My newly-restored (re-capped) 1950s vacuum tube voltmeter is now operational...but there's one further issue. The meter gives correct readings when lying on its back, but when oriented vertically- as I'd like to use it - the readings are wrong and I can't even calibrate it. Images follow measuring the same 10K resistor.
The meter movement should have balance weights to counteract the pointer's weight, so the meter reads the same regardless of position. It’s likely that those weights became dislodged, so you’ll need to take on the delicate task of rebalancing the movement.
The zero adjust screw won't fix this, it only changes the preload on the meter's spring. You can see this in the photos: although the meter is centered when horizontal, it moves to one side when vertical. That can only be caused by mechanical imbalance.
More here: https://www.nutsvolts.com/questions-and-answers/meter-metrology
And a discussion thread here: https://www.vintage-radio.net/forum/showthread.php?p=1119938
You will find a variety of arms projecting from the moving coil, some with wire wrapped round them as weights (copper or possibly even solder as a field repair), often forming a cross shape. First picture here shows the left and bottom arm, the bottom arm has a sleeve weight on it.
One or more of these weights may have fallen off; if you are very lucky they may be rattling around inside the case or on your desk.
These are adjusted on test, or after repair, to balance the meter in all orientations, not just on its back or vertical. Adjusting them is quite an art, a bit like timing a mechanical watch "in positions".
See Page 14 of the AVO 8 manual for a balancing procedure for another well known analog meter! (This is not detailed enough to follow without other experience and training) However, at least for the AVO 8, the meter should be zeroed on its back (Position 1) and then balance should be adjusted for standing up (Position 4).
The AVO 8 MkV belongs to the subassembly replacement era; instructions for an earlier model are more detailed; see items 18, 19 on page 11 of this manual on hairsprings and balancing, and items 8,9,10 on pages 8 and 9 on pivot troubles.
As the procedure hints, another possibility is damage to the bearings, which are probably jewelled bearings like a watch balance staff. This would show up as stiction, rather than a smooth sweep from one position to another.
Analogue meters might have a preferred measuring orientation. If the meter is DIN certified, the orientation will be indicated on the faceplate.
See top right, under Nennlagen.
(Example found here: https://www.eevblog.com/forum/testgear/exploring-the-symbols-on-a-analog-multimeter-quiz/)
It's not an issue. Analog meters have an orientation they are designed to be used in (or are designed to be used in any position but need to be zeroed specifically for that position).
That needle is a lever with weight on one end.
We're spoiled now in that we can balance our digital multimeter onto a nearby pile of junk and take readings with it.
The meter needs to be calibrated to zero by rotating the spring balance adjustment (which is the thing that looks like a slotted screwdriver head at the bottom center of the meter face). It's not really a screw, it is an adjustment mechanism for setting the "zero volts" position of the meter. Because it's a mechanical spring system, it is sensitive to gravity and thus orientation. In most meter movements I've seen, the resting location is on the left side, but on your meter it says "center zero", so I assume that is where the meter needle should be adjusted to, when the meter is in position but powered off.
There is a good diagram of the internal workings of a d'Arsonval meter movement in this answer here: https://electronics.stackexchange.com/a/331372/35022
