Erratic needle movement after gauge switch

A couple of days ago I replaced the instrument gauge faces on my '98 RC46 from mph to kph (British import to Denmark) and now the needles move quite erratic. At the same time I replaced the seven W5W with LED W5W.

Symptoms are:
- The needles move to about 20-50 kph and 1500-2500rpm if I turn on the ignition only (not starting)
- It's as if they "hang" at some angles, especially the speedometer. Speeding up fixes that with a sudden jerk up to a new speed (not the correct speed though)
- When the needles are set at 0 (even with ignition on after several tries) they don't return all the way back to 0 after testing speed and revs. Repetitive ignition on/off can drop the needles to lower speed/revs, but almost never to 0
There's a video (72MB) of one of my first tests here: https://drive.google.com/file/d/1s-8gRRGiVwHqHGX-xKKg0BYgi2R_e-iR/view?usp=drive_link

Does anyone have an idea about what's the issue? Did I somehow brake both stepmotors (?) at the same time? It's probably the first time the needles have been removed making it a possibility...
Is there a magic trick or spell that I have to say when reinserting the needles? Should I use a bit of oil on the tip of the metal rod of the needles? It feels like they touch some rough surface at the bottom of the stepmotor. Could the different power draw of the LEDs have a anything to do with this?

Reading this post https://vfrworld.com/threads/new-gauge-faces.44026/ I should be able to set the needle correctly at 0 in both gauges with some patience. It's the erratic movement that annoys me and makes it hard to keep the correct speed.

 

are the needle bases rubbing on the gauge faces? try pulling them up just a hair.

since my bike is currently a basket case i dug my cluster out of a tote and took it apart to see how it worked. the needles can be moved with almost no effort, and gently and smoothly return to zero when released. my suspicion is your needles are too down low on their shafts, causing them to rub. hope this helps.

 

Could be electrical noise from LEDs affecting instruments.