Starting with your cylinders around arbitrary minor axes, these are largely well done. You've shown considerable skill both with drawing well controlled, evenly shaped and confidently executed ellipses, and in keeping their alignments quite close to your intended target. For this section I have just one minor issue to point out.

Our cylinders, due to perspective, undergo two shifts. You're aware of both of these and you demonstrate them throughout your cylinders. One is a shift of scale, where the far end gets smaller than the closer end. Sometimes it does so quite dramatically, sometimes it shifts only very slightly. The other shift is one of degree, where the farther end gets wider than the end closer. Similarly, sometimes it's by a little, sometimes it's by a lot.

The reason I'm explaining this is because there is a relationship between the amount to which both of these shifts occur - you shouldn't ever be seeing situations where the scale shifts by a lot, but the degree only shifts by a little, or vice versa. Most of the time you get this right, but there are a few cases, like the bottom right and middle-left of this page where you start getting slightly into the territory of shifting too much in degree but not enough in scale. It really is quite minor in your work, but it does come up here and there, so I figured I'd point it out.

Moving onto your cylinders in boxes, there's one key issue that jumps out at me immediately: all of your boxes appear to be drawn with extremely shallow perspective. It appears as though you're actively striving to draw the boxes such that their sets of parallel lines converge as little as possible. This is not an issue that was present in your box challenge - there you respected the presence of your vanishing points (whether they were closeby or very far away), rather than seeking to get rid of them altogether.

Now, despite this you have largely still performed the exercise otherwise correctly, and as such have moved closer to its intended goal. That is, this exercise is actually largely about improving how you draw boxes by specifically improving your ability to maintain proportionally square pairs of faces. Since the additional tests we perform on the cylinder's own major lines (minor axis, contact points, etc.) require the ends to represent circles in 3D space, then it would ostensibly also require the planes that contain those circles to be squares in 3D space. So by testing these additional lines, we gradually improve the student's ability to estimate those proportions, a skill that will be quite useful as we move forwards.

Now, I do want to make sure you get this box thing sorted out, so with a bit of a change in process, I'm going to ask you to draw 10 boxes. You don't have to put cylinders inside of them (you can if you want to), but do be sure to apply the usual box challenge line extensions.