Issue with Avenger R8-RS headset

[Pehem]

Hello,
I would like some thoughts about an issue with my headset. I am unable to have zero play in the headset. I have to screw the headset to its maximal torque to remove as much play as possible but I still have some minor play. The screw is a M6 titanium countersunk head which I found a max torque of 8 Nm for. So with the stem loose, I screw the headset screw to 8 Nm and then I screw the stem to 8 Nm (2x M5 titanium), that seems a lot but it is what the manufacturer gives. When I firmly push the front brake, I can still feel some play in the headset, I checked that it doesn't come front the front rotor which is torqued to spec.
- There is still 1,5 mm of space between the top of the expander of the lower end of the top plate.
- I checked that the screw is not topping in the thread. When I remove the plate, I can insert the screw way more down.
- At 8 Nm the bearings are absolutely free, which is coherent with the fact that there is still play.

At this point, the only thing I would try is putting a 1 mm spacer between last spacer and the stem if I messed up the measurement but I can clearly measure space between top plate and expander.
Do you have any ideas ?

[Serge_K]

that stem torque sounds high, but if it's what it says on the tin...
I would torque the top cap more, and the stem less (or the same).
I have never heard of gear destruction from high torque on top caps - but carbon stems do break fast. Point here is that i dont think you have play because the stem is under torqued, but it's likely because the top cap / whole vertical column is under torqued.
ofc, if you torque higher, makes sure the compression plug isn't slipping up the steerer tube.
Once you've removed headset play, you can back off the top cap torque, as stuff is hold in place by the stem at that point. all the more reason to temporarily torque up that stem cap bolt.
years ago velobuild had a C ring in metal, solid piece, prone to play. They then shipped one a slot, solved many problems. Since then the C rings i ve seen have been made of plastic. if your bearings are sitting clean in the frame (no obvious tolerance issue), as long as the bearings don't start to grind / rub, i think you can continue compressing that column a fair bit. I wouldnt be too concerned until at least 12nm, because you'll back off once the stem is torqued.
Ofc, you're torquing the stem AFTER you've compressed the column, right?

[Pehem]

Thanks for your answer. Yes, I'm tightening the stem after headset compression. So I tried to go 12 Nm on the headset, I was afraid to do so but it seems like the stem went down a bit more. I was concerned about the expander but it didn't move (nor in rotation nor vertically torqued to spec 12 Nm) and I gained 0,3 mm, now only 1,2 mm space vs 1,5 mm with still free bearings. I tightened the stem to 7 Nm (instead of 8 Nm) and then released the headset a bit and torqued it back to 6 Nm. It seems to be better but I'll have to confirm with a ride, and especially when riding on the rough concrete part next to my house.
I think I'm understanding what's happening. I must have a lot of friction going on in the stack below the stem. I'm actually using a lot of spacers (x6) and they are plastic. Each interface gives friction and the plastic may continue to soften after compressing the headset resulting in play coming back.

[CoffeeNbikes]

Did you solve your problem long term? I have same issue. Avenger torque spec says 4 nm top cap not 8nm. Anyhow can seem to find a solution yet. It seems fine in the workshop but headset loosen during ride.

[Pehem]

I didn't know about the torque spec, I didn't receive the manual. When I asked them, they gave me much higher values 8 Nm for the stem (it is written on it but I couldn't figure out whether it was 3 or 8 Nm) and 12 Nm for the expander. Torque spec of the top cap is not really relevant because the torque needed to remove the play is different for every bike. The best torque is as low as possible with no play. The value of 4 Nm they gave is maybe an average value, but certainly not the max torque the screw can withstand.
So at the beginning, I screwed the top cap (probably lower than 4 Nm) enough to have no play in static with the front brake test. After a ride, I had play. I progressively screwed the top cap higher and higher until 8 Nm (max torque I found for this type of screw), it was better but the play came again after a few rides. I tested then 12 Nm (I'm pretty sure the screw can withstand 20 Nm before rupture), again much better. But still not perfect. The play is less and it has taken more time to comeback but I still have a slight play. My guess is that the plastic spacers (I use 6) are creeping with load, they reduce slightly in thickness and thus preload of the headset dissapears. Now I'm thinking of repeating the procedure I mentioned in my second post. Maybe the creeping will reduce over time if the plastic becomes harder. I'll look if the play takes more time to come back. If I don't resolve the issue I'm thinking to modify the spacers : remove material in the middle to put hard carbon spacers and use the original plastic spacers as a "cap" for aero and aesthetics.