I’ve been off the grid for over a week. The banks would generally run down to what the SG thought was low 60s, but the amp hr counter thought was low 70s. Recharging took various forms, from running the generator to intermittent engine use. Twice I ran the generator until Absorb acceptance amps got as low as 0.5%C. I let the bank rest, for about 30 mins, after the generator was shut down, and manually reset the SG to 95%. Despite this forced learning, it would only read 85% on subsequent charges back to Absorb=0.5%C. It just seems to like that level, as noted above.
Odd thing happened yesterday. I had to run the engine for 4 hours, underway. I noted that acceptance amps declined to well under the 0.5%C level, more in the range of 0.1%. I do not have a smart alternator, so it essentially maintains a permanent Absorb level charge. The SG, after those 4 hours, now indicated 93%. Still, it was clearly 100%, but I found it interesting that the SG would only recognize charge over 85%, if Absorb went far past where it should.
Like you, I do not have a "smart" regulator on my alternator (Leece-Neville 8MR series). I have 2 x 110 AH FLA wet cells connected in parralel as my house bank. I have been playing with a 60 Watt solar panel (actually 4 x 15 Watt panels), and the MPPT controler that was included with the panel to "charge" my house bank. My observation is that the solar panel does not bring the house bank voltage up to 14.x VDC when I connect it. Instead, the voltage starts at 12.xVDC, and as the battery slowly charges from the solar panels, the voltage displayed on the SG slowly raises. As a result the SG does not seem to recognize this as charging the battery at all! The SOC stays right where it was when I connected the solar panel to the batteries. However, if I start my motor, the 90A alternator quickly jumps the voltage over 14 VDC, and the SG SOC starts to rise.
Interestingly, the SOC reported by the $30 BlueTooth battery monitor (BTBM) does
recognize the increase in SOC when the solar panel is charging the batteries. Frequently the SG will report 78% SOC and the BTBM will report 80%. While they don't always match the reported SOC, they never seem to be off (in my limited experience) by more than 10%, and are usually within 5%.
Another observation is that the BTBM seems to recognize anything above 13.2-ish (I am not really sure of this number) VDC as charging. Therefore if my batteries are fully charged, such as when I first shut my engine off, this el-cheapo battery monitor will identify the surface charge as charging until the surface charge drops. The BTBM does NOT report SOC while the batteries are charging - this is something that the SG does.
My working hypothesis is that the SG looks at the battery bank voltage only. If the battery bank voltage is over some seceret threshold the SG believes that the battery bank is being charged. The SG then quantifies the TIME above that threshold, and begins to increase the reported SOC as time elapses. If the voltage falls below another threshold then the SG starts to decrease the reported SOC. The further below that threshold the faster the SOC decreases.
Unlike you I am not trying to figure exactly how the SG will report with some exoctic battery chemistry. My goal is simply to keep my house bank viable for as long as possible. For $30 the BlueTooth device seems to meet that need.