If you want to go a little further with a fan replacement mod, here is what I ended up doing with my SDS2104X plus.
As others have noted, the stock fan has a very unpleasant sound for me, with a "ticking" type sound that really grates on my nerves. It was always a notable relief whenever I turned off the scope in my otherwise quiet lab.
I bought a Noctua NF-A9 PWM fan that runs at 2000RPM @12V. The stock fan runs at 1600RPM, and was actually quieter than the Noctua at 2000RPM. I first used the Low-Noise-Adapter that came with the Noctua to lower the speed to 1600RPM, and the noise was then about equivalent to my ears. However, the characteristics of the noise were much better. More white noise, and no "ticking".
Not good enough though, and not really worth the swap in my mind, so I went further.
I wanted full PWM control, but the scope hardware does not support that. I saw other posts of folks using a small PWM controllers, so I picked one of those, and dived in.
https://www.amazon.com/dp/B07L1DZMKS?psc=1&ref=ppx_yo2ov_dt_b_product_detailsThis little gadget allows you to set the baseline duty-cycle, and a linear extrapolation between two start/end temperatures. I had shell access to the scope (via telnet), and was able to collect temperature data from two internal sensors, and the NCT thermistor that came with the PWM controller.
/sys/devices/soc0/amba/e0004000.i2c/i2c-0/0-0048/hwmon/hwmon0/temp1_input
/sys/devices/soc0/amba/f8007100.adc/iio:device0/in_temp0_raw
I can't speak to the absolute interpretation of these sensor readings, but I do believe the relative values are viable, and they track with the reported temperatures in the "Developer" menu for temperatures.
I first baselined the stock fan as it warmed up in my 20C lab, and ran some FFTs, and other processing to warm it up a bit more. I then swapped in the Noctua fan with LNA (no PWM), to try an apples to apples comparison. At steady state, the Noctua fan kept things cooler by 2-3 degrees C.
I then installed the PWM fan controller, and started playing around with various settings, and monitoring the temperatures. I found what I figured was the optimal settings, that didn't allow the temps to reach more than ~5 degrees C over the stock setup.
It was quieter, but not really by much. Still a lot noisier than I wanted. While I was taking things apart and putting them together, I discovered that the majority of the noise was really coming from the air whistling through the Aluminum intake and plastic exhaust vents that the fan was sandwiched between. At that point, I decided to kiss my warranty goodby.
I took off the aluminum cover, cut-out the grid, and also cutout the plastic exhaust grid on the back of the scope
A little more tweaking of the PWM settings, and I was finally happy. It really is quiet now. The majority of the residual noise is from the various intake vents around the perimeter of the scope housing. Not willing to do anything about that, and it's not bothering me at all. I no longer have that relieved feeling when I turn it off, and barely notice when it is on. The steady state temperatures are about 2-3 degrees C higher than the stock fan in a 20-21C room. The PWM runs at about 55% duty-cycle most of the time, and runs up to 65% if it's really working. Still has lots of head room if the ambient temperatures get hotter. Given that the scope is rated to 40C ambient, I running a little warmer is just fine.
Below are some pictures of the mod. Notice where the thermistor is mounted on a small piece of the original aluminum grid. I have some plots of the temperature data if someone is interested.
Craig