DMX frames drop when high i/o

Ask a question about the usage of QLC+ with the Linux operating system
Post Reply
juliaotran
Posts: 2
Joined: Mon Apr 25, 2022 3:21 pm
Real Name: Julia Otranto Aulicino

Hi guys.

I'm using QLC+ 4.12.3 on ubuntu 21.10 with a FT232RL usb interface (https://www.phlighting.com/lci-512-usb).

However when I start to do some activity that requires high I/O, the lights gets black and then turns back on, by a few milisecs.
The interface also has a orange led who flashes on every DMX frame, and when this problem occurs this led also gets fully black by a few ms (it is flashing so fast most of time).

I know that it is a high i/o problem because when I run `stress -i 4` the problem starts to occur. Stopping stress stops the drops, except a few ones that i suppose to happen due to other opened programs.

So, I have lots of questions:
Is this a QLC bug?
Is this a driver bug (maybe could be something related to my usb controller)?
How could I know what is?
User avatar
GGGss
Posts: 2732
Joined: Mon Sep 12, 2016 7:15 pm
Location: Belgium
Real Name: Fredje Gallon

juliaotran wrote: Mon Apr 25, 2022 9:37 pm stress -i 4
is not QLC+ related - I presume an utility on Ubuntu?

Now, you have to understand that with your interface, the processor of your computer has to take care of the DMX timing. Depending on the number of DMX channels used, this can indeed result in packet losses.
Or you can try to lower the DMX frequency, or you can try to limit the number of channels to be sent to the universe.
In the case of a DMX interface with an inbuilt processor, the DMX timing is done there and the only communication between the computer and the interface is a list of values to be sent. As you might understand, a total another workload.
All electric machines work on smoke... when the smoke escapes... they don't work anymore
juliaotran
Posts: 2
Joined: Mon Apr 25, 2022 3:21 pm
Real Name: Julia Otranto Aulicino

Well, I used wireshark and usbmon to see if the packets are being sent with delay, i could notice that most of they are sent with 29ms interval. however a few ones had take 38ms, maybe this is the really the cause of the problem, but what it is strange is that stressing CPU with
```
stress -c 8
```
doesn't cause issues, however stressing I/O with
```
stress -i 2
```
a few secs after starts the problem.

Also, I was using 25 fps, after increasing to 35 the flashing amount reduced.

I will try to use some artnet to dmx usb driver and see if gets better
Post Reply