New Sandevices Version 5 Firmware

[kidcole]

Jim has a lot of changes in progress. So programming these controllers will catch up to all the other brands very soon. Plus a new power saving feature that looks very promising. This new firmware will be for all of his controllers except the E6804. Here's his first post:

Version 5 firmware is in the works and is actually a bit ahead of schedule. Questions, comments and suggestions are welcome.

Initial release will be for the E682 and upgraded E680s and E681s, in other words, this is for the 16-output controllers.

The primary goal of this release is to overcome some significant addressing limitations of the current firmware. With existing firmware, on all controllers except the E6804, you specify one length, one start address, and one color order for each group of 4 outputs, and addresses for those 4 outputs are always assigned sequentially.

This is not convenient for users who have many different string lengths.

With version 5, you will be able to set any start address and any length on each of the 16 outputs. In addition, color order, pixel grouping, and zigzag can be set per-string (as well as per-string nulls and reverse which you can already do). The zigzag function has been improved so that you can specify 2 values: the number of pixels until the first zig, and the number for all subsequent zigs. This allows the zigzag feature to be used on a matrix or megatree even if the string length is not an even multiple of the matrix height.

Version 5 will support RGBW pixels in 2 modes. In RGBW mode the sequencing software is expected to send 4 channels of data to each pixel. In derived white mode (idea from Joe Hinkle), the controller only receives the R, G, and B channels, then it derives the final value for the RGBW channels to be sent to the pixel. This allows use of RGBW pixels with standard RGB sequencing software.

I expect (not 100% guaranteed yet) to be able to handle 16 universes of pixel data with version 5, with the potential to do even more , but possibly requiring a minor hardware upgrade. In addition to increasing the total number of pixels that can be controlled from 2,040 up to about 2,700, it also allows users to configure the controller as "1 output = 1 universe". The concept of universes is confusing to many users, and a lot of folks think that outputs and universes are the same thing, and have trouble dealing with 16 outputs but only 12 universes.

Version 5 will support 512 or 510 channels per universe. I'm not sure exactly how that will be done, whether global (entire board), per output group, or per universe.

By the way, the output type will still be assigned per groups of 4 outputs. That's a hardware limitation right now. I am freeing up at least one CPU so that opens up the possibility of a 5th output group in a future design.

Multiple DMX modes: DMX Pixel mode allows use of features such as color order, grouping, reverse, zigzag, and also length is specified in pixels, and standard DMX mode where the length is specified in number of channels. Renard output 'length' will also be specified in channels also, and there will be support for multiple baud rates for Renard outputs.

Multiple timing options for 1804/2811 class pixels.

Output addressing settable as universe-channel number or just as a relative channel number.

A controller name field on the web page, and a short controller name field that will appear in the browser tab, to simplify controller identification in multi-controller setups.

Not coded yet but on the wish list:

Support for SACN Priority, and SACN HTP (highest takes precedence).
Undo command to undo multiple levels of configuration changes.
A name field for each individual output.
Auto addressing of outputs.

[kidcole]

A few more things (copied from doityourselfchristmas.com):

GECE will probably not be in the first beta release of version 5. Support for the GEs and TLS3001s will require another software module and that's not part of the first release.

I am attaching a screen shot of one of the web pages on the version 5 user interface. Because of the additional configuration capabilities, it's necessary to go to 2 pages.

Here's a description of what's on this page. This page is for configuring the 16 individual outputs, 1 line per output.
Output number, eg 1-1, and output type. Output type is still done per groups of 4 outputs, and that selection is done on the other web page, display only here.

Length is the number of pixels for pixel outputs, or the number of channels for channel-based outputs. Renard is always a channel=based output. DMX can be done both ways, as a pixel-based output or as a channel-based output depending on whether or not you're running DMX pixels (or dumb RGB pixels via a dumb to DMX controller) or other non-pixel DMX devices.

The color order drop-down varies by output type. Standard RGB pixel outputs have the 5 possible RGB color orders. RGBW pixel types have 12 RGBW color orders to choose from. Non-pixel DMX and Renard have no color order drop-down.

Start Universe and Channel. Much like before although there is now the ability to specify either 510 or 512 channels for each universe. 512 channels makes more sense for RGBW pixels or non-pixel DMX outputs.

Ending Universe and channel are straightforward. but notice that universes 13-16 have been set as 512 channels per universe. For example, the 128 RGBW pixels on output 2-1 use all 512 channels of universe 13.

The Reverse check box indicates that the output is reversed, in other words data that normally goes to the first pixel of the string goes to the last pixel. For those not familiar with the feature, imagine a roofline with a pixel controller located in the center a 2 strings of pixels, one in each direction. By flagging the left-hand pixel string as reversed, the sequencing software can address the entire roofline sequentially from left to right.

Null pixels, same as before, except now you cannot add nulls to DMX or RENARD outputs.

Group size is used to reduce channel count by driving pixels in sequential groups instead of individually. This is in the existing firmware. One new feature is the "Chase" flag. When set it tells the controller that it should use groups as a chase. In other words, if your group size is 5 and the CHASE is set, the string will appear to be 5 pixels long. Pixel 1 data will also appear on pixels 6, 11, 16 etc. So if you run a "chase" pattern across the first 5 pixels it will chase across the entire string.

The zigzag feature has been improved. The zigzag feature is designed for use with a pixel matrix or megatree where the string lengths are larger than the matrix height, so strings are snaked up and down the matrix. The previous zigzag feature would ony work if the matrix height divided exactly into the string length, say a 10-high matrix with strings of 50 pixels. The new zigzag adds a "first zig" entry so that you can use zigzag even if matrix height doesn't divide evenly. Say the matrix is 12 high. The first 50-pixel string will have 2 pixels on the 5th column of the matrix. On the 2nd string you would say first zig=10 and other zigs=12.

Finally, the DIMMER entry allows several selectable intensity levels for each output.

Tomorrow I will poet the other web page and run through it. One new feature is a "power limiter" that you can use if your power supply isn't capable of running all pixels at 100% white. the power limiter keeps track of the cumulative intensity levels for all pixel channels, and from that it calculates total power consumption. If this exceeds the settable threshold the board will automatically dim the entire display to keep total power below the threshold.

[Tripplett]

Looks great and dimming my whole pixel tree would be a nice addition, that sucker is bright. That would be great to do at the controller level so I can just drive the sequencing software at 100%

[kidcole]

Looks great and dimming my whole pixel tree would be a nice addition, that sucker is bright. That would be great to do at the controller level so I can just drive the sequencing software at 100%

Yes, a very smart feature. The others will probably follow with that feature over time.

[ggusta]

not sure where this news belongs so this seemed like the most relevant discussion.

e1.31 and e1.33 public review (btw, for some fat fingered reason I am gguZta at dychristmas.org)

[FrankP]

Has anyone installed version 5.0 yet?

[kidcole]

Has anyone installed version 5.0 yet?

I have not updated mine yet. But I do plan to break free from my garage-building captivity around the end of June. By mid July I should have my entire display redefined in DMX channels. I'll be updating firmware around that same time, when I set up my 4 Sandevices controllers. I'll be contacting some of you guys to get some help with converting LOR to xLights and setting up several FPPs.

[TGabriel]

I have update my 2 E682 to 5.003 and waiting on the versionfor the E681s

Thomas

[kidcole]

I have update my 2 E682 to 5.003 and waiting on the version for the E681s

Thomas

Please keep us posted if you see the 681 update. I am running two 681s that have the upgraded network card. Mine should be candidates for the new update too ...

[Papa-Ho-Ho]

Where can i find the firmware download?

Lowell

[kidcole]

Where can i find the firmware download?

Lowell

Jim is using DIYC as his beta testers for the new firmware, so it's not on his site yet. Go to this link for the info thread and you can download the file from Post #1

SanDevices version 5 firmware information thread - Page 3

Looks like Post #25 has a lot of detail. Also, the E681 upgrade to V5 is going to require a crystal change. I may not need this new firmware on my E681s, but if Jim offers the crystal in a kit with instructions, then I'll upgrade my two. I've always kept them up to date with the E682s, so they don't fall behind ..

[Papa-Ho-Ho]

Thanks Denny - I have been kind of out of the loop for almost 4 years, I did get my show back up last year at the new location, but it was far from complete. The new neighborhood was in awe. Thats fun, because "they ain't seen nuttin yet!"

[Tripplett]

Well let me tell you I just spent the last 3 hours farting around with this and I'm not happy. I have an E682 and really really really wanted the dimming on the controller. I have a few purchased sequences I don't want to mess with in XLights to dim them and my tree is close to the street (read as it blinds the neighbor across the street). I was running firmware 4.033. i read a few posts of people having problems with this V5 firmware.

I started by updating it to the latest V4 of 4.081. That didn't give me the dimming unfortunately. I'm actually not sure what it did change - it all looks the same. So I went ahead and upgraded to V5 since people said they were able to go back from V5 back to V4. It wiped all my settings out - it was like a brand new controller. The IP address went back to 192.168.1.206 from 10.0.0.205 so I couldn't even find it on the network. I moved it to my home network (from show network) since that was luckily 192.168.1.x. I found the controller and updated it back to a 10.0.0.205 address. Then I had to painstakingly update all the outputs back to WS281x, starting universe, length, GRB order, non reversed, etc etc. Finally I had it all done. I ran a few of the test sequences on the controller against my ribbon Pixel tree laid out in the living room floor.

The result was a mess. Solid white test pattern... Strand 1 would only turn on the first 3 pixels then the 4th was yellow with the remainder out. Strand 2 wasn't much better with every few pixels being yellow. The remainder of the strands seemed ok. I was now thinking I had damaged some of the ribbons when I put them away last year. That was until I tried all red. Strand 1 and 2 were all messed up still. Strands 7 and 8 were green, not red. When I changed to an all green 7 and 8 were red. What? I had read that V5 test patterns were messed up so I fired up XLights, started a new sequence, and tried the same test. The result was the same. i now was worried again that my strips were messed up and I'm very careful to put them away nice and gentle. Oh and I went back into the E682 setup page and verified the RGB order, reversing, everything. I even tried to change them, save, and change them back. No luck.

Last ditch effort - roll back to V4. All the problems went away. Good grief I'll never get those 3 hours back. I'm supposed to be pulling stuff out of the attic today. Anyway folks, I'd stay away from V5 especially this close to season. Sadly dimming will have to wait. I'll check XLights updates and maybe I can do it at the model level there.

[EricR]

You can do it at the model level in xlights. I will have to look at Xlights to tell you exactly but you can change model attributes on the layout screen so that your tree model is set to what ever brightness you want.

Called "Dimming curves"




Sent from my iPhone using Tapatalk

[mpurser]

if your ruing a falcon brightness can be controlled from the controller