller (and a little less expensive) ECG-PIXAD4 for the times when you only need 4 strings of control. Customer comment would be appreciated. We are also designing an enclosed ECG-PIX4 in a similar case to our DR4 units. This is for the more consumer oriented clients that are scared off by bare board technology.

Everyone wants to know the price. We are targeting $90-100 for the ECG-M32MX/ECG-PIXAD8. The ECG-M32MX/ECG-PIXAD4 should be about $15 less.

We are in final design for our first pro offering. The ECGpro-DMX1 will be a small (6cm x 8cm) aluminum encased product available in either POE or external adapter models. It uses Nuetrik XLR style RJ45 and 4 pin connectors and includes transorb protection on the DMX signal lines. We hope to have working prototypes in the next 30 days.

Thanx for all your interest in our products. Stay tuned, you ain't seen nothin' yet!

-Ed

[chrisl1976]

Sounds great Ed.

With the 8  outputs, is that going to be 1 universe per output? (170 rgb node)?

[ɟɐsʇǝppʎ]

Excellent news Ed, looks like you have been a busy man and we look forward to the ECG-PIXAD8 & 4

[j1sys]

Initially we will probably just map one universe, or part of it, to the string.

Very soon we will have the universes form a pool of data that can be used in any size block with any starting point and it will 'roll' over into the next universe in the pool.

There are no technical reasons strings of 3000 pixels, with their own power injection layout, could not be driven. however the user must consider update rate etc. The html pages will provide estimated timing information for the different protocol and speeds selected.

Like many of the pixel solutions this solution will be have two SPIs round robin servicing the 8 stings (4 each). We have a design on the boards that will have 8 dedicated SPIs (1 per string) that could be used if high bandwidth updates and LARGE string arrays require the extra horsepower.

I forgot to mention that we could also scale up the PIXADx series and produce a PIXAD16. So you will have the option of 4, 8, or 16 string adapters down the road. Each one driven by an ECG-M32MX.

The M32MX is the most powerful master processor board we have built. It is spec'd for UDP transfer by Microchip at 8723 KBps (kilobytes per second) vs 2115 KBps for the processor/ethernet combination in the DR4. The processor has 512KB of flash and 128KB of RAM for programs and data buffering. Additionally we have a 2MB SPI flash chip on board for configuration and possible html page and/or data storage as needed for future growth. Right now we keep all the html pages in program memory for all of our products. But if the program code grows we can move the html to the external flash and free up a lot of memory.

-Ed

[mrpackethead]

Good luck ed with this project.   

[j1sys]

Thanx Andrew.

To All -

I have sent an email containing either an invoice or a request for more information to all of the people I think should be in the queue. If you think you are in the queue and didn't receive an email from me in the last few hours please advise.

-Ed

[dmoore]

It's great to see you offer a commerical class controller at reasonable price level. 

[lithgowlights]

Thanks for the updates & glad to hear you are recovering ok. Invoice received and paid

[ɟɐsʇǝppʎ]

We expect to receive a prototype run of ECG-PIXAD8 PCB early next week and will build them up and test them and post photos as rapidly as possible. The ECG-PIXAD8 is a 3oz copper PCB signal and power adapter board that sits under an ECG-M32MX to provide fused power distribution connections to 8 pixel strings. The strings are split into two groups of 4 with a fuse on each output as well as the input of each group. The traces and parts are spec'd to handle 8 amps per string and 32 amps total per group. We will probably provide 7.5amp and 30amp fuses or possibly offer the more conservative group of 5amp and 20amp fuses. It uses standard automobile ATO fuses so field configuration and replacement should be easy.

We plan on initially supporting 6803, 2801, and 1804 pixels and will support either ArtNet II or E1.31 inputs. We will provide support for other pixel chips as demand warrants and time allows. The system will include our usual html configuration screens.

Ed, you comment on providing output to 8 strings, does this also mean that it can handle 8 Universes of lighting control.

[j1sys]

Eddy -

short answer: yes

long answer: (Your mileage may vary, and all other disclaimers. TANSTAAFL - Their ain't no such thing as a free lunch (Heinlein))

I will probably arbitrarily limit you to servicing 32 universes over E1.31. The average use will be about 4. When you come to using RGB pixel strings you have a lot of timing considerations along with your power distribution problems.

Note: ArtNet II is limited to 4 universes by design

(Run Tabor, Ed's going to do math)
(b = bit, B = byte)

Everyone knows the simple 170 pixels = 1 universe rule of thumb. The actual number is: 512/3 = 170.66

So assuming you've solved your power distribution problem and you want to hook four strings of 170 pixels to one group of connections let's look at it's refresh rate. (There are two groups of four on the board)

All refresh rates are based on the clock frequency (baudrate) used to send the data to the pixels. Due to noise and wire length people have found that keeping the baudrate below 1Mbps works well. Let's use 800Kbps as an example. The software will allow you to set your target baudrate on a per string basis to fine tune your layout. All calculations will round to whole numbers and the interrupt latency will be ignored but the numbers should be within 10% of real world.

800Kbps / 8 = 100KBps
100KBps / 510 = 196 universe per second
196Ups / 4 = 49 Ups round robin on 4 strings
1sec/49 = 20ms refresh rate

so at 800Kbps you can probably just barely keep up with a 25ms update rate coming from your sequencer. Not all pixels and protocols can run at 800Kbps so you will have to change something: baudrate, pixel count, number of strings, update rate to make it match.

So if you have solved your power distribution and are willing to use 50ms or 100ms update rate or your pixels are real close to the controller and you can crank the baudrate up you COULD double or quadruple the length of each pixel string and use 32 universes total for all 8 strings at 4 universes per string.

You also may want to use short strings (50 pixels) and crank up the update rate to 10ms. All settings are up to you and your dreams and desires. If you really need speed take some of your gold bars out of storage and ask me about ECG-M32MX/ECG-SPI8/ECG-SPIPIX8 three board solution with a dedicated SPI for each pixel string.

Each time you update your configuration the system will display the timing information for your settings so you can see the final result.

Always look to the left and see my motto: There are no rules and those are the rules.

-Ed

[j1sys]

I may have scared some people with the 170 pixel string calculation. here is the calc for 50, 100, 150. all of them are well within the 25ms update rate. Only the 170 is a little tight with some ethernet overhead.

For 50 pixels:
800Kbps / 8 = 100KBps
100KBps / 150 = 666 universe per second
666Ups / 4 = 166 Ups round robin on 4 strings
1sec/166 = 6ms refresh rate

For 100 pixels:
800Kbps / 8 = 100KBps
100KBps / 300 = 333 universe per second
333Ups / 4 = 83 Ups round robin on 4 strings
1sec/83 = 12ms refresh rate

For 150 pixels:
800Kbps / 8 = 100KBps
100KBps / 450 = 222 universe per second
222Ups / 4 = 55 Ups round robin on 4 strings
1sec/83 = 18ms refresh rate (probably will get done in 25ms update with overhead)

The more refreshes that can be done per update the better. this will keep latency for updates down. IMHO the 100 pixel count is probably the sweet spot. It can probably be done with double ended power injection and will refresh twice in each 25ms update to have a 12ms latency at worst.

The configuration screens will do this math for you based on the protocol, baudrate, and pixel count. It will even handle different protocols, baudrates, pixel counts across the strings.

-Ed

[mschell]

Thanx Andrew.

To All -

I have sent an email containing either an invoice or a request for more information to all of the people I think should be in the queue. If you think you are in the queue and didn't receive an email from me in the last few hours please advise.

-Ed

Ed, I assume you're talking about pending DR4 orders, right, as far as the email goes.

Definitely put me down as interested in the PIXAD8, and I can certainly see some interest in the PIXAD4.

The density of a 16 output pixel controller seems to be limited to either a megatree type fixture, or perhaps in a LED curtain arrangement, I'm thinking.

Glad to hear that you've had a good recovery from surgery, etc...

[smartalec]

i know now is not the time to place an order,
but i'll have a boxed version of the PIXAD8 when there ready
thanks alec.

[tpboyce]

I would like to express interest in the 16 string version potentially for a mega tree.  I may have additional interest as well.

[ɟɐsʇǝppʎ]

For guys thinking of building a pixel mega tree this controller looks like the goods, you can effectively use a 100 string (2 x 50 strings joined together) for a total of 300 channels for each output, which is a total of 2400 channels which achieves a refesh rate of 12ms.
The power can then be injected at both ends of the 100 count string at the bottom of the tree with a data input at the beginning of the 100 string. The tree can then be made in a zig zag with power and data supplied at the bottom and a 50 pixel count string ran to the top of the tree and then joined with another 50 pixel string for the run back down with power connected at the end of this string which will go through the fused outputs of the ECG-PIXAD8
The below diagram should make it a bit easier to understand



[ Guests cannot view attachments ]