I'm about at the end of what I can diagnose remotely...
Battery charge monitors are very simple in concept[1]:
- Instantaneous power (watts) = voltage x current.
- Power over time is (essentially) measuring instantaneous power rapidly (typically 100 times per second) and adding it up (integration). This is is the watt-hours that your battery is absorbing (if the current is positive) or supplying (if the current is negative).
Current is measured by the measuring the voltage drop across a calibrated resistance (the shunt[2]). Voltage is measured directly. When monitoring a battery, both the current and the battery voltage vary over time. (Trivia: amp-hours is a mostly meaningless number if you don't have the actual voltage when those amps are being retrieved.)
If you have two batteries in parallel but the current measurement shunt is only on one of the battery negative terminals, the battery monitor will only see half the actual current because the two batteries split the load. The sharing will be particularly even when the batteries are new and essentially identical.
What is interesting in a suspicious way is that your battery monitor indicated almost exactly half the watt-hours (-22.3 W-H) that you calculated it should have registered (43 W-H). If you have the batteries connected in parallel (1-2-All switch in "All" or the "+" posts tied together) but the current measuring shunt is only on one of the batteries, then half of the current into and out of the battery bank will be going "behind the Link-1000's back".
If you have both "+" battery posts tied together or you are running the A-B-All switch in the "All" position and then off to the boat DC breaker panel, then you must hard-wire the battery "-" posts together, then wired to the current shunt, and then the other side of the current shut goes to the DC breaker panel.
Referencing my diagram above (post
#26), your symptoms match wiring the current shunt per the "Traditional" diagram with the batteries wired in parallel. If that is the case, you need to change the "-" wiring to wire both batteries to one post of the shunt and the run your boat ground off the "load" side of the shunt ("One Big Bank" style).
[1] In practice there are several unknowns that have to be configured, calibrated, or estimated:
- Nominal battery capacity has to be input. Actual battery capacity can be measured (some effort required). Actual capacity over time will go down; the capacity loss over time must be estimated.
- The efficiency of charging and discharging can be input but is mostly estimated. Some of the electrons that go into the battery get turned into heat and other losses. Not all the chemical reactions turn into electrons; again, some of it gets turned into heat and other losses.
[2] Shunts have extremely low resistance (500A shunt probably is 0.0005 ohms) so they don't waste power.
- Shunts are manufactured very carefully in order to be extremely accurate.
- The shunt sense wires must be connected directly to the shunt (per manufacturer's recommendations) or your current measurement will be inaccurate.
- The signal level on the current sense wires is very small (~0.05v for 100A, ~0.0005v for 1A) so that wire must be routed carefully to avoid noise coupling into the sense wire. It should be a twisted pair to minimize noise coupling; definitely not two random wires run independently.