Aurora Support Forum: sticking/stuck channels

Author

Message

Topic Search

Topic Options

Pony_God

Senior Member

Joined: 01 Sep 2008
Location: Naples, FL
Online Status: Offline
Posts: 551

Quote

Topic: sticking/stuck channels
Posted: 03 Dec 2009 at 7:22pm

I'll start with the point, then give the background.

The point: I'm thinking that Borialis is sending out erroneous commands causing channels to get stuck.

So here's my thoughts.

1. Durring the 2008 season, we did have sticking channels that cleared themselves up the next time that the channel was used. These stuck channels were fairly minimal, far between, quite random.

Aurora .41, F/W 1.15beta, single network, 6 ACx16s, 1 Firefly.

2. This season we have quite a few stuck channels that still clear themselves out on next use. These channels are much more predictable, I know what songs, and what chases that they will get stuck in almost every time.

Aurora .69, F/W 1.15beta, F/W 1.16, dual network, 12 ACx16s, Firefly not up yet.

3. We have had the last two controllers on network 1 loose thier IDs (even though we didn't loose any IDs last year) and yet they are getting command at times. The times that they are getting the commands are the exact same times that other channels get stuck on other controllers on other networks.

4. The channel that's getting activated on the two controllers are the same channel ID (or possibly one off)

5. Now, if those controllers do not have the correct ID, how are they getting a command? And why are they getting a command at the same time that other controlles are getting stuck channels?

5. Is there a command going out to _all_?

6. Since half of the firmware is last year's I would expect those controllers to have the same frequncy of stuck channels, but it's not. One of the controllers that has a lost ID has 1.15beta (from last year) and the other has 1.16 (new for this year)

It just seems that thier are only two changes, a network addition and software update.

What's everyone else's thoughts?

Fine. You're so smart you rig up the lights.
D-Light users Unite!

deweycooter View Drop Down

Members Profile

Send Private Message

Find Members Posts

Visit Members Homepage

Add to Buddy List

Development

Joined: 14 Oct 2007
Location: League City, TX
Online Status: Offline
Posts: 674

Quote

Posted: 03 Dec 2009 at 8:15pm

For you - what's keeping you on .69? Why not go back to .41?

For the issue - I assume the lost ID are confirmed by using the Discover Units in the DL Utility, or by commanding a chase to a specify ID in the DL Utility and not seeing anything? Do they reset to a value or just have no ID? What would you estimate the total cat5 length? I assume the sequences have changed since last year, so that clouds the issue a bit when comparing year to year.

An interesting test would be to assign all controllers the same ID, then compare behavior differences down the line. Stuck channels....? That's a bit of work for one test, but could let you know if there are network length/quality issues.

Software stuck channels (as far as I understand) are channels that receive an on command and don't receive/obey the off command - for whatever reason. (I say software because I have had bad triacs that lead to stuck channels.) The question is whether or not the command is getting out at all, or is just not picked up by the controller. Another task that would be a LOT of work is to get a serial sniffer out and watch the commands - huge pile of data to dig thru...

Edited by deweycooter - 03 Dec 2009 at 8:18pm

Aurora Lights Wiki

Pony_God

Senior Member

Joined: 01 Sep 2008
Location: Naples, FL
Online Status: Offline
Posts: 551

Quote

Posted: 03 Dec 2009 at 9:01pm

I thought that there was a DB change between the two. Can .41 load a .69 sequence? If Borialis can run .69 sequences I can downgrade the server.

Disconver can not find the two controllers.

With the hardware utility, and ALL checked, ON/chase/twinkle/etc all get to the controller.

If they have an ID, I could not see it. I went from ID 1 though 20 and tested each with ON, and did not get them to respond. (One of the IDs is 12, I don't remember the other off-hand.)

cat5 100'? more? I can measure tomarow. It's less than last year because of the seperate networks. Net 1 is not terminated, net 2 is.

Sequences did change, I copied last years, and then imported this year's channels. I can send the one that I know has the error.

Durring any hardware tests, everything seems fine, but that's totally different software, speed, etc.

And that's how I understand a stuck software channel, partially. I see three options.

1. Software - sending bad data down the wire

2. Firmware - not obeying the commands

3. Hardware - triacs, etc.

As I see it.

3. The hardware seems fine, it can act normall and to firmware tests, and HWU tests.

2. Possibly, but some have the same firmware as last year and did not have as many stuck channels. If it was firmware, I would expect simmilar results.

That leavs option 1 again.

Fine. You're so smart you rig up the lights.
D-Light users Unite!

deweycooter View Drop Down

Members Profile

Send Private Message

Find Members Posts

Visit Members Homepage

Add to Buddy List

Development

Joined: 14 Oct 2007
Location: League City, TX
Online Status: Offline
Posts: 674

Quote

Posted: 03 Dec 2009 at 9:20pm

I don't think there was a change between 41 and 69 - I just recently upgraded (last week). The hard part is the uncertainty that has been seen with up/downgrading. I'd try on a spare machine first, though it may be as easy as doing the same process as upgrading: uninstall current version via control panel, reboot, install preferred version, reboot.

Aurora Lights Wiki

Pony_God

Senior Member

Joined: 01 Sep 2008
Location: Naples, FL
Online Status: Offline
Posts: 551

Quote

Posted: 03 Dec 2009 at 11:13pm

Okay, so I did do some more testing since I was up and it was bugging me.

The net 1 cat5 is 150 foot. net 2 is at least 225 foot.

Test was done with HWU with ALL and fastest chase.

I let it go and just watched the arch and the tree (last two controllers on the line, and the two that lost the IDs.)

Ever 8 seconds or so a channel would stick. The same channel on both of the last two controllers.

I moved the cable even further away from power, I'll try to cut a new cable and see if everything clears up.

I did happen to try another usb-RS485 adapter and it acted the same.

That doesn't explain the first two controllers of net 2 sticking though. (Yes, I did run the HWU on net 2 and nothing stick with that.)

Fine. You're so smart you rig up the lights.
D-Light users Unite!

JohnnyL

Pre-Order User

Joined: 17 Dec 2007
Location: Merced Ca
Online Status: Offline
Posts: 67

Quote

Posted: 03 Dec 2009 at 11:58pm

what happens when you run a single network? are you using the v3 adapter or v2? are you certain your cat-5's are good?

"In God We Trust" all others pay cash

Pony_God

Senior Member

Joined: 01 Sep 2008
Location: Naples, FL
Online Status: Offline
Posts: 551

Quote

Posted: 04 Dec 2009 at 12:01am

HWU ran on a sungle network.

Have not ran the show on a single network.

It's not damaged, and looks good its entire length, but I'll try and cut a new calble tomarow to test that.

Fine. You're so smart you rig up the lights.
D-Light users Unite!

JohnnyL

Pre-Order User

Joined: 17 Dec 2007
Location: Merced Ca
Online Status: Offline
Posts: 67

Quote

Posted: 04 Dec 2009 at 12:23am

i had sticky channels last year too. 1.16 cured it. 1.17 brought the stickies back. so i went back to 1.16. has anyone else besides yourself had issues with the new adapter?

"In God We Trust" all others pay cash

Pony_God

Senior Member

Joined: 01 Sep 2008
Location: Naples, FL
Online Status: Offline
Posts: 551

Quote

Posted: 04 Dec 2009 at 8:08am

.16 cured it? I thought that was only the LED fix.

I know that other people have sticking channels, but not sure about the adapter version.

I'll have more time over the weekend, so I'll try to determine better what firmware is sticking, make the new cable, re-set the IDs, try downgrading to .41, and upgrading the 1.15beta arch that's sticking.

See if any of those changes make a difference.

Fine. You're so smart you rig up the lights.
D-Light users Unite!

JohnnyL

Pre-Order User

Joined: 17 Dec 2007
Location: Merced Ca
Online Status: Offline
Posts: 67

Quote

Posted: 04 Dec 2009 at 4:50pm

yes it cured it for me. as soon as 1.17 was installed i noticed sticking channels on 3 controllers and the ramps looked terrible. the difference between 1.16 and 1.17 was night and day. i had one controller lose its ID on 1.16 i reassigned it and swapped out the Cat-5's just in case and have been trouble free (fingers crossed) ever since.

"In God We Trust" all others pay cash

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version
Pony_God Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 01 Sep 2008 Location: Naples, FL Online Status: Offline Posts: 551	Quote Reply Topic: sticking/stuck channels Posted: 03 Dec 2009 at 7:22pm
	I'll start with the point, then give the background. The point: I'm thinking that Borialis is sending out erroneous commands causing channels to get stuck. So here's my thoughts. 1. Durring the 2008 season, we did have sticking channels that cleared themselves up the next time that the channel was used. These stuck channels were fairly minimal, far between, quite random. Aurora .41, F/W 1.15beta, single network, 6 ACx16s, 1 Firefly. 2. This season we have quite a few stuck channels that still clear themselves out on next use. These channels are much more predictable, I know what songs, and what chases that they will get stuck in almost every time. Aurora .69, F/W 1.15beta, F/W 1.16, dual network, 12 ACx16s, Firefly not up yet. 3. We have had the last two controllers on network 1 loose thier IDs (even though we didn't loose any IDs last year) and yet they are getting command at times. The times that they are getting the commands are the exact same times that other channels get stuck on other controllers on other networks. 4. The channel that's getting activated on the two controllers are the same channel ID (or possibly one off) 5. Now, if those controllers do not have the correct ID, how are they getting a command? And why are they getting a command at the same time that other controlles are getting stuck channels? 5. Is there a command going out to _all_? 6. Since half of the firmware is last year's I would expect those controllers to have the same frequncy of stuck channels, but it's not. One of the controllers that has a lost ID has 1.15beta (from last year) and the other has 1.16 (new for this year) It just seems that thier are only two changes, a network addition and software update. What's everyone else's thoughts?
	Fine. You're so smart you rig up the lights. D-Light users Unite!

deweycooter Members Profile Send Private Message Find Members Posts Visit Members Homepage Add to Buddy List Development Joined: 14 Oct 2007 Location: League City, TX Online Status: Offline Posts: 674	Quote Reply Posted: 03 Dec 2009 at 8:15pm
	For you - what's keeping you on .69? Why not go back to .41? For the issue - I assume the lost ID are confirmed by using the Discover Units in the DL Utility, or by commanding a chase to a specify ID in the DL Utility and not seeing anything? Do they reset to a value or just have no ID? What would you estimate the total cat5 length? I assume the sequences have changed since last year, so that clouds the issue a bit when comparing year to year. An interesting test would be to assign all controllers the same ID, then compare behavior differences down the line. Stuck channels....? That's a bit of work for one test, but could let you know if there are network length/quality issues. Software stuck channels (as far as I understand) are channels that receive an on command and don't receive/obey the off command - for whatever reason. (I say software because I have had bad triacs that lead to stuck channels.) The question is whether or not the command is getting out at all, or is just not picked up by the controller. Another task that would be a LOT of work is to get a serial sniffer out and watch the commands - huge pile of data to dig thru... Edited by deweycooter - 03 Dec 2009 at 8:18pm
	Aurora Lights Wiki

Pony_God Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 01 Sep 2008 Location: Naples, FL Online Status: Offline Posts: 551	Quote Reply Posted: 03 Dec 2009 at 9:01pm
	I thought that there was a DB change between the two. Can .41 load a .69 sequence? If Borialis can run .69 sequences I can downgrade the server. Disconver can not find the two controllers. With the hardware utility, and ALL checked, ON/chase/twinkle/etc all get to the controller. If they have an ID, I could not see it. I went from ID 1 though 20 and tested each with ON, and did not get them to respond. (One of the IDs is 12, I don't remember the other off-hand.) cat5 100'? more? I can measure tomarow. It's less than last year because of the seperate networks. Net 1 is not terminated, net 2 is. Sequences did change, I copied last years, and then imported this year's channels. I can send the one that I know has the error. Durring any hardware tests, everything seems fine, but that's totally different software, speed, etc. And that's how I understand a stuck software channel, partially. I see three options. 1. Software - sending bad data down the wire 2. Firmware - not obeying the commands 3. Hardware - triacs, etc. As I see it. 3. The hardware seems fine, it can act normall and to firmware tests, and HWU tests. 2. Possibly, but some have the same firmware as last year and did not have as many stuck channels. If it was firmware, I would expect simmilar results. That leavs option 1 again.
	Fine. You're so smart you rig up the lights. D-Light users Unite!

deweycooter Members Profile Send Private Message Find Members Posts Visit Members Homepage Add to Buddy List Development Joined: 14 Oct 2007 Location: League City, TX Online Status: Offline Posts: 674	Quote Reply Posted: 03 Dec 2009 at 9:20pm
	I don't think there was a change between 41 and 69 - I just recently upgraded (last week). The hard part is the uncertainty that has been seen with up/downgrading. I'd try on a spare machine first, though it may be as easy as doing the same process as upgrading: uninstall current version via control panel, reboot, install preferred version, reboot.
	Aurora Lights Wiki

Pony_God Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 01 Sep 2008 Location: Naples, FL Online Status: Offline Posts: 551	Quote Reply Posted: 03 Dec 2009 at 11:13pm
	Okay, so I did do some more testing since I was up and it was bugging me. The net 1 cat5 is 150 foot. net 2 is at least 225 foot. Test was done with HWU with ALL and fastest chase. I let it go and just watched the arch and the tree (last two controllers on the line, and the two that lost the IDs.) Ever 8 seconds or so a channel would stick. The same channel on both of the last two controllers. I moved the cable even further away from power, I'll try to cut a new cable and see if everything clears up. I did happen to try another usb-RS485 adapter and it acted the same. That doesn't explain the first two controllers of net 2 sticking though. (Yes, I did run the HWU on net 2 and nothing stick with that.)
	Fine. You're so smart you rig up the lights. D-Light users Unite!

JohnnyL Members Profile Send Private Message Find Members Posts Add to Buddy List Pre-Order User Joined: 17 Dec 2007 Location: Merced Ca Online Status: Offline Posts: 67	Quote Reply Posted: 03 Dec 2009 at 11:58pm
	what happens when you run a single network? are you using the v3 adapter or v2? are you certain your cat-5's are good?
	"In God We Trust" all others pay cash

Pony_God Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 01 Sep 2008 Location: Naples, FL Online Status: Offline Posts: 551	Quote Reply Posted: 04 Dec 2009 at 12:01am
	HWU ran on a sungle network. Have not ran the show on a single network. It's not damaged, and looks good its entire length, but I'll try and cut a new calble tomarow to test that.
	Fine. You're so smart you rig up the lights. D-Light users Unite!

JohnnyL Members Profile Send Private Message Find Members Posts Add to Buddy List Pre-Order User Joined: 17 Dec 2007 Location: Merced Ca Online Status: Offline Posts: 67	Quote Reply Posted: 04 Dec 2009 at 12:23am
	i had sticky channels last year too. 1.16 cured it. 1.17 brought the stickies back. so i went back to 1.16. has anyone else besides yourself had issues with the new adapter?
	"In God We Trust" all others pay cash

Pony_God Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 01 Sep 2008 Location: Naples, FL Online Status: Offline Posts: 551	Quote Reply Posted: 04 Dec 2009 at 8:08am
	.16 cured it? I thought that was only the LED fix. I know that other people have sticking channels, but not sure about the adapter version. I'll have more time over the weekend, so I'll try to determine better what firmware is sticking, make the new cable, re-set the IDs, try downgrading to .41, and upgrading the 1.15beta arch that's sticking. See if any of those changes make a difference.
	Fine. You're so smart you rig up the lights. D-Light users Unite!

JohnnyL Members Profile Send Private Message Find Members Posts Add to Buddy List Pre-Order User Joined: 17 Dec 2007 Location: Merced Ca Online Status: Offline Posts: 67	Quote Reply Posted: 04 Dec 2009 at 4:50pm
	yes it cured it for me. as soon as 1.17 was installed i noticed sticking channels on 3 controllers and the ramps looked terrible. the difference between 1.16 and 1.17 was night and day. i had one controller lose its ID on 1.16 i reassigned it and swapped out the Cat-5's just in case and have been trouble free (fingers crossed) ever since.
	"In God We Trust" all others pay cash