Best under-desk cardio for bad knees

This page exists because our own walking-pad page on the same subject runs out of evidence. Nobody in the walking-pad category publishes a cushioning measurement — not a deck deflection figure, not a durometer rating, nothing you could use to compare two decks. So that page ends by naming the one difference that is structural rather than claimed: on a seated machine your feet never leave the pedals, so there is no footfall at all. Then it handed readers to a ranking built on desk clearance, which is the wrong criterion for someone whose constraint is a joint. This is the right one.

1. An ellipse or a circle.These are two different mechanisms and the difference is geometric, not marketing. A crank carries the foot around a circle, so the knee is lifted by the full crank diameter at the top of every turn. An elliptical linkage flattens that path, so the foot travels further forward and less upward and the knee stays lower through the stroke. That much is mechanical fact. What nobody will tell you is by how much — stride length, crank diameter and step height are the three figures from which range of motion could be derived, and not one maker we track publishes any of them, from companies that will state a product weight to a tenth of a pound.

2. How low the machine is. This is the figure most buyers get wrong on a knee purchase, because it sounds like a storage question. It is not. Your desk height is fixed. Every inch of machine under it raises your knee by an inch inside a gap that did not change, which means a tall machine forces a more bent knee for the whole session. The clearance arithmetic, with every published height. The published spread here is small and it matters: 9.5 in for the Cubii Move, 10 in for the JR1 and the DeskCycle, 11.38 in for the Stamina.

3. Whether the foot is held on. On the upstroke of a pedal circle the pedal is coming back up at you, and if nothing retains the foot you either push it down or lift the leg to follow it. One machine in this category publishes a solution: VIVO states adjustable pedal straps for the DeskCycle, re-read on 2026-10-17. That is the only published foot-retention feature in the category.

4. How hard the hardest setting is.A level count is not a magnitude — it tells you how finely a range is divided, not where the range ends. Exactly one maker here states where it ends: VIVO publishes 8 settings up to 39 lb of resistance for the DeskCycle. Every published resistance figure in the category, ranked on magnitude rather than on count. For a knee, the useful end of that range is usually the bottom, and the bottom is exactly what nobody publishes.

Three of the six publish no dimensions whatsoever, which means you cannot check them against the gap under your desk before they arrive, let alone against a knee. That is the single most common finding on this site and it is not a research gap on our side: we opened the maker’s own page in every case and recorded what was absent.

One number in that table deserves to be read sceptically rather than trusted. The Cubii JR1’s 150 lbmaximum user weight comes from its retail listing, and Cubii has never published a maximum user weight for any model in its own range — re-checked on 2026-10-17. It is the lowest limit of anything on this site. We are not going to tell you it is wrong, because we cannot verify a different figure from a primary source; we are telling you where it came from.

We have not used any of these machines, we have measured no joint load, and we are not clinicians. There is no EMG study of any machine on this page, no force-plate data, and no published stride length to derive range of motion from. Every figure above is the maker’s own or its retailer’s, read from the page named in the sources on the date given. The full method.

The one clinical source we will quote is about activities rather than about machines. Harvard Health, reporting a BMJ analysis of 217 randomized trials covering 15,684 participants, states that lower-impact aerobic exercise — walking, cycling and swimming — came out on top for relieving pain and improving function in knee osteoarthritis, and notes that the strength of the evidence was of moderate certainty. That sentence is about cycling as an activity. It is not a statement about the DeskCycle, and we are not going to present it as one.

If your knees hurt, the useful next step is a clinician who can look at the knee. A specification table cannot do that, and the ranking above is a ranking of published documentation, not of outcomes.