You may have something foreign in a QD that is blocking or restricting the fluid flow. You mention it's intermittent, so the hydraulic fluid flow may be moving it around so it works at times, and then doesn't work other times.
Is your setup configured to allow easy QD swap between the pair of ports for the loader and the pair of QD ports for the bucket?
Are you able to disassemble the suspect QDs on the (4) individual QDs used at the demarcation for bucket control to visually look for something foreign inside the fitting? They're generally pretty basic and straight forward. Check the male sides first, look for foreign material in each of the orifaces.
Apologies if the following is already well known by you. It sounds complicated written out, but it's pretty straightforward when the objective is realized.
The goal is to separate the components in your configuration which are proven to be working normal, and the components on the suspect list.
So if one function appears to be working normal...say, the Loader Arm control (the loader arms go up and down normal), then all of the components in that lineup are OK to be used for testing the components supporting the function not working...say Bucket Control.
There is a loader disconnect point, call it the demarcation point, typically with (4) QD ports on your tractor, which connect with (4) associated QD lines on your loader to facilitate easy separation of hydraulics for removal of the loader.
They operate in pairs and the flow in each pair of QDs is typically activated by separate hydraulic valves using the joystick.
One valve and set of QDs operate the loader arms. Another valve and set of QDs operate the bucket.
This is where it gets specific to each unique hardware setup: my Kubota facilitates easy rolling of the pairs at the demarcation. Yours may not.
Your description seems to be an intermittent problem affecting the bucket control. So likely safe to say the valve and the 2 individual QDs on each side of the demarcation point (total of 4) supporting the loader arm operation are working normal and can be used to isolate the problem with the bucket control.
The objective, if your QD configuration allows, is to use the good working components such as from the loader arm to isolate the issue with bucket control by simply rolling each pair of QDs at the demarcation. (Example is at the demarcation point, plug the pair of QDs from the loader side which are used for loader control - > into the QDs on the tractor side used for controlling the bucket).
Use what is working to test and identify what is not - using thoughtful logic to swap known good components in the line-up (ideally only one at a time, to minimize variables) with suspect components to finally isolate the one(s) at fault.
Always before starting, make sure you physically tag each QD component in your original configuration so you can always return it to normal...colored tiewraps or tiewraps and labels will work...and take "before" photos just in case. Write it out on paper if needed, plus the swap plan.
If your male/female QD use is consistent and allows:
To prove if your bucket valve at the joystick is good - ideally at the demarcation point, you would simply reconnect your pair of QDs operating the loader arm into the pair of QDs on the tractor side which operate the bucket, and do the same roll with the remaining QDs - plug Bucket QDs to loader arm QDs on the tractor side.
So now your loader valve should operate the bucket and your bucket valve should operate the loader arms.
Test to see what works and what doesn't. Your original joystick motion for the loader should now control the bucket.
The joystick movement originally controlling the bucket should now operate the loader arms.
If both work normally, run it that way for a bit due to the intermittent symptoms mentioned, to determine if a problem still exists and what is affected.
IF the issue moves so the loader arms are not working, it could be your QDs originally controlling the bucket on the tractor side at the demarcation point, or the valve at the joystick controlling the bucket.
IF the problem stays with the bucket, it is likely something affecting one of the bucket QDs on the loader side of the demarcation.
Return all to the original configuration.
To further isolate, try then swapping individual QDs at the demarcation one at a time, depending on which function and which side of the demarcation seems to be affected using "known good" QDs from the loader arm lineup.
Return to original configuration.
The key is to step back and think through what pieces and parts seem to be working OK, and use them to test and isolate the suspect components. What makes yours a bit complex is the intermittent nature, and difficulty recreating the symptoms on demand.
