You're in the perfect spot to answer quesitons about the AGM batteries, then! I'll trade questions with you....
The stock battery voltage on these CB750 bikes is in the range of 12.6v at idle to 13.6v at 3500 RPM, typically. This presumes the battery has been used within the last few days, and a good digital meter is being used to take the readings.
The "break-even" point, i.e., the RPM where the battery discharge stops and alternator charge starts, is around 2000 RPM (sealed beam headlight, lights on). This can be seen on a voltmeter as a sudden rising reading at this RPM, reading taken across the battery. (An easier way: connect an ammeter across the fuse terminals and remove the fuse, run on the meter instead, read current directly.)
Typically, installing extra loads causes this "break-even" point to rise in RPM. A halogen headlight like my H4 (50/55W) makes "break-even" happen at about 2250 RPM, compared to a sealed-beam headlight (extra 250 RPM). Adding 3-ohm Dyna coils forces this "break-even" point to rise further, to approximately 3100 RPM on my bike. This is equivalent to almost 3 extra amps of current added to the system.
This extra RPM isn't a terrible, immediate problem, if you're not forced to idle in traffic a lot. Lots of idling will drag the system voltage down considerably: also, if your riding style is to shift early and putter along, this problem will become worse. If you ride with revs up, passing 4K on every shift, the voltage will be better. Low batteries die early deaths.
Long term, the 3-ohm coils have shown another issue: heating of the splices in the BLACK wire distribution of the Honda wire harness. This causes them to crystallize their solder and become poor connections, sometimes failing in "soft" ways, like hidden voltage drops. Your method of shunting past these with the extra wire is a good one to prevent this problem. On mine, I have added an extra 20 AWG BLACK and 18 AWG ground to the original harness, carrying extra current around those splices and rejoining the harness in the headlight. This accomplishes the same thing, but without the relay and its additional current draw and complexity, which wire novices tend to avoid or not understand very well.
The Resistor Pack is the easiest way I could think of (and used to be Dyna's recommended way) to adapt the 3-ohm coils to these bikes without causing problems. These coils generate more than 3 times the Honda coils' voltage (if you open the plug gaps to force it up higher), so losing about 15% of it to an inline resistor is not much of an issue, unless you're running nitro or alcohol. In the "old days", at Dyna's direction, the recommended installation was to go to an auto parts store and get the Resistance Wire found in Fords and Chevies (1.1 to 1.8 ohms), coil it up under the battery box, and feed it inline to the coils. Those parts are hard to find today, and made ugly installations, anyway...so I made this version.
Something 'new' has happened at Dyna lately (like in the last 5-8 years) regarding their "installation recommendations" people, I'm not sure what... In years past, my discussions with Dyna engineers (actually, just the one who designed the "S" triggers) yielded the advice that "S" triggers were able to handle 4 amps on a fully-heated (255 degrees) CB750K engine. Since the 3-ohm coils pass (13.6v/3)=4.5 amps, the combination of "S" and 3-ohms were specifically not recommended. The Dyna III was specifically designated to solve this problem, with its separate 6-amp amplifier for the "S" triggers. Today, however, I am hearing that riders are advised that "S" with 3-ohms is an OK combination. Also, however, I know of 3 such systems, installed last winter, that died this year with shorted (i.e., overheated) "S" triggers before summer even arrived. So, it seems the 'old guys' knew better what they were selling...
In any case, the "S" triggers raise the total current consumption for any coils by a little over 80% (averaged over time), because they are ON most of the time. Points are ON (or OFF) for only 50% of the time, while "S" is only OFF for 27% of the time, best case (17% at RPM >4000). This extra "time current" comes straight off of the system charge, showing up as lower operating voltage.
A typical case just wrote me this last week: he has a CB750F with Dyna 3-ohm and an onboard voltmeter. He has long put off long-distance riding on it because of low voltage problems. Last week, he installed a Resistor Pack: he wrote back that the next day his system voltage was over 13.0 volts, and he was finally planning his first long trip, realizing what was long causing the problem. Really, it's just physics...
OK, now it's your turn:
I was thinking of one of the AGM batteries for the Hondaman Special #1 I'm building. (Or, maybe for mine, next time around...). Which one replaces the 12N14A for the CB750K, and does it fit? What is the system voltage at full charge? The "gel-cel" batteries out there can let the system voltage rise over 15 volts when cold, which can then damage the stock Honda (15 volt) rectifiers, which then requires a better rectifier...I don't want to get into that routine with a bike I'm going to sell to someone.