Exactly which version QGC and AQ FW are you using here.
Apart from the odd Gyro plot (which is probably some logviever vs fw version issue) and the 0 value on the acc align, things are looking normal for the first 3 steps far as I can tell.
For the gyros, try and plot voltage 0, 1 and 2 instead.
For the dynamic datas you are not allowing any still time at the beginning and end of the dynamic. Its important to have clean data in the beginning and end of the dynamic for at least 10 seconds...I usually let 20 pass.
500 MB is too long for sure. 90 mins produces around 350 Mb and thats the max I ever used. I have seen a tendency that too much data can screw things as well. Take a look at the temp1 curve and stop when the line becomes more or less flat. Thats usually in 60 or 90 mins depending on the container used and the lowest temp of the freezing cycle. You can cut them down from the back using the logviever cutting tool.
You can also try just using 1 or 2 static files.