I don't understand how camera (talking hardware) quality can be evaluated from a pcb picture where pcb is not perpendicular to camera axis, in low and warm light and mostly white background around it.

Light
The first crap to avoid to make/test pictures is LED. The color rendering is terrible, (exception to professional models).
Il misses wavelengths and exesses in other ones.

Camera
There are several distinct things we can test in a camera:
-luminosity (loss of light into the lens)
-sensibility (sensor asa)
-sharpness
-exposre measurement
-color rendering
-vignetting (in color, contrast, brightness)
-deformation
-sensibility to flairs
...

So what are we evaluating with these pcb pics?

First picture:
It looks to me that the first one is quiet sharp on the lower part but seems overall unsharp for two possible reasons:
-moved because the exposure time had to be long because of the miss of light, so a tiny move makes it unsharp.
-Out of focus as the focus seems to be near of the camera, the rest of the subject going out of the depth of field because more far from the cam. The more you are close, the less depth of field (sharpness zone size) you have.

Second picture:
Totally moved or out of focus.

Both:
Both seem a bit underexposed. Depending of how exposure is calculated, a big piece of white in the picture tends to make the camera underexpose to compensate the big amount of white. Resulting in an underexposition of the rest of the subject.
Color balance is strange. The depron foam should look white but it seems white-orange. Maybe the leds or wb on "day".

A camera quality must be evaluated on a whole bunch of several different and widely variated tests. Not only (but why not also) fast fast fast under a bad, a trap, light condition.

I'm not saying the camera is good or bad but I think it'd deserve a more deep test before concluding anything.
It's new, give it a chance!
The crew behind the camera also deserve patience, rigorous tests and understanding of their choices.

Also, what do we want? Honest rendering or totally fake-and-always-beautyfull-more-than-real standardized rendering like most devices seem to do nowadays?

About the n900 cam:
I use it a lot and "developped" pics into darktable. Most of the time starting from raw format. I love it, but it almost doesn't deserve the name of camera.
Is is a real... hmm.. not so perfect piece of hardware. Just look at the huge awfull vignetting almost uncorrectable because not only brightness changes but also big color and contrast shift, the more you go to the broders of pic.

Once again. Until I'll see raw neutral interpretation of the images, compared to the same scene with another cam in raw, I'll not make my mind on this cam.

I'd love dogs to leave rather that kind of things in my street!