Odd behavior on F4V3 with long strings after upgrading to FPP 8.0

[Reys Dad]

I've got a bit of a strange issue that I'm thinking is likely a bug. Here's whats going on
I'm running a F4V3 on v2.59 with an expansion card
I have a string of icicles that is 252 nodes running off an older differential receiver. It worked fine last year.

If I run a display test from the F4V3 all 252 nodes light up and cycle through the rgb test perfectly.
However, if I run a color wash from the master FPP to the slave FPP connected to the F4V3 I get the following results:

252 pixels - They start white and freeze. No color change, no shutoff at the end of the color wash sequence
180 pixels - Same as 252 pixels although once they flickered and did the color wash and then stopped
120 pixels - Works fine for several tests.

I double checked my xlights and controller setup for both the FPP and the F4v3 and the start and stop channels match.

This leads me to believe that somehow there is a pixel limit per string for data being passed through the FPP in version 8 on the F4v3 that wasn't in the 7.x versions. I have F16v3 and it doesn't have this issue with a 440 node string, so seems specific to the F4v3. Can someone check if this is a bug. If you need logs or anything let me know before I revert back to 7.5
Thanks

[JonD]

The Falcon and FPP test functions are a very handy tool, but are not to be trusted.  Have you tried running a sequence to confirm the issue even exists?

How many lights does the F4v3 show are assigned to the port?  252 is a lot of standard lights even at 30% power.  Are you power injecting? Flickering and color issues could mean they do not have enough power towards the end.  What is the voltage measurement at the end of the light string?

I have an F3v4 with expansion card running a 299 light string no issues, but I am power injecting midway down the line. 

[Reys Dad]

The Falcon and FPP test functions are a very handy tool, but are not to be trusted.  Have you tried running a sequence to confirm the issue even exists?

Yes, I'm running a simple colorwash effect on a sequence.

How many lights does the F4v3 show are assigned to the port?  252 is a lot of standard lights even at 30% power.  Are you power injecting? Flickering and color issues could mean they do not have enough power towards the end.  What is the voltage measurement at the end of the light string?

It shows 252 lights and I am power injecting. I'm almost certain it's not a hardware issue as it tests fine using the F4V3 string test. All 252 light up and change color no issue. It will run a sequence just fine if I reduce the pixel count to 120. It will not run a sequence if set the pixel count to 180 or higher. So long story short, all 252 work if data signal comes directly from f4v3. Will only run 120 when the data signal comes from the master via the connected fpp. I don't know the exact count where it breaks but somewhere between 120 and 180 when running a sequence.

[Poporacer]

if I run a color wash from the master FPP to the slave FPP connected to the F4V3 I get the following results:

252 pixels - They start white and freeze. No color change, no shutoff at the end of the color wash sequence

I would bet that you are using 12V Pixels? There is a somewhat common issue when using 12V pixels when you don't have a sufficient ground. The voltage on the ground gets too high to be able to register the proper ws2811 signal and this typically happens after an all-white effect. Did you try testing just RGB from FPP instead? 
There are a few solutions to this, one or all might work.
1. Increase your ground wire...Larger wire or an additional wire
2. Reduce the brightness
3. Reduce the number of pixels
4. Use an F-Amp
5. Don't use all white in your effects

120 pixels - Works fine for several tests.

So item #4 worked, so I think what I mentioned is likely your problem

This leads me to believe that somehow there is a pixel limit per string for data being passed through the FPP in version 8 on the F4v3 that wasn't in the 7.x versions.

No, that is not true, I have V8 running 840 pixels on one of my ports without a problem.

I have F16v3 and it doesn't have this issue with a 440 node string, so seems specific to the F4v3.

The controller might have a slightly different output or the pixels themselves could be drawing a slightly larger amount of current. I use the exact same controller connected to the exact same pixels year after year and on 2 occasions I had pixels that would not light correctly like they did the year before and then needed power injection. I chalked it up to the pixels just getting a bit older and using more current than the year before.

Can someone check if this is a bug.

I don't think it is a bug, it sounds more like a hardware issue.

Yes, I'm running a simple colorwash effect on a sequence.

But that color wash probably includes white.

It will run a sequence just fine if I reduce the pixel count to 120.

Can you create a sequence using the bars effect and just use red green and blue and see how that plays?
What brightness are you using?

So long story short, all 252 work if data signal comes directly from f4v3. Will only run 120 when the data signal comes from the master via the connected fpp

What happens if you test from the FPP connected directly to the F4?
Are you sure the outputs on both FPP devices are correct?

[JonD]

It is actually pretty common to have to fix a light string or two every year.  You roll them up, throw them in a box, and something happens.  I have learned never to say never, but I would guess it is a light string issue. 

[Poporacer]

You roll them up, throw them in a box, and something happens.

And they are not designed for that, I am actually surprised that there are not more failures!

[Reys Dad]

4. Use an F-Amp

This seems to be the issue. I reverted back to 7.5 and still had freezing issues. Adding an F-Amp straightened it out. The run didn't seem all that long to the first pixel maybe, 10 feet or so but I guess it was enough and as God as my witness I thought turkeys could fly. Anyway, I'll upgrade back up to 8.0 in a day or two and post back but it should work now.
Thanks for the help

[JonD]

If your cord is only 10ft long, a f-amp might only be a temporary fix.  Mice will chew and wind sway will rub holes.  Might thoroughly inspect the cable for any exposed wires.  Hope that resolves it for you, but at least you will know where to look if you have any more issues.

 

[Reys Dad]

Just to confirm. 8.0 is fine now that I added the F-Amp