USB-problems

Ever since I first connected a computer with Rig-Manager to my profiler there as been problems. Random lockups and transfer/synchronisation that never finishes has been the order of the day. And if I happen to forget that the computer is attached and it finally goes into hibernation or other form of powersave-mode I can be certain that the KPA will lock up sooner rather than later.

It shouldn't be like this. Any computing-device with a USB-port should be able to take any abuse thrown at the USB-port, except for deliberate misuse of implemented communications protocols. You can not for example bar an application that pretend to be rig-manager and deliberately alters or deletes content on the KPA. Incomplete, incorrect or unknown activity on USB OTOH could possibly trigger some form of error-message, but should not be allowed to interfere with normal operations. Any other behaviour is a bug that needs addressing.

I just hope that the problem is in Kemper's code and can be fixed. However, since the bug(s) affecting USB have survived for so long I suspect that the root cause may be found in whatever platform the KPA is building upon, possibly never to be fixed.

Quote from michaelmellner

However, please provide more info about what you're working with

There are scattered reports on the forums over the time RM has been out about people having these kinds of problems, myself included, so it appears not to be isolated to any particular platform. The USB-A-port saw no use, except possibly in the LAB. prior to the release of the RM.

My 30 years in the IT-industry has taught me a thing or two and I've tried changing every component possible in the chain (computer, cables, software, profiler) and it has so far not made a difference. I have used 5 different KPAs over the period RM has been available. 3 rack-units and 2 heads., as well as windows and macs, and been swapping cables. I'm currently using a Macbook Pro (2012) running macOS-10.12.6/RM2.0.19 and a non-powered KPA-head w/Kemper-remote running firmware 5.3.1.

What I see when a computer is connected to the USB-A-port is:

• Random lockups with no indication of anything wrong in the display. Most frequently it happens if the computer has been unused for a while and gone to sleep/hibernation. Maybe not an issue for casual use, but I like to leave everything connected during recording-sessions.
• Lockups with a message about sync of performances in progress on the KPA display. No indication of such activity on the computer.

None of this happens when there is no computer connected. Whether RM is running or not on the computer does not seem to matter. I suspect that activity related to the computer discovering attached equipment when waking up from hibernation is among the things causing the KPA to lock up. It must be over a year since I reported the USB-problem to support. There has since been no other feedback than acknowledgement of the issue.

A beta-release of the firmware some time ago exposed parts of the network-stack on the KPA. A scan of the unit while connected to my home LAN indicated that it may have some form of embedded windows at its core. The exposed network stack would not match any other embedded system that I've come across. I still hope it is not windows, because I have spent thousands of hours over the years troubleshooting embedded devices of various forms and windows is IMHO not fit for purpose. I don't usually care what platform embedded devices use other than that I'm convinced that any device where latency is a potential issue needs to have a real-time-system at its core. This obviously includes anything built to handle audio or video. Microsoft's approach to real-time-solutions is no more than patchwork at best.

Quote from Dragonsf

Any crashes on my KPA indicate an ARM processor as microcontroller. Why should there be any Windows (CE) be involved?

Because the protocol-stack in a beta released during the development of the remote (the one that had a http-server with a remote-simulator) exposed something resembling a full microsoft-server protocol-stack including, if i remember correctly, active-directory. That is not something one would see on any of the proper realtime-platforms I know.

Windows CE (later renamed to Windows Embedded Compact) runs on ARM. Up to version 6 it supported also SH4 and in Compact 7 MIPS. Latest version is named (confusingly) Windows Embedded Compact 2013 and Microsoft is committed to support it up to 2023. It's a different kernel (real-time) but provides some compatibility at the API level with desktop and server versions. I question its realtime-properties are IMHO questionable.

Quote from G String

@heldal: please contact support through the contact form. i'm sure we can help you getting this to work. assuming all this will not help you much.

There should be a ticket in your system already, from way back. I'll see if I can dig up the #, provided that the response still is in my archive.

Quote from Dragonsf

Couldn't it be Windows RT then? CE and consorts have no real-time capabilities AFAIK.

Possibly. Details are just speculation anyway. I saw a MS-server-like network stack and concluded that it may be some MS product, but I'm not even certain of that. It's rather irrelevant to this thread anyway. It's about the KPA freezing randomly when a computer is attached to its USB-port.As a user I shouldn't really care what OS the KPA is based on. It's just my curiosity as a part developer part network-engineer and my personal experiences/preferences with real-time-platforms shining though.