I like to work with original components as much as possible, tweaking them as needed and only replacing them when creating something from scratch makes it easier to accomplish my goal.

For a long time I've believed that the stock filter bypass may cause problems when a sandwich plate adapter is installed for a remote oil cooler.
Just by feel, it seems that the spring is too soft and allows the valve to begin opening at a fairly low differential pressure.
I think a remote cooler circuit might have enough pressure drop that that the differential exceeds the cracking pressure of the valve, causing the oil to bypass both the filter and the cooler, contributing to a hot running engine in spite of the remote cooler.

Full flow modifications that ensure all the oil goes thru the filter and remote cooler all the time are popular but I'm not really in favor of eliminating the stock filter by tapping into the front of the case or using a pump cover with a fitting.
In the past I've disabled the bypass in the stock mount by replacing the spring and ball with a solid plug.

However, IMO the full flow mod has a significant shortcoming.
A couple years ago we had a cam/lifter failure in a customer's engine and the fine metal debris quickly clogged the filter to the point that the engine lost oil pressure.
When we opened the filter, the element had collapsed from the pump pressure. Someone else had previously disabled the bypass to make a full flow system but
if the bypass had been functional, at least the engine would have maintained oil pressure and the damage might have been less severe.

So I've decided that rather than disabling the bypass, a modification to increase the cracking pressure is more desirable.

My first plan was to measure the stock opening pressure using air, but I discovered that the valve in the filter mount I was testing leaked even when applying air pressure to the back side of the valve.
So rather than making a direct measurement I disassembled a bypass valve and measured the spring force at the installed length using a small digital scale.
By my calculations, the cracking pressure is very close to 5 psi.
At least that's higher than I imagined it might be.

But I still think a remote cooler circuit could have greater than 5psi pressure drop at higher flow rates, especially if small diameter hoses are used.

I thought I would replace the spring with a heavier one, but looking at a listing of available springs, I didn't find anything that was just right.
Instead I made a shim for the stock spring that roughly doubles the cracking pressure.
I made a new press fit retainer for the assembly and re-installed the stock bypass, modified to open at more like 10psi.
That should be enough to keep the valve closed under most operating conditions.
If the filter gets plugged, there will be a pressure drop of 10 psi which will cause the oil light to go on at idle but at least the bearings will still get oil at higher rpms.

Would the additional compression on the factory spring cause early metal fatigue?

I doubt it.

I like the idea as i just installed dual front coolers and did not modify the bypass yet.

I have also thought a lot about rerouting the output of the oil pump to a remote filter/thermostat housing, then right back in the port on the bracket like stock.

I am a little nervous about the oil filter, the 12AN sandwich adapter I used, and the hose weight all hanging on that aluminum filter bracket.
I have wanted something beefier, that has no bypass, and can accommodate 12AN fittings.

I have wanted to just make my own block off plate, but there would be some angles that I would want to be more accurate than I can do with just a drill press.
I even thought of modifying a spare filter bracket, there's just not that much meat on there.

The alteration procedure is simple enough that it can be done with the filter mount on the engine, in the car.
It only requires two small pieces of aluminum turned on a lathe.
The hardest part is probably removing the steel retainer.

I've never seen a sandwich plate assembly come off unexpectedly, but the one we were just working with only engaged about 4 threads on the filter mount - not exactly confidence inspiring.
At least the hoses were secured with adel clamps so not much weight hanging off the sandwich plate.
The engine we just put in this 912E came with a non-thermostat sandwich plate, and the hoses (about 14 ft) were only dash 6.
The simplest/cheapest way to improve the system was to change the sandwich plate to a thermostatic one and change the hoses to dash 10.

I created remote oil cooler packages that take off from the stock cooler mount instead of a sandwich plate,
partly because I like retaining the stock filter mount for the oil filter but I don't think its a great place for a remote cooler take-off.
By removing the added load from that location, the bypass function isn't compromised.