[...] if the cheap safety switch were to fail short and say the exhaust were plugged, it would still seem safe. [...]
The furnace's firmware does a self-check. The sequence is:
read the switch (contacts should be open)
start the draft inducer blower
wait a bit
read the switch (contacts should be closed)
If OK then proceed and do trial for ignition
A stuck open or shorted switch, or bad input on the controller board, or failed blower, blocked tubing, jammed diaphragm etc - would get detected because the firmware is looking for the change. This gives functional-safety coverage for many pieces of the hardware, which is what you need to prove for regulatory.
The switch contacts are likely low spring tension so you can use smaller diaphragms (cheaper), I'm glad Microswitches are being used instead of homebrew switches.
I would have expected it to work something like this. When I was sorting out the problem and it seemed to not care about the initial state of the switch, only that it had closed some time after the blower was active and that it remained closed. I was thinking, that seems a bit unsettling.
If the switch was opened while the igniter was active, it would disable. If I closed the switch fast enough, it would immediately enable the igniter. Once the gas was turned on, if the switch was opened, it would turn off the gas valve and latch it off. Now, that said when I get ready to change that switch, I will have another look to make sure but it seemed like jumping that switch was enough to get it running.
One thing I noticed is if the switch remains open, it appears they will continue to cycle the blower, forever. When the furnace first stopped working, I went to have a look an could hear the blower cycling but the furnace wouldn't start. I grabbed the blower motor and it was very hot to the touch. I'm not sure why they wouldn't just set a code and lock it out after several minutes of retrying.
I'll post once I retest some of these conditions.
********************
Just installed the new part. First I shorted the circuit with the power off, then turned on the furnace. Sure enough, it came up with a different error code. So it does check the initial state. I let it try and start up with the circuit open and indeed it does seem to try to restart, I'm guessing forever.
You can see the difference between the new and old designs. The new switch required the larger spade.
I checked the contact resistance of the old part and it still seems fine. Once I am certain the new switch is isn't going to have a problem (sad that I have no trust in new products), I will cut the old part open and get some good pictures of their design.