Joining two brass components - advice?

I have a 8mm long, 10mm round brass shaft (A) to which I need to join a 7.7mm splined shaft (B) - [U]as strongly as possible.
[/U]
Thus far I have tried:

Threading shaft (A) to 5mm (fine 0.5mm), threadlocking into 5mm threaded socket on B - sheared when looked at heavily!
Turning down splined shaft (B) to 6mm, drilling B out to 6mm and inserting a 1.5x10mm springpin - sheared at pin drilling! BUT the drilling was only 2mm from the shoulder.

The weak point seems to be the shoulder all the time. So, my next thought is turning the splined shaft down to its inner diameter 6.8mm for 7mm then drilling part A for a 2mm brass pin say half way up and using a cyanoacrylate adhesive on the whole thing. It does not need to be demountable but it does need to be simple to assemble...

Alternatively, I thought a 6.5mm fine thread on the splined shaft, tap part A to 6.5 and then use a lock nut to lock it to part A?!

What's wrong with my ideas? Does anyone have any better please?

What is the intended use of the two components? Have you tried using a collar instead of joint pins/threading? What about TIG welding the two parts? Welding would be awkward for such small parts and I'd be worried about heat deformation but it's something else to consider. Can't really give any more advice without knowing the use.

You might consider turning down the end of the splined shaft, drilling the round shaft and silver soldering the two together.

Have you determined what the loading, stress, strain and deflections are actually required?

Are you sure your design isn't adequate?

Thank you peeps.

COD. Hmm. I have to be careful describing the use due to wanting to protect what I'm doing. Suffice to say, the side with the splines on will have a tap head attached and the other end something to do with a tap valve.

JA. We have considered that. Unfortunately, the intended assemblers are trades people who would rather walk away than do something as 'complicated' as silver soldering which means no take up!

KB. Do my answers to the others help you? Manufacturers of the orignal items tell us that the shaft had been tested to 12 Nm and had not failed. When you translate that force to how you'd use a typical tap head, that level of force is quite difficlut to replicate in normal use. We do not need a specific force as our design has other issues, but I am concerned that we have designed in an inherrent weakness (the join) so wish to demonstrate we have done the very best possible within the constraints of the designed use...

Aplogies for the secrecy, but I hope that all makes sense.

[QUOTE=YorkshireDave;11944]KB. Do my answers to the others help you? Manufacturers of the orignal items tell us that the shaft had been tested to 12 Nm and had not failed. When you translate that force to how you'd use a typical tap head, that level of force is quite difficlut to replicate in normal use. We do not need a specific force as our design has other issues, but I am concerned that we have designed in an inherrent weakness so wish to demonstrate we have done the very best possible within the constraints of the designed use...

Aplogies for the secrecy, but I hope that all makes sense.[/QUOTE]

No, does not help. Joining two brass parts together? I doubt you're doing anything truly special...

I'm not doing anything 'special' I agree. Tap in this instance is your faucet, so the forces involved are turning and typical of those involved in the daily use of a faucet.

As I'm looking to patent the thing I dare not say any more without compromising any application.

It does not sound like a very complicated component - why not just machine the whole thing from solid ?

It's not complicated. However one part must be changeable. Hence the join.