Yes, it is the converter. Yesterday I finally had time to do some proper diagnostics.
- 120 volt AC power off (not even plugged in)
- battery negative lead connected
- battery positive lead disconnected
- all fuses pulled out except main
- all devices switched off
- ammeter probe connected to oscilloscope measuring current in battery + lead.
- voltmeter probe on battery + cable
- for test purposes the main fuse was replaced with a twelve volt, 30 amp circuit breaker
At the instant the battery is connected, there is a current spike peaking around 50 amps, lasting about 200 milliseconds and dropping to about 20 milliamps. This is what is blowing the fuse. If I wait two minutes and repeat the test the results are the same. If I repeat the test within a few seconds, the results are similar, except the spike is smaller, the faster I repeat the test, the smaller the spikes become.
The voltmeter showed voltage on the plus cable jump to 13.1 volts when connected, and when the battery was disconnected, voltage on the + cable dropped to zero over the space of about 20 seconds.
Anyone familiar with power supply electronics will recognize this as the battery trying to charge filter capacitors in the converter. A fully discharged filter capacitor (brand new, good ones, in perfect condition) normally act as a short circuit the instant they are connected to a power source. The current drawn by the capacitor diminishes as it charges up to the applied voltage. The biggest spikes come when the capacitors have had some time to discharge. Smaller spikes are generated when the capacitors are still partly charged.
On the voltmeter end, the capacitors will briefly store a charge, until it is dissipated by an external load (something else in the charging circuit) or by their own internal leakage. This is why the voltmeter shows voltage on the + cable for some time after the battery is disconnected
There should be a diode in the converter to isolate the battery from the charging circuit and filter capacitors when the charging circuit is not powered up, and that diode would prevent the sudden spike of current. Either WFCO cheaped out and omitted the diode (in which case I would think other people would have experienced this as well) or the diode is shorted as Boondocker suggested. Or WFCO screwed up on the main fuse, labelling it wrong and installing the wrong size, as Grampa Jim suggested. Maybe the fuse should have been 30 amp which would tolerate the spike better. At any rate the converter either has a design issue or an outright component failure.
Knowing that, and with the big trip of the year 3 weeks away, the temporary fix is to use a pair of fuse taps to connect a 100 ohm resistor across the main fuse terminals. Pull the fuse out, connect the battery, let it charge the capacitors through the 100 ohm resistor (which would limit current to at most a measly 120 milliamps). When the capacitors are partly charged, then install the fuse. Note that this won't fully charge the capacitors because the stereo memory and detectors are trying to draw current too. But it should get the spike down to a level where the 20 amp fuse is happy.
Long term fix is to get rid of the junk and put in an industrial 12 volt 50 amp supply, with separate chargers (shore power and solar) and automatic switchover.
2011 Grand Caravan