The Great question: which ESR meter is the best for in and out of c.

[TimFox]

VNAs work with a "reflection bridge", and require calibration to short, open, and 50 ohms with well-known standards.
LCR meters do a phase-sensitive measurement of the voltage and current on and through the unknown and compute the relevant parameters from those data.
They require only a short and open calibration, with no important transmission-line effects at relatively low frequencies.
Especially when measuring extreme cases (high C and low C for this case), a fresh calibration is needed for both types of units.

[LinuxHata]

Well, I calibrated it again for sure.
Zero changes.
caps 1000uf and below - ok, bigger can be measured only on 100 and 120hz, above that - 0L
I also had Victor 4080 LCR bridge, which has color screen and does not have that issue (but you still need to press several buttons before you reach ESR)
But it started to fail, so I discarded it.

[Shonky]

caps 1000uf and below - ok, bigger can be measured only on 100 and 120hz, above that - 0L

As per the specs. Have you looked at the manual?

It will do Cs measurement up to 2000uF at 1kHz

[TimFox]

I calibrated my unit carefully, and measure the 1500 uF capacitor according to the DE5000 specs.
Note that "calibration" here really means computational compensation for the series and parallel impedance of the test fixture, and needs to be done not at the front panel but at the point where the DUT is connected.
For my TL21 with small alligator clips, that means doing the "short" at the same place on the clips where the capacitor will be connected.
(This type of LCR meter relies on an internal precision resistor to set the scale factors.)

[Shonky]

Here's a brand new United Chemi-con EKMG250ELL222MK25S 2200uF/25V I just purchased from element 14 last week very quickly tested on my near new DE-5000 at 1kHz. i.e. it just under the 2000uF maximum range.

[TimFox]

Looks right to me, and is consistent with my similar measurement.
By democratic vote, we may conclude that LinuxHata's calibration is incorrect.

[EPAIII]

You are talking about 40 year old electrolytic capacitors! You can, of course, do as you wish, but in my experience with consumer equipment the capacitors are often sourced at the same time, from the same manufacturer and they often dry up at the same age, plus or minus a year or five. When I go to the trouble of taking an item apart and accessing the components for testing, it is not much more trouble to just replace all suspicious electrolytic capacitors. And my time is certainly worth much more than the cost of the replacement parts.

That's what I do. YMMV.

Yes, solid state radios 70s and 80s.

I am planning to replace filter capacitors in PS and also if there is another bigger one on some board.
Do you think it is neccesary to replace all electrolytic caps in radios from 70s and 80s even smaller ones? or just the ones that are out of spec?