Experience using wideband O2 sensor with simulated narrowband output to DME

[ed323i]

As I still have some issues with the adaptation values (turns very negative, and have already tried and replaced a lot of things), I was thinking about replacing both O2 sensors on my 1997 M52 323i.
Those are quite expensive alas.. Parts stores easily ask 100 euro a piece.

Then I remembered I already have a Bosch LSU 4.9 that was left over from when I was Megasquirting my ex E28 528i (was great fun).
At 14point7 they sell wideband O2 controllers (with LCD display!) for only $45 (need to add some soldering and wiring though). So, that's cheaper than the OEM Titanium sensors..
So I think about buying another set of the SLC Free 2 controller, which includes an LSU 4.9 sensor ($95 for the set).

So, then I could replace both narrowband sensors with two wideband sensors, see life AFRs (or Lambda-values) in the cabin, and feed the narrowband simulation signal to the DME..

BUT... Alas the MS41 uses the Titania O2 sensor which outputs 0-5V, whereas the SLC Free 2 outputs a 0-1V signal (most common O2 sensor).
So, the SLC Free 2 is probably not an option then, if I want to use the narrowband output to feed to the DME.

I learned someone used another controller (LC-2 I think it was) which was able to send the required 0-1V signal. So that's another option.

But then I did some more research and some people say they advise strongly NOT to replace the stock narrowband O2 sensor with wideband sensors, because there are some issues:
* At warm-up the DME expects to see a certain signal from the narrowband, to know it's still warming up, but many wideband controllers output a simulated narrowband signal that makes the DME think the mixture is too lean, which results in the DME enriching the mixture more and more, up to 25%!
* If the wideband controller crashes (which seems to happen from time to time), it will either revert to a 1V or 0V signal and it will cause the DME to enrich the mixture 25% or lean it 25%.. Problematic.

So, that's all theory. So, my question:
Has anybody tried this: Using wideband O2 sensors instead of the Titania narrowband sensors (M52B25, M52B28), and then sending the simulated narrowband signal to the DME? Does this work well, or is it indeed problematic?

Here are some threads in which people show problems with using this setup:
https://www.s2forum.com/forum/techni...-into-motronic
https://forum.hptuners.com/showthrea...ow-Band-How-To
http://www.zeitronix.com/questions/NBpoint.shtml
https://pcmhacking.net/forums/viewto...=3289&start=20

And here's a guy who got it to work on an MS41 DME:
"It seems to work for me. I have it set to output 0.6v at 14.55:1 and 4.7v for 14.65:1" .. At least with the LC-2 controller..