relays for electric fan in truck

Good point. I tend to think of the "not uncommon above 200V" spikes that can occur (hence prefer next 400V rating) and think of that spike superimposed as a +ve going spike, hence exceeding rev breakdown voltage. But that may not be valid. (Ah - memories!) I know in practice that other things are likely to absorb such +ve transients whereas even a smallish -ve voltage transients can do massive damage (say -0.3V for chips or -0.6V for most other circuits, where minus means below 0V = ground). But usually spikes >50V are rare. (But I know 200 Amp spikes with halogen bulbs are not rare!) Now, will a forward biased spike that exceeds the diodes PIV rating arc the diode? And .....   Ah - memories!

Well, if it's a forward biased spike, it really can't reach the PIV, because it'll start conducting to clamp it down right away. If it's a huge current surge that has enough energy to exceed the peak current limits of the diode, then it might be trouble. But that would be exceeding the current limits, and still not related to the peak inverse voltage. If the diode is used for AC rectification of 120V, it'll be important to make sure it can hold off the large reverse voltages. But that's somewhat rare for 12V automotive stuff.

Ah - so you can never exceed the PIV in a forward direction? Hmmm? (Even Hi-Z circuits & chips?) And whilst I definitely recall 200A spikes, I only recall 200V spikes from reading, though I think I have seen 200V transients on CRO pics/dumps. But I won't vouch that the 200V is not a misprint or misunderstanding of 200A, or based on some normalised 1-Ohm cable resistance. But I agree that a 50V rating should be enough. But 50V is now difficult to find here - IN4004 is the minimum. And maybe they specify >200V in expectation of electric welding repairs LOL! (In case they isolate the batt- instead of the batt+ ....)

Don't worry - it's just banter. In my case on things I haven't looked at for ages..... and pointing out that forward biasing only clamps if there is a low-resistance path to a power rail (ie, if a coil turns on, its +ve 50V or 250V spike is likely to damage IC/chip inputs if they are unpowered (ie - not internal clamping to the +ve supply)). Not that a relay quenching diode is forward biased in that situation... Forget it - late night speculation....

Not for spike suppression - you had that correct in your diagram on the last page - ie, diode between #85 & #86 with the line (Kathode) towards the more positive - namely #86 by convention. What you have drawn in the bottom of the diagram above is that current will only flow from #85 to the ECU (the current flows out thru the line end of the diode). Hence for the relay to switch, #86 is positive (say +5V or +12V etc) and the ECU output connect to GND to turn the relay on. That is a very common switching method - aka GND switching or Open-Collector output. It is common because it doesn't matter if the relay is 12V & ECU 5V etc (an Open-Collector output normally handles 20V or more). That diode doesn't do anything to protect the ECU - I think you are getting confused with the spike suppression or quenching diode between #85 & #86.

How would it change if I reversed the diode?  I want to protect the engine computer from anything that might accidentally go back up the line.

Then it only allows the ECU to supply current to the relay. The 85-86 reverse biased diode quenches relay-coil induced spikes in case the controlling circuit does not have it. But if an ECU is designed to control a relay, the ECU should have that protection anyhow. The in-line (series) diode us unlikely to protect the ECU - it will not stop over-current etc. To be certain, you need to find out what your ECU is intending to switch, or whether it is open-collector (grounding) or supplies +ve current (of so, at +5V or +12V etc?), and what current it can source or sink. I'm not sure if you have supplied such info... I can;t recall - Id have to reread the lot. But do you have any reason to think the ECU cannot supply the load (relay) you intended? (Ah - it's for a fan...)

I guess it is open collector since it provides a ground.

I have 3 different sources of a ground that will trip the relay: engine computer, a manual on switch, and the AC trigger.