Will Elec said:

the instructions state 3Nm for neutral

Don't they say 3 Nm for the 'neutral link' screws, all incoming/outgoing terminals are I believe listed as 10 Nm, at least according to this instruction leaflet: https://www.eaton.com/content/dam/eaton/products/low-voltage-power-distribution-controls-systems/lt835-instruction-leaflet.pdf ?

Yes they do, the terminal in the pictures is one of the incoming/outgoing. The 104gc doesn't have the neutral link but I just tried the 3Nm to see how it felt in terms of clamping effectiveness on the cable. 10Nm with such a small screw just feels very hard to achieve!

I think the only people who can answer this are Eaton themselves.

Have you discussed the breakage with them?

I have yes, they insisted on 10Nm but the tech i spoke to seemed to be just quoting the pdf you have linked to.. I guess it begs the question of how close you should follow the manufactures advice when in the real world it seems incorrect..

If you tighten to a lower torque, you reduce the pre-load (compressive force on the wire) in direct proportion, minus a loss for thread friction.  The bolt looks flat nosed, so the pre-load will be the force over the area more or less as a flat.
So at 3nm you are holding the wire with more or less 1/3 the clamping force - if that is enough to  puts a good full area 'flat' onto the cable you are using, then you are on the right side of OK. If the screw force is not enough to flatten the cable to make good contact, then the torque is not enough, or it is the wrong kind of cable.

I think the regs actually say 'take account of instructions' not 'follow blindly' so if the cable is being well clamped feel free to tell Eaton their advice is incorrect.  And ask them if the can send you some spare screws.

If you know the thread size on that screw  (or diameter and length  plus  of turns over the length ) I can show you how to calculate the forces on it if you like.
Mike

I have experienced exactly the same problem, but with a 63 A unit.

The flat bit on my wrench could have had a better profile and it tended to chamble the slots. I had to replace them with hex drive grub screws.

One also has to be careful not to damage the bakelite.

I was surprised to find this problem given that the Glasgow devices are otherwise very well made.

Could the issue be that it's a 5/16 inch screw - so the metric slotted screwdriver bits are 'wrong' ?

It will depend slightly on 5/16 by what thread form, but in that leaflet it seems to attract the same 10Nm torque as the M8s, despite being a touch smaller.  Now in tables I have here, M8 in brass with 63% copper is listed as absolute max 9Nm while  10Nm is in comfortably in the middle of the range for steels. Now of  course it may not be that exact brass, but it is clearly sailing towards the top end of the possible in terms of stripping, let alone head shear- I do wonder if the 'steel' figure has been picked up in error when the datasheet was written.
Mike

I can confirm they are indeed M8 not 5/16, something tells me that the instructions have been repurposed from older units (made in England with better components!) and not be updated. Eaton are sending more screws but as you say Chris some M8 grub screws would be far more appropriate if the 10Nm is to applied

Will Elec said:

made in England with better components!)

Or made for USA market ... 5/16 is a common US thread size and Eaton sell switchgear there.