Sgt_Mike
Well-known member
Interesting ,,,
What was the recommended method, has been the method I've used in working up loads for a good bit. (0.3gr increments, 0.003" seating increments, usually best loads are around / within 1ish grs of listed max for the upper node, in my experience)
Although the components I used in the Remington PCR 700 (6.5 CM) was Winchester 6.5 StaBALL powder Barnes 145gr Match Burners, in SRP Starline Brass.
I asked grok using the components listed above just to see how close the response would be to my load.
Grok's response was start at 54.0grs, now a quick review of Hodgdon website shows the max for that powder to be 43.3grs with a 147gr bullet (2grs heavier bullet than my listed bullet). The real life node I found was 0.5gr wide, 42.0 to 42.5 grs.
The bullet seating was within 0.010" of what I had the bullet (I was longer).
Not quite sure why it showed 54.0 gr for a start, or where that came from. Heck I honestly don't think 54.0gr will fit in the case. The only thing I can think of was the manner in which I asked the question? I did list the model of the 700 which is Remington 700 PCR (Precision Chassis Rifle) which may have caused the fluke being so close to the 6.5 PRC round.
However it did list the specs exactly for the rifle in twist, length, barrel profile, including the 6.5 CM cartridge.
Just looked up the 6.5 PRC after typing the above, yeah 54.0grs of Winchester StaBALL HD would be a safe start load with a 147gr bullet. StaBALL HD is quite a bit slower than 6.5 StaBALL on the burn charts. The PRC case has more volume. Grok had to have given the StaBALL HD 6.5 PRC load data, even with identifying the correct caliber.
GIGO --- Garbage in Garbage out.
If it is that sensitive to input yeahhh maybe not for me.
Outside the suggested start load, like I said it did align with the methods I already employ so that is good in my opinion. Understand I'm not slamming Grok /AI just sharing the experience / observation with it. Had I just listed Remington 700 and the caliber then components, the response might have been more in alignment with reality.
What was the recommended method, has been the method I've used in working up loads for a good bit. (0.3gr increments, 0.003" seating increments, usually best loads are around / within 1ish grs of listed max for the upper node, in my experience)
Although the components I used in the Remington PCR 700 (6.5 CM) was Winchester 6.5 StaBALL powder Barnes 145gr Match Burners, in SRP Starline Brass.
I asked grok using the components listed above just to see how close the response would be to my load.
Grok's response was start at 54.0grs, now a quick review of Hodgdon website shows the max for that powder to be 43.3grs with a 147gr bullet (2grs heavier bullet than my listed bullet). The real life node I found was 0.5gr wide, 42.0 to 42.5 grs.
The bullet seating was within 0.010" of what I had the bullet (I was longer).
Not quite sure why it showed 54.0 gr for a start, or where that came from. Heck I honestly don't think 54.0gr will fit in the case. The only thing I can think of was the manner in which I asked the question? I did list the model of the 700 which is Remington 700 PCR (Precision Chassis Rifle) which may have caused the fluke being so close to the 6.5 PRC round.
However it did list the specs exactly for the rifle in twist, length, barrel profile, including the 6.5 CM cartridge.
Just looked up the 6.5 PRC after typing the above, yeah 54.0grs of Winchester StaBALL HD would be a safe start load with a 147gr bullet. StaBALL HD is quite a bit slower than 6.5 StaBALL on the burn charts. The PRC case has more volume. Grok had to have given the StaBALL HD 6.5 PRC load data, even with identifying the correct caliber.
GIGO --- Garbage in Garbage out.
If it is that sensitive to input yeahhh maybe not for me.
Outside the suggested start load, like I said it did align with the methods I already employ so that is good in my opinion. Understand I'm not slamming Grok /AI just sharing the experience / observation with it. Had I just listed Remington 700 and the caliber then components, the response might have been more in alignment with reality.
Last edited: