@Bud
Thank you for your excellent efforts. I recall you criticizing the PLL when the jitter problem was first discovered. I think your findings prove without a doubt that your criticisms were all correct. The PLL fail cannot be blamed on building down to a price point. Since you changed only component values, I doubt you impacted manufacturing cost at all in the PLL circuit and by no more than a few pennies overall. The tantalums are a bit more expensive but I'll bet that with sufficient effort an equivalent cost solution could be had. There is no excuse for Rigol not to have gotten this right. I do agree with what folks said about translation from engineering to production. I've experienced that myself. It could explain the wrong ESR caps on the LDOs, but not the wrong values on the PLL.
One thing, most of us do not have the SA to confirm our fixes. We would be doing it blind and taking it on faith that our scope performance was improved. Replacing caps is relatively easy, but those hand-fabricated inductors are likely beyond me. It would be of benefit to many, I think, to swap back to chip inductors of the proper value and retest.
@IanJ
Great opportunity for a "repair" video! I've really enjoyed to ones you've done before. If you do it there are several places that you could add value. One, if you have a 1054Z it would be good to perform the fixes and IDing the component locations on that model. Two, you could use chip inductors instead of hand fabricated. Three, if you could perform tests that demonstrated performance improvement using more common tools (DMM, another scope, signal reference), that would allow more of us following along at home to verify our results.
@The folks who question the value of the fixes
I agree with the person who said that ripple is not likely to cause crashes. By its nature digital logic is fairly robust. It's far more likely that crashes are caused by flawed software. Besides, the LDOs Bud fixed were for analog circuits. The impact of those would be things like offsets just as Bud wrote. Also things like trigger sensitivity and dare I hope that cleaning up the supplies would improve the noise floor. I would love to see before and after captures of a 10 MHz 2 mV signal.