Replace 20+ years old bulk caps or not?

[NaxFM]

I recently bought an Agilent 34401A and i know it was manufactured at least in 1999.
This means the electrolytic caps are more than 20 years old.
The meter works beautifully and it seems accurate, tho i'll send it for calibration when i have time (and money).
I don't know if it's really worth to replace the caps given that they show no sign of degradation. Still, i know well from experience that when they fail they tend to do lots of damage, and these are quite old.

What would you do? What is your rule when it comes to replacing caps on old equipment? You just leave them there untill they fail or replace right away?

The big electrolytic on the 34401A are just for power supply stuff and away from the delicate analog circuit, changing them should not modify the calibration

[Haenk]

If they are all branded caps (Nichicon etc.), I would not replace them.

[mikeselectricstuff]

If they don't see significant heat, most electrolytics will last almost forever

[Per Hansson]

What is your rule when it comes to replacing caps on old equipment? You just leave them there untill they fail or replace right away?

I always Google the item on forums like this and Badcaps.net and see if people did find problems with the caps.
If they did then I do a full preventive recap, if not and a visual inspection of them and cursory in-circuit ESR checks good then I leave them alone.

[floobydust]

On a linear power supply, I just leave the caps alone if they are good quality and not in a really hot environment.
On a SMPS, they experience higher stress so I tend to outright replace them if they are el cheapo brand or a cheap design (minimum size=high stress=short life). A low value cap on a SMPS can cause it to fail, IC or mosfet to pop so this is preventative.

The 34401a I'd leave the filter caps alone, they're Nichicons. Some tantalums have roasted for some people but they smoke to let you know they need replacement.

[abdulbadii]

"What would you do? "

Obtain quality LCR meter, or just C meter if it'd save money, one that has on-board testing meter capability

Find out every cap's ESR and its function validity one by one

[james_s]

Inspect them for signs of leaking electrolyte, if they look ok and the meter is still working I'd leave them alone. Any time you replace a part it risks damage to the PCB, and I've seen people replace older but good quality capacitors and other parts with new cheap garbage that doesn't last.

[BrokenYugo]

I have some consumer grade junk twice that old running fine on all original caps. I'd let it be unless it's back in the tube era or a known weak point e.g. those 90s SMD electrolytic caps that tend to fail destructively.

There's a lot of undue paranoia around electrolytic capacitors. I guess from the plague in the 00s and shit tier consumer electronics in general, but about the time tubes fell out of favor they got quite reliable if properly made and specified for the job they're doing. If money is of little concern, like if you're designing a piece of top quality professional test equipment, it's often easy to over spec one to a point where the lifetime calculations give nonsense answers of 20+ years of continuous operation to wear it out, sometimes over a century even. The lifespan is more or less limited by the quality of the seal at that point, and many have held on for 40+ years at this point.

[bdunham7]

If you have a capacitor problem with the 34401A it probably won't be the PSU filter caps, but rather some of the SMD caps on the main board.

[NaxFM]

Thank you everyone.
I've verified and they are all original nippon chemi-con caps that look absolutely perfect, so i guess I'll just leave them there