PDA

View Full Version : FPP to 2 E682 controllers = massive network/sequence errors



Raabaman
08-16-2016, 02:57 PM
I'm trying to connect my 2 E682 controllers to my RPi running FPP. When I run the controllers individually from the FPP all is well. When I connect the second controller, through a unmanaged switch it all kind of falls apart.

I'm not saying that the controllers don't work when both connected up. They do. They just have pauses that are super noticeable. When I put the FPP into test mode and run a simple chase it goes correctly (on both controllers) for 1 second, then pauses for a bit, then runs again for a second or two, then pauses..... It's not the chase. If I run the same chase on one controller it's seamless. It just runs continuously as expected.

I can see that data is getting to the controllers by seeing the "Packet Count" go up on the controllers web-settings page. It also shows that there are numerous errors in the data.

The controllers have been updated to the very latest firmware 5.018 (**correction** I just noticed there is now a version 5.018-01) Both are set to the same configurations. I am only enabling 4 strings of 100 pixels. Each string is 1 universe - 510 channels.

I have also seen the same type of problem when I try to connect my laptop to the switch. It will not make a sustained connection. It connects then disconnects every second or so.

This has got to be a network setting problem. I just don't think I understand it well enough. I am also assuming that the problem is in the way I've configured the FPP network setup.

The RPi/FPP is connected via wifi to my home network(192.168.1.***). I then set the ethernet port on the RPi/FPP to be a separate subnet(192.168.100.***).

What have I done wrong??

It's confusing so I'm including a few pictures to help get the point across.

Cheers!

sporadic
08-16-2016, 03:06 PM
Turn your sequencer off?

Raabaman
08-16-2016, 03:09 PM
Turn your sequencer off?

Not sure if you mean the controller's built in tests. Both of the controllers have the "Test Pattern" set to "Off".

ThinkAgainAgain
08-16-2016, 07:33 PM
Yes, you can't have your sequencing software running while you're running your falcon

Ruppro
08-16-2016, 09:56 PM
The Network configuration you have looks sound and should work fine. Sounds like the switch is having issues, both the controllers and your laptop are having problems when connected to the switch, replace it and see if that makes a difference (you can always return it if not)


I'm trying to connect my 2 E682 controllers to my RPi running FPP. When I run the controllers individually from the FPP all is well. When I connect the second controller, through a unmanaged switch it all kind of falls apart.

I have also seen the same type of problem when I try to connect my laptop to the switch. It will not make a sustained connection. It connects then disconnects every second or so.

jchuchla
08-16-2016, 10:29 PM
Two things come to mind. You generally don't want your sACN source being a wifi node. Wifi is really meant for the wireless clients to be data consumers and the occasional request going back upstream. In general the devices pushing the data out should be on the wired network segment. This is even more important if the listeners are also on the wireless segment. The wireless client is not going to guaranteed enough timeslices to get all of its data out. If the traffic comes from the wired segment onto the wifi for listeners to consume, it will have a more even share of the airtime to get the data onto the airwaves.
Also something to keep in mind is that multicast works better over wireless than unicast. There's another recent thread where I go into detail on this topic (it was named something about choosing an access point)
However, the Ethernet chips used in the 682 controllers work better on unicast than multicast (unless there's been a recent 682 firmware change that I'm not aware of)


--Jon Chuchla--

Sent from my iPhone using Tapatalk

Ruppro
08-16-2016, 10:39 PM
Jon

Look at his 4th Picture, He has the FPP and Controllers connected on the Wired network. He is using WiFi only to manage FPP from his home network.

33298

algerdes
08-16-2016, 10:43 PM
That TP Link 5-Port 10/100 Desktop Switch uses their "Green Ethernet Technology" which, according to them, " saves the power up to 68%".

I had all kinds of trouble with a couple "green switches". Until I put in a reliable, always on, I had no idea that the switch was making a mess of my efforts.

If you have a standard, non-green, to try, I would.

jchuchla
08-16-2016, 10:44 PM
Jon

Look at his 4th Picture, He has the FPP and Controllers connected on the Wired network. He is using WiFi only to manage FPP from his home network.

33298

Good call. I must have read the OP wrong.
In that case, I'd agree with the others earlier, try a different switch.


--Jon Chuchla--

Sent from my iPhone using Tapatalk

Raabaman
08-17-2016, 02:55 AM
Thanks for all the feedback everybody. I'm just my day started so I'll have a few hours after work to find a new "always on" switch and test it out. I'll let you know a bit later how it goes.

...I knew this was going to be confusing. A picture speaks a thousand words and all that! ;-)

Cheers

Raabaman
08-17-2016, 02:31 PM
The results are in ....

It was the network switch! I replaced it with a new Netgear FS105 and everything just came alive. I can't tell you how relieved I am at this point.

Now, for the **full disclosure**
It probably wasn't the TP-Link switch's fault. It was entirely mine. I accidentally hooked the switch up to my 12v supply and it's a 5v switch. It may be salvageable but since I already sacrificed it's cable to the enclosure it was in and I don't have a spare 5v cable with the correct end on it, it could take a while before I know.

Doesn't matter now. It works and I'm on to the next thing, which in this case happens to be a mega-tree. I'm sure that will generate some more "wtf have I done wrong this time" questions!

Thanks for the help everybody.

Ruppro
08-17-2016, 05:33 PM
Glad to know you got it working right.

jchuchla
08-17-2016, 06:02 PM
For reference, the Fs105 switches come with a power supply that's something like 7 or 9v. But they do work fine on 12v power.they have an internal regulator that handles the 12v fine. I use these units(and their newer brother the gs105) all the time and regularly share a 12v DC supply to power them. They get very warm on 12v, that extra power needs to go somewhere. But I've yet to fry one or see any odd behavior.


--Jon Chuchla--

Sent from my iPhone using Tapatalk

algerdes
08-17-2016, 06:44 PM
I second what Jon said. The power supply can vary, for sure.

Under "normal use", these "green" switches can be fine in a non-critical environment (read "home") or a small office.
One thing I have noticed, but have not verified, is that some of the algorithms used to test if a port is being used require occasional bi-directional data to be present within a specific amount of time. A lot of what we use in DIYC is push only. Very little return data. This caused my green switches to close out the port, even if only for a few seconds, and therefore cutting the signal.

Troubleshooting can be a bear!!!

Glad to hear you found it.