where to go ? any ideas ?

I wouldn't recommend the HTB-1100S at all.  I got about two weeks of use out of the first pair I bought before one of them died*.  So I ordered another pair from the same seller with the intention of using the extra as a spare.  They arrived in a few days, but replacing the dead one didn't solve the problem.  The new one would power on but couldn't connect to the other. (I was pretty sure that I connected them properly because I used to work with laboratory optics.)  I tried replacing the converter at the other end of the fiber with the spare, and that yielded about 5 minutes of use before the connection dropped.  After resetting both devices, the connection would drop roughly once per minute. Switching back to the OEM power supply didn't fix it, and the reliably degraded rapidly each time unplugged the fiber cable and reconnected it. (I only did that three times while trying different combinations of the three remaining media converters.)  At this point it won't connect at all regardless of the combination.

It makes me wonder if these really need an SC-APC cable, but it seems more likely that there's a bad retaining clip or something similar.  It's also possible that the cable is the culprit.  Either way I spent $60 and a month waiting for parts, and I have four dead devices and maybe a damaged cable.

(*One of the connectors at the battery came loose momentarily, and I re-set it. A few seconds later I smelled smoke, and after confirming that it was dead I took it apart and found that the optical device inside was fried.  The polarity and voltage were both correct, and there was an LM17 regulator between the battery and the HTB-1100S that should have protected it, at least on the positive terminal.)

Sorry to hear it Karl :( From your description it sounds like a cable problem, since you describe it getting worse over time. I think it was mentioned in this thread earlier that there is a possibility to damage the optical surfaces with a mismatching cable model. Could that be the issue?

I have the yellow cables with the blue connector, and haven't experienced any issues. I even have a patch cable in the middle to increase the length. Unplugged everything many times, and it always connects. Perhaps I need to be more careful.

But the quality for the price of the boxes is of course apparent. One of the connectors is slightly off center for example, and I have to wiggle the cable to get it seated properly. But other than that, perhaps I have just been lucky with my pair. I only use them through my own 5V regulators on both ends.

I only used the SC/PC cable, the same type as yours.

In what way does yours not seat properly?  The dust plugs on mine were a little bit difficult to remove, but my cables were only slightly sticky.

Karl wrote

In what way does yours not seat properly?  The dust plugs on mine were a little bit difficult to remove, but my cables were only slightly sticky.

The receptacle on the receiver has a small inner tube where the actual cable is supposed to go. This tube is slightly bent to the side in one of the media converters. So I can't just push the cable straight in, I have to sort of jiggle it a little to one side so it hits the inner tube properly. If I don't do that, you can't even get the whole connector half way in. So it's very obvious when it happens.

Once everything lines up, the connector slides all the way in without any problems.

Well, fiber is probably still the way to go.  But it would be good to have some comparisons between devices.

I bought a different model from yours in part so that we can compare them.  It sounds as though yours is similar?  You didn't say whether you have opened the case.  If I buy others, I will post info.

A disadvantage of fiber is lack of physical compatibility between the numerous types of cables. Additionally there's a minimum radius they can be bent along, and the ends are hard to remove/replace, so I'm not sure how easy it will be to pass them through existing walls.

Yes I opened mine also. I looked at your picture and as far as I remember it looks very similar to that one.

I even remember seeing plain circuit boards of those models (or similar) on AliExpress. Maybe you could find just a replacement board to get the price down further, if you're still planning to try this stuff.

You're right about the cables. There is a learning curve to those. I also saw cable mating equipment on AliExpress. As far as I could tell it's all done with a simple machine, but probably still a bit more expensive than how you can crimp your own Ethernet cables.

Thanks for the reminder.  Yeah, I could try to replace the boards.  That's probably what I should have done in the first place when the other one got fried.

casper wrote

 I always thought you needed to build an RC filter, and calculate all the values and all of that hocus pocus.

Casper, I did some reading and experimenting with RC filters in the last few months...  They work pretty well for currents up to 1 ampere, but at higher currents a 5-10 Ohm resistor will get really hot and burn up a lot of power.  I think a fuse or circuit breaker is mandatory.  Also, you need a special type of resistor called a "power resistor."  (Like this one:  https://www.digikey.com/product-detail/en/te-connectivity-passive-product/SQPW1010RJ/A137407-ND/2365764)

If the resistors get too hot, you can run more than one in parallel, but it has the side effect of reducing the resistance.

So far I've built four 10-14 Ohm RC filters and hope to have some videos soon.

Another thing I learned recently is that I should have used polypropylene film capacitors rather than polyethylene film capacitors in my filters, because PE film caps work much less well at frequencies above 100 KHz. Polystyrene or ceramic or silver-mica capacitors should work at high frequencies also, but only the ceramic and PP ones are readily available.

Very interesting Karl. Thanks for the update. I will be fun to see what you have done. I bough an enormous industrial EMI filter on eBay, since I found one selling for quite cheap, however I have not had time to test it. It may be a dud, as we have previously noticed that industrial users do not always care about the same frequencies we care about. Not even sure what is inside it, as it is completely sealed in metal. Bigger than a closed fist. I'm sure it filters something though...ha. To be continued...

Have you tried Arco, Idaho? Out there by Craters of the Moon, cell phones wouldn't work (possibly due to natural magnetic disturbances). Not sure if that would help. I guess that's not Australia, though, but maybe Australia has a similar area.

That was a while ago when I was there, though; so, maybe it's full of 5g or similar, now.

This article talks a bit about uncovered areas in Australia: https://cellphones.lovetoknow.com/cell-phone-guides-how-tos/places-that-exist-without-cell-phone-service

Sorry I didn't find anything more helpful.

Karl wrote

I did some reading and experimenting with RC filters in the last few months...  They work pretty well for currents up to 1 ampere, but at higher currents a 5-10 Ohm resistor will get really hot and burn up a lot of power.  I think a fuse or circuit breaker is mandatory.  Also, you need a special type of resistor called a "power resistor."

Since posting that, I learned something the hard way: Almost all of these resistors are "wirewound" - they have a small coil inside. Unfortunately each loop of the coil is so close to the next that an electric field in one can cause the electrons in the next loop to move.  At high frequencies the electrons can 'jiggle' back and forth along the length of the coil rather than going all the way around each loop, and that greatly reduces the resistance :-/

OK, that's the bad news.  Maybe I can just use a few inches on Nichrome wire instead.  "Carbon composition" resistors are supposed to work for frequencies up to a few MHz, but they can't handle the 0.5 Ampere current of the lights I was trying to filter.

I found some articles about which types of resistors work and why, and I can post links if anyone is interested.

casper wrote

Karl, for the regulator I got these. Prepare your wallet for the huge expense :)
They call it a buck regulator, but it's just a regular linear one.

https://www.ebay.com/itm/5PCS-AMS1117-1-2-5V-DC-Jack-Step-down-Buck-Voltage-Convertor-Power-Supply-Module/163081375537 

I bought a few and like them. The pins have the same 0.1"/2.54mm spacing as a solderless breadboard or a computer fan connector, which is handy sometimes. Also the voltage drop is less than an LM317, so you have more freedom about what type of battery to use.

I also blew up one regulator in the process, so it's good that they come in packs of 5.
Protip; don't connect a high amp power supply with plus and minus reversed to the regulator.

I had the same problem. These boards are more sensitive to reverse voltages than the LM317, and once you fry them the light stays on and they stop limiting the output voltage (possibly frying the device connected to them if you don't notice right away).

I went back to using these for devices with terminals that are easy to reverse accidentally:

https://www.ebay.com/itm/LM317-Voltage-Regulator-DC-Linear-Converter-Buck-Step-Down-Power-Supply-Module/183417780567