If a dichromatic mammal had better night vision than a human, including tapetum lucidum, could they see large (red) and medium (green) light wavelengths but not small (blue)? Would something similar to tritanopia be possible, and if so, how likely? Most mammals that I could find couldn't distinguish red, is that because rods can't pick it up as easily because its wavelengths are large? The animal I'm trying to design can't see uv wavelengths.
Modern mammals lost two different color receptors (cone cell types) to make room for more rods in the eye, to get better night vision. you can easily have the same thing happen.
Mammals lost the UV and green receptor, most importantly without the UV receptor we can't see that end of the spectrum at all, if the blue receptor was lost as well we would lose the color blue entirely.
Just to be clear most mammals can see red, they just cannot distinguish red from orange, yellow and green because they all only activate the one type of cones, whereas we can distinguish them because they activate two different receptors at different levels.
The non-TL;DR version: It all comes down to photoreceptors
Better night vision can be a mix of several things: foveal acuity, tapetum lucidum, and available visual spectrum are the ones that easily come to mind. Foveal acuity could easily be selected for, whether naturally or through selective breeding. The tapetum lucidum assists night vision via retroreflection
The bottom line is that without light in the visual spectrum, you'll need to look elsewhere (pun intended). UV light largely gets absorbed by the atmosphere, but more importantly there won't be nearly as much present at nighttime.
That leaves infrared, which quite a few animals can sense, so evolutionary pressures can select for IR photoreceptors. Since some animals lose their eyes on evolutionary timescales, it's not unreasonable that others might lose a vestigial photoreceptor, causing species-wide tritanopia