Bad W231 powder ? velocities up/down

I have been casting 130GR plain base HP lead bullets (sized .358) and using commercial Lyman #2 alloy.  Using a S&W revolver K frame with Clark 1-10 Douglas barrel.  The bullets were shooting great groups at 50 yards in the 1.5" to 2" range...but recently they are opening up to the 3" range.  (Double action rested)

My competition power factor goal requires approx 950 fps, using two different chronographs my six shot strings have been:

867,828,904,827,899,798 fps    -and-  895,869,823,799,774,775 fps 

Using a Lee Loadmaster with standard disk powder drop.  I just did several sample powder drops and they consistently drop at 4.5 gr, so the powder appears consistent.  The bullets are consistent at 130gr.

When firing double action I (think) I can feel slight variations in recoil??? I'm not sure what an acceptable FPS would be.

Need to increase the powder drop but have any ideas why the FPS is not consistent?

What primers?
Neck tension?
Crimp?

Most likely culprit is not enough crimp.

For me, Federal Gold Metal Match primers tightened groups beyond 25 yards shooting PPC!

's

Crimp...And W231 is very temperature sensitive.

I'll experiment with my old Lyman taper crimp...had good luck with it years ago. Using Federal Premium Match small pistol primers.

Using the Lee factory crimp right now.

Have you recently changed crimp or the 231powder lot and these problems arose?  Was the temp different?  You have over 100fps extreme spread in a 6 shot string.  Has something changed ? - if so, look to the change(s) first before you start to make modifications to a previously acceptable system.  A few days ago I had a discussion with a HM regarding differences attributable to powder position and ES i.e., "powder up or powder down" which is known to have an effect on velocity, especially small loads in a larger case like 38 Spl.  Powder position needs to be consistent, and firing double action at 50yds is not the best way to assess accuracy, and be able to demonstrate a significant difference between 1.5 and 3" groups.  If you can, get to a Ransom Rest, and retest.  Please don't take offence, but I often see folks chasing their own tails, making unneeded modifications to their loads/process without a good reason.  I hope you find your solution.