Bolt catch snaps for one of two reasons.

The first, the bolt is retracting back too far from the catch is has too much of a run at the catch.
To firgure out if this is the problem, pull all the way back on the charging handle and look where the bolt stops against the back edge of the ejection port window edge.  The correct distance should be 1/8" to 1/4" of the bolt stopping in front of the ejection port window. If your rig has the face of the bolt stopping even with, or behind the ejection port back edge, remove the buffer and recoil spring , and stop dropping quarters down the tube as back of tube spacers to get the bolt face to stop the correct distances on a full charging handle pull in front of the back of ejection window edge.


The second reason that a bolt is correctly stopping the correct distance in front of the back of window edge (correct distance behind the bolt catch itself), is if the buffer is not controlling the back of stroke to create the end of stroke stall, and instead the buffer is hitting the back of the receiver extension to sling shot the bolt returning forward even faster isntead.  Hence the buffer has weights in it to create a dead blow effect, both when the bolt is cycled all the way back, and when the bolt is fulling locking up (so the carrier does not slam into the face of the barrel extension to bounce back to unlock the bolt, Hence bolt bounce).

So yes, if needed, you can use a stronger or reduced tension recoil spring, but since they may be creating a problem with the bolt cycling back too fast and the buffer not have a dead blow effect to stop the buffer from sling shooting off the back of the tube (instead of dead blow effect stopping for a brief micro second) a heaver buffer has to be used isntead.

Trust is, with both a extra tension recoil spring and the use of a T2 buffer to get the rig to cycle correctly (hell of lot of force that is tagging the bolt catch on the way forward to stop the bolt), this is just screaming that the barrel gas port is to large and the problem at hand isntead.

So myself if the problem is not just the bolt retracting too far back from the catch to begin with, would attack the problem focusing on the gas port problem itself; instead of the heaver buffer or increased tension recoil spring to band-aid the problem that way instead.


Welding the gas port closed to allow it to be re-reamed to the correct needed smaller size is a option, but only when you have the needed equipment to do such correctly to begin with.  If such is not the case, then either an adjustable gas block or just installing a pig tail gas tube would be a good starting point isntead (lessen the amount of gas, or slow it down before it reaches the carrier chamber to unlock the bolt).   Hence here, it not the ST buffer so much that we want to get out of play (very good at it's dead blow abilities), but to get the extra tension spring out of play isntead (slow down the speed in which the bolt is running at the bolt catch from it normal back position isntead).  And, with the gas unlocking the bolt at the correct time (instead of it unlocking the bolt too soon), this too will slow the bolt cycle speed as well (so you don't need the heaver buffer or stronger recoil spring to slow the bolt unlocking too fast isntead.

The fact that you already have an adjustable gas block in play, guess that you did not start with just a standard 2.9oz buffer and standard tension recoil spring, and tried to adjust the gas block down to get the rig to cycle with them from the beginning.  This should be your first attack plan to the end of line problem of the bolt catch snapping,since when you do this, going to find that the rig is going to have a hell of a lot less felt recoil as well.

Simply put, dam near all the felt recoil of the rig is the buffer moving in the receiver extension, and when you go to a heaver buffer and stronger recoils spring, all you are doing is increasing the amount of felt recoil instead.  Heaver buffers are needed in some rigs, but it's do to the ammo that may be used in then instead. In a standard carbine,standard buffer should all that be needed.  But in the case of a M-4 (why it has a 1/7 twist barrel) it also designed to run with Mk262 ammo as well.
Hence in the Mk262 ammo (69gr), slower burning gun powder is used for this load, which moves the peak burn dwell of the powder closer to the gas port, and this cause higher barrel gas port pressures and the bolt to unlock the bolt too soon.  So here in the Colt M-4 with 14.5" barrel, the heaver H  buffer is used to slow the bolt unlock down a touch (but still run with m193/m855 ammo), and then a heaver extractor spring is used on the extractor to make sure that the spent case is extractor correctly as the bolt is still unlock a touch too fast with the Mk-262 ammo (residual bore pressure a touch too high, spent case pressured welded to the chamber a touch to much at bolt unlock with the Mk262 ammo, even with the H buffer in play).


One last thing to point out if you have not figured this out yet,
With the gas port so close to the chamber, the 9.5" barrel is going to ammo sensitive as all hell.  
Its not as bad as a 10.5" barrel with the gas port so close the muzzle, but neither the same, with such a short barrel, a change in the gun powder burn rate with too far of a swing from the norm, is going to cause you nothing but head aches.

If you are shooting Nato ammo, then it's designed around a fixed barrel gas ports placements and sizes, and the power for the most part has the same burn rates for it (in the case of M193 and M855).  If you are using civilian or imported ammo for say a bolt action gun with a much lighter or heaver bullet, then can bank that the gun powder burn rate will have a wide swing from the standard gun powder burn rates used in say nato 55/62gr ammo, and this will cause either higher or lower gas pressures through the gas ports.   The fact that you have an adjustable gas port, will allow you to dial the rig in for these ammo, but don't expect that your going to come up with a 1 setting for all that is going to work for all of it isntead.