Sir bobbintb . . . . .
I must say that model almost has all of the bells and whistles, and I NEVER have seen a manufacture be as generouis with ALL of the tech info.
I initially am just using the simplest block diagram / photo.
It also appears to be the main power relay for the magnetron that has time/ service eroded and pitted contacts and its resultant "iffy" contact action. Sometimes not closing to good contact connection and sometimes making a temporary "spot" welding of the contacts. . . .ergo a continual running then..
Either of those fault modes could be analyzed by simply monitoring the coil of that relay to confirm if there is never being contact continuity on the top two push on contacts without 12VDC relay coil power presence . . . .OR . . . . .an open contacts circuit with 12VDC relay coil power.
Consulting the full schematic for further analysis, reveals a door locks tandem arrangement of switches that either inhibits operation of the magnetron power or in its other mode, blows the main fuse. So . . . .I guess . . .that that fuse has never blown ?
Can you view your board, or you may remember from a previous viewing . . . .if there is just the sole presence of relay R202 ? . . . . and not the additional RY203.
Also note the SM's error on the text balloon . . . which I corrected . . . that had the relays I.D.'s being reversed, compared to the boards symbolization.
AND did 'ja notice, that its absolute " RELAY CITY" over to the right area !
I would suggest taking two insulated wires from the the solder connections of the relay coil connection to remote outside the unit to be able to use a meter for monitoring 12VDC presence to that coil.
Then, if you can repeatedly do turn on short test cycles and catch the unit running when it SHOULD BE OFF , but with no 12VDC relay coil voltage presence . . . the relay contacts are at fault.
The other situation is the magnetron unit not running when turned on for a cook cycle.
In this case you could perform a DOUBLE test by having one the remote wires coming out while the other is routed a bit differently.
You pull off the two switched wires that plug into the top of the relay.
You now have a power dormant set of relay contacts, one of which will receive the OTHER remoted 12vdc wire into one vacated top relay contact and a new wire connects to the other relay contact.
You now effectively have one remoted wire remaining as it always was for testing , the other wire connected to one contact at the top of the relay, the newly added wire goes to the other contact and down to the monitoring position as the other needed wire.
There is a shortcoming of using either an analog meter or DVM in monitoring in their response time. (Vice the aural rsponse of the DVM in its diode / continuity test
IF being an available option ? )
Should you use a LED with a series 470 current limiting resistor . . . . with the pair installed with correct polarity for the incoming 12VDC voltage.
A visual monitoring of that lit LED will then be able to perceive a relays contact power disruption for as short of a duration as ~100 milliseconds.
Shorter time than that is a fuzzy situation . . . . . in the respect of the human eyes persistance of vision kicking in.
This preparation all leads into the final relay testing aspect of having the top cover access open to be able to reach the RY202 relay case .
You then key in a good known cook cycle , but don't expect the MAGNETRON to respond BUT DO know that R202's coil should be having a 12VDC presence and the LED indicator is lit .
You then use the plastic handle of a medium size screwdriver *** to tap the side of the RY202 relay case in the same directional plane as the nearby large BLUE filter capacitor pair are.
(*** Using either a conventional Phillips or Standard slotted blade . . . . . .but I would be a bit wary of a new fangled Torx. )
Proceed thru a moderate tap to a harder tap to a FIRM TAP, to see if relay contact is disrupted by LED light dropout for that long /or / short of a duration.
Now if you posess the mechanical aptitude and technical acumen, go
DOITTOIT !
Or . . . . . . ask for any clarifcations . . . .
Thaaaaaaaaaaaasssit . . . . . .
Le TECHNO-NIBBLE . . . . . .

73's de Edd . . . . . .
Don't EVER find yourself ending up in getting between a dog and a fireplug / utility pole.
(Hmmmmmmmmm ?. . . . . but then, there ARE times, when a dog thinks that YOU are a fireplug / utility pole ! )