Barco 9xx flash memory back up

Does not ring a bell to me. Never heard about the term.
What "key usage" stands for ?

Well this is all assumption but in this instance it could mean that the menu has advanced item selections or more range of adjustment or particular adjustments. Possibly for a special projector installation where more control or manipulation is required. The EKU must stand for something, however, is it in English or some other language? I don't know.

You can change them to suit yourself for what you want there, but only on Cine 9 software. Its very convenient to have.

I carefully listened with the headphones at both videos and indeed it sounds like in the first clip, the sound is a bit lower.
However, for best comparition, you should swap the controller only. Else, other things can influence the noise (e.g. another front fans, another aluminium tray with his own vibrations and if i remember correctly, the aluminium air-deflection cone infront of the focus board was missing in one 919 tray ).
Then you can go with what ever soft. version is the most silent. Also, don't forget that making those metal upgrades will reduce the noise further.

Update.

First reason I made this programplug adapter was so I could go back
to previous file if an update is not as expected and this works well.
Second reason gain knowledge for repair ,I have one controller board that works through rs232 (remote control prog) but not with internal or external remote and I could not find any hardware malfunction except from the reset chip that did not generated resets anymore.From this I suspected software corruption ,I replaced reset chip and made a copy of main program I6 and I7 from other working controlboard out of my cinemax and programmed it in the defective board but still no remote control.

The memory off the control board consists of four memory chips (I6 I7 I8 I9),I never considered I8 and I9 important because I thought they only hold mainly data for the settings and they are not accessible throug the programplug alone , the chip enable line for both chips are still missing.

But recently I went for it , made an extra adapter for I8 and I9 .
The chip enable lines are available at chip I4 (isplsi2032-80)
So I used an plcc socket 44pin cut out the back plastic and now you
can put it upside down on I4 making good contact so no need to solder the extra wires.

The defective controllerboard has seos prodas on it and flashed to v2.21 (before I had this progger so v3.12 was lost)
To test first I made a copy of the I8 and I9 chips from the working cinemax controllerboard with v2.21 (cine9) on it.
I first replaced I9 ,and bingo the remote works again (and everything else as far if I could see) so there is data related to the remote in I9, also the seos prodas in the menu is gone and is now cine9 , so menu id is also in I9.

Now I replaced I8 and the hour counters and serial number are now the same as on the cinemax controller ,so hour counters and serial number is in I8.
So replacing all four memory chips from one board to another you can clone a board this works with my boards backwards and forwards.
But doing the same with the boards of Hulio the hour counter values where a little bit of reason not known yet.
There is still alot that needs to be checked though.
And the boards I use are different revisions.
Looking at the data file from the chip I8 I can not find any hexnumbers that are the same as the hour counters or serial number to chance them in the file.

So controllerboard finely totally working again and with the gained knowledge the work was more then worthed.

Some new findings.
The core databus is 16bits ,I8 and I9 are 8bits memory,so barco used two 8bits chips instead of a 16bit one.I had there relationship a bit ignored because the hardware of the board makes it possible to read and write them separately and there is no reason for that if they act as one chip.Looking at there content ,they contain several banks with data with empty banks inbetween the bank addresses from I8 and I9 match each other perfectly.There is however a difference in the addresses used depending what firmware is used , so there is a difference when cine9 firmware is used or the br909 v3.12 firmware.
that explains why the hour counters where not zero with the boards from Hulio we had swap files for I8 and I9 so we made some FRANKEN data.
Reparing the board that failed from Hulio (replaced I9 chip) I left it blank (total content FF) to see what happens .
Seems the program does check the content of the data memory and a blank memory makes the program to totally rewritte I8 and I9 resulting in a clean start up ,hour counters zero ,serial number _ _ _ _ (same as a replacement board I learnt from Hulio) and all settings mid point.(this also removes seos and brings it back to Barco when cine221 is used).So erasing I9 is a way to bring the board back to virgin state (software ways).

I also found the two 16bit words that represent the serial number
I'am now able to change this number without affecting anything else
not important but fun to do.

With the help from Hulio who let me extract (programplug) the br909 v3.12 firmware from one of his boards and Justin who wrote a byte swap program I
managed to convert the file back in to a flash file ,and it seems to work fine ,I just flashed (rs232) my board back to a br909 v3.12.