Very slow voltage rise, cap forming, VoltCraft TNG35 linear PSU

[electr_peter]

VoltCraft TNG35 PSU (rated for 0-30V, 2.5A) showed strange response to higher voltage setting once, but does not reproduce fault. It was used as isolated PSU in 2-15V range, never in >15V range.

[ISSUE] I tried setting voltage to higher level (~20V) and to my surprise voltmeter needle raised quickly to 15V, then stopped and very slowly moved up to 17V. Turned down voltage pot reduced output to zero quickly. Turned pot up again, output jumped back to 17V, but stopped and slowly moved up. DMM readings matched panel voltmeter. Potentiometer was turned all the way up several times during slow voltage rise with no speed-up effect.
This was repeated in steps several times till meter showed 30V after few minutes. After this voltage output and panel meter reacts quickly in full range 0-30V with no unnatural slow down.

[TESTS] Testing with scope does not show anomalies. Components were checked in circuit, all caps have nominal values +/-20%. Reverse diode, small cap and 10uF electrolytic cap are connected on the output. This output cap tested as 10uF in circuit (with DMM and LCR meter), but shows ~40 Ohm ESR on LCR meter (not sure if it is real due to in circuit measurement).

[HYPOTHESIS] Based on schematic (not same version of linear PSU, but very similar), my hypothesis is that capacitor across output became sort of zener diode at 15V level, then "reformed" few times. This behavior does not repeat again.

Caps will be replaced. For output cap to fail first is strange because it does not see high temperature or ripple.
Is output cap "forming" an issue here as it was not used for long duration in 15-30V range? Can this be caused by other components failing?

[electr_peter]

UPDATE: Bigger capacitors were checked out of circuit - values were only 5-20% worse than that of new ones, ESR was OK. With new caps, PSU works same as before, voltage rise issue does not repeat. Source of the original issue is still unclear.

[floobydust]

I have seen the linear PSU "creeping voltage" and it was due to damaged semiconductors, either the pass-transistor(s) or the op-amp. Try load it down to say 17V and see if it can still regulate.
Your PSU it could be a leaky C6 1uF 40V did you replace it? It only has five electrolytics or maybe a tantalum so I would just replace them all. ESR tells you zilch about leakage current.

[electr_peter]

C1, C2, C4 and C7 were replaced. Resistive discharge was added to C7 as it contains 47V for days (solder wire exploded on terminals ). C6 is still in the circuit at the moment, but I will replace it as well.
Leakage testing can be done with old caps just to be sure what is what.

Preliminary load testing with electronic load did not show regulation issues. PSU can't quite reach 2.5A at any voltage (2.35-2.4A max, then drops out), but 29V at 2.3A is possible.

I also noticed that after switching off, voltage drops gradually to 1/3 of set voltage, waits for 0.2 s, then drops to 0 quickly and then current meter briefly goes negative(!). Same for 30-10-0 or 10-3-0, which should be related to op-amp behavior at lower voltages.