Apparently a Chinese transformer is likely to fail prematurely. I have a china made lamp that had the power supply fail on it. It was a complex circuit with unmarked components but after googling for the voltage controller IC, I saw a "recommended configuration" in the datasheet that looked exactly like the circuits in the PS. So now I knew how the components were supposed to behave. It turned out that it was the transformer inside the circuit that had failed. I could not find a replacement transformer anywhere. So I bought instead a new MeanWell PS and fed the voltage from it directly into what I believe must be some kind of constant current driver circuit. I knew from the datasheet that the voltage is supposed to be 12V. That was a lot cheaper than if I would have to buy the components. It is still working, and if the PS should fail, a small 12V replacement will not be hard to find.
Now this LED driver is a bit trickier because, 1. space is rather limited unless I'd let the PS be outside the lamp and 2. the current PS is sealed with some kind of glue that looks impossible to take a part without destroying it.
I don't know much how LEDs operates, but it seems that voltage doesn't matter to the LED. What does matter is the current. That makes me look upon electricity and voltage from a slightly different perspective. But that confuses me a bit as I believe that even LEDs are sensitive to ESDs. So I'm not entirely sure how it all fits into the equation.
But I guess what matters here is that the amperage is 900mA ± 5%. For that to be achieved, 38.7V was measured over the circuit. That would make me assume that the impedance, or resistance of the load is 43\({{\Omega}}\).
Now, this is a bit tricky because since the driver maintains a constant current and if that current is too strong and I choose to drop the voltage with say a series resistance circuit, then the power supply will try to drive up the voltage to compensate for that load. So I guess that I will have to match the amperage.
When thinking about it, maybe a parallel resistance may help drop the current. Kirchhoff says so. I found for example an "LED power supply 50W 6-48V 1050mA ; MeanWell, LPC-60-1050". It most likely is too big to fit but let's use it for calculational purposes. We know that the voltage should be 38.7V, at this voltage we want to deduct 150mA from the circuit. So the parallel resistor should be 258\({{\Omega}}\) and be able to handle a power of at least 0.6W. A resistance like that may be hard to achieve, so then we have the question of what amperages can the LEDs tolerate? Perhaps it is better to hit a lower threshold than a higher. I guess that the more current you drive through the LEDs, the lower the life span.