Vibration Log problems

Since I had a crash, which destroyed my main Glass fibre blades, I’ve had a vibration problem. I rebuilt the copter, with new carbon fibre blades, which weigh less than the glass ones, new main shaft, main gear ( it had lost a tooth) and tail blades. The blades have been balanced, and checked. The heli is fitted with a Pixhawk, mounted in an extension under carriage.
I’ve enclosed a two bin logs taken from the SD card.
The ‘6 bin log’ is with Heli suspended from elastic bands, with No main blades attached and spooled up.
As you can see there is very little vibration.
The second ‘8 bin log’ is with blades attached and doing a test flight. Now the ‘Y’ axis is showing excess vibration, and the ‘X’ axis has hardly changed.

Can anybody suggest how I can cure this.

The big “vibration” is 2hz. It is really caused by your roll controller oscillating. You don’t seem to have much vibration at all.

Thanks for your reply
On the bin 8 log it shows -10 to +10 for the ‘Y’ axis or am I reading it wrong. I understand that anything over +3 or -3 is to much vibration.

[quote=“OAPpilot”]Thanks for your reply
On the bin 8 log it shows -10 to +10 for the ‘Y’ axis or am I reading it wrong.[/quote]
Well, you are and you aren’t. You see -10/+10 on your graph, yes, but what you don’t see is the frequency content. Vibration would look like white noise - just random spikes everywhere. But if you zoom into your log, it is not white noise, it is a sine wave at 2hz. Looking at the X (roll) gyro, you can see that same sine wave. You don’t have a physical problem here as far as I can tell, you have a tuning problem.

The Y accelerometer is acceleration along the lateral axis (positive is to the right). The X gyro is rotation about the longitudinal axis (roll).

Yes, if it is actually vibration. It is really much more complex than just magnitude - the frequency content matters more. If the vibration is fast enough, it can cause aliasing which wrecks things really badly.