I'm sharing here my successful conversion of my Norcold 1200 to "residential" freon compressor operation. When the original cooling unit failed, I decided to explore converting the fridge to conventional freon compressor cooling like residential type refrigerators use. There is at least one manufacturer of a freon-based cooling unit that is sized and physically configured to replace the ammonia absorption cooling units used in Norcold and Dometic RV refrigerators. These units draw less than 100W when running, and since the compressor draws power only when cooling, the actual power used over a 24 hr period will typically be much less than 1KWhr, so you can still dry camp using the house batteries.
The caveat is that the control electronics (power board) of the refrigerator is not designed to work with a compressor. It was designed to work with a vapor absorption system and control the heat sources (propane flame or electrical heating elements). The conversion of the RV refrigerator from a vapor absorption system to a freon compressor system requires the freon compressor to be hooked to the terminals that used to provide power to the heating elements of the vapor absorption system boiler. The instructions that came from the supplier of the conversion system state that the refrigerator's warning beeper must be disabled and the users must simply ignore the flashing error codes on the fridge front panel that indicates the heater elements have failed. (It will give this error continuously because the freon compressor draws only 1/10th the power that the heater elements did.) I deemed this ham-handed approach (disabling the warning electronics) unacceptable! Therefore I decided to re-engineer the fridge control electronics so that it accepts the much smaller power (load current) of the compressor as a valid operating level.
My goal, therefore, was to raise the sensitivity of the AC Heater current sensing circuitry so that it would accept the 100W load of the small freon compressor in place of the 1500W boiler heater, and not have the control logic give an error/fault indication of Open Load/Sensor ("Sr OP" was the fault displayed by my particular board).
Having no schematic for the control electronics, and having been told by the manufacturer of the control board that the schematics and circuitry are proprietary and cannot be shared, I began the task of reverse engineering the control electronics. I soon determined, by tracing out the pathways on the printed circuit board, that the heater load power was sensed via a 500:1 ratio current transformer.
"Aha!", I thought, "All I have to do is increase the value of the 'burden resistor' on the output of the current transformer to achieve the desired increase in sensitivity. (Since the output of current transformers behave essentially like constant-current sources, their secondaries are typically loaded by a burden resistor to achieve the desired volts-per-amp output (i.e. turns-ratio x burden resistance = volts per amp). However, and much to my surprise, I discovered that the burden resistor footprint on the circuit board was not populated on the board (no part was soldered thereon)! Without a burden resistance the secondary of a current transformer will try to rise (trying to push current through an infinite impedance) until it reaches the saturation voltage of the coil! (Not the usual practice, to be sure.) I did note that current transformer's secondary was provided some loading through a diode and capacitor feeding a voltage divider (comprising, I believe, a passive integrator/filter circuit to translate the signal to an appropriate trigger level).
Not having the option of increasing the burden resistance (since a missing resistor is the same as a resistor with infinite resistance) to achieve the increased Volt-per-amp sensitivity, I was still left with the option of altering the turns ratio on the current transformer to achieve the desired sensitivity. This necessitated unsoldering the transformer from the board, taking it apart and replacing its one-turn primary with six turns of 18AWG magnet wire. This lowered the turns ratio to ~83 and increased the sensitivity, allowing me to populate the burden resistor footprint with a 5.6K burden resistor to achieve a sensitivity of ~2.5V across the burden resistor per amp of primary current.
The board now functions properly with the compressor as its load, and still has all the load-sensing functionality and other features intact! The refrigerator cools more quickly and efficiently than ever before, and the ice maker produces much more ice than we could ever use.