Months after building the first Fireball, I finally got a chance to chrono some loads and it nearly didn't even happen. I hadn't had a chance to sight in my rifle and the first round came within an inch of clipping the chronograph. 
The following is out of a 10.5" barrel.
Supersonic
110gr bulllet, H4227 - 2048 fps
125gr bullet, H4227 - 1998 fps
150gr bullet, H4227 - 1693 fps
One thing I noticed about these loads is the velocity spread is generally quite low. For example the 125gr load had a standard deviation of only 6 fps. I suspect this is due to the small case being quite full of powder.
Subsonic
- all with 200gr uncoated bullet.
H4227 - 846 fps (obviously this is slower than optimal but the rifle cycles fine with this load - needs a bit more work)
H322 - 997 fps (fricken PERFECT - I am going to drop 0.1 gr and go with this load)
Lil Gun - 1028 fps (slightly strange results as I got higher velocity with less powder)
HS6 - ~1100 fps (this load is too hot and still doesn't lock the action back on an empty mag.)
W296 - 983 fps (fricken PERFECT #2 - slightly more velocity spread than H322 but still a good load)
Notes:
H4227 and H322 both work well with subsonics but they tend to produce a bit more muzzle noise than the faster powders. However they are reliable performers and fill the case enough to be safe. They can also both be used for supersonic loads meaning one powder can be used for every kind of ammo.
Lil Gun: I only had a small amount of this stuff and won't be buying any more. I think it could work well but I have enough powders on my bench at the moment.
HS6: This stuff is better left for bolt guns and manual actions. It is not reliable in a semi-auto. I wanted it to work because it is likely to be quieter than the slower powders and I have a can of it I'd like to use up but I also don't want to blow myself up or wreck my rifle and short stroking is just a drag.
W296 may be the best choice for subsonics. It should be reasonably quiet yet produces easy and safe subsonic cycling in an AR rifle. A load 0.3 gr less than the one that produced 983 fps still cycled and locked back the rifle reliably and I suspect it could go even lower reliably.
The following is out of a 10.5" barrel.
Supersonic
110gr bulllet, H4227 - 2048 fps
125gr bullet, H4227 - 1998 fps
150gr bullet, H4227 - 1693 fps
One thing I noticed about these loads is the velocity spread is generally quite low. For example the 125gr load had a standard deviation of only 6 fps. I suspect this is due to the small case being quite full of powder.
Subsonic
- all with 200gr uncoated bullet.
H4227 - 846 fps (obviously this is slower than optimal but the rifle cycles fine with this load - needs a bit more work)
H322 - 997 fps (fricken PERFECT - I am going to drop 0.1 gr and go with this load)
Lil Gun - 1028 fps (slightly strange results as I got higher velocity with less powder)
HS6 - ~1100 fps (this load is too hot and still doesn't lock the action back on an empty mag.)
W296 - 983 fps (fricken PERFECT #2 - slightly more velocity spread than H322 but still a good load)
Notes:
H4227 and H322 both work well with subsonics but they tend to produce a bit more muzzle noise than the faster powders. However they are reliable performers and fill the case enough to be safe. They can also both be used for supersonic loads meaning one powder can be used for every kind of ammo.
Lil Gun: I only had a small amount of this stuff and won't be buying any more. I think it could work well but I have enough powders on my bench at the moment.
HS6: This stuff is better left for bolt guns and manual actions. It is not reliable in a semi-auto. I wanted it to work because it is likely to be quieter than the slower powders and I have a can of it I'd like to use up but I also don't want to blow myself up or wreck my rifle and short stroking is just a drag.
W296 may be the best choice for subsonics. It should be reasonably quiet yet produces easy and safe subsonic cycling in an AR rifle. A load 0.3 gr less than the one that produced 983 fps still cycled and locked back the rifle reliably and I suspect it could go even lower reliably.
