Last modified:
Scooba has been provided with a swivel-caster that functions very much like a furniture caster-wheel, at least with regard to accommodating direction changes. Since Scooba is programmed to not go backward more than a very short distance, the designers had no need to allow the full 360-degrees of swivel motion that most caster are permitted. Similar to furniture-fitted casters, Scooba's wheel is retained by a friction detent working in a grooved swivel-shaft. Figure 1 illustrates that shaft-groove, and shows the wheel pulled out of the robot, so it can be maintained.
Figure 1. Swivel-Caster Easily Pulls Out for Maintenance
See those vertical ribs on the forward wall of the wheel-well; they are limit-stops that limit swivel-action to slightly less than plus and minus 90-degrees.
Once we look more closely at this Front Wheel Assembly, it proves to be a cleverly designed mechanism which does more than just support Scooba's front. An interior Push-Rod does double duty to both actuate the safety-wheel-drop switch for this wheel, and to actuate a sensor that reports each revolution of the wheel to the robot's controller -- at least, that is what it appears to be doing!
FYI & FWIW, the outside diameter of each tire (Fig.-3 gives a good view of the pair) is 28-mm.
Notice the knee-action wheel-suspension indicated in Figure 2. Well... not really a 'knee-action-suspension' -- Scooba has no plans to jump curbs or smooth out pot-holes in the road! What we see here, in Figure 2-a, is the hinged wheel-shroud, being pushed down by the built-in wheel-drop spring. Figure 2-b illustrates the normal running position of the wheel-shroud when Scooba's front-end weight forces the wheel UP.
Figure 2-a. Wheel Dropped Figure 2-b. Wheel Elevated
Pay attention to the difference, shown in 2-a and 2-b, of the amount of push-rod stick-out. The difference (approx. 6-mm) is the dynamic range due only to the wheel-drop action.
As mentioned a moment ago, revolutions sensing is being done. To accomplish that, the push-rod's vertical position is modulated as the wheel turns! Figure 3 give a view of the wheel, as seen by a bug on the floor, that makes it possible to see how that modulation comes about.
Figure 3. Crank-Pin Between Wheel-Halves Reciprocates the Push-Rod
This view makes it clear that the front-wheel is actually a split-wheel with something in the gap! Spanning that gap you should see what looks like a rod, or post -- nearly on center. That is a "crank-pin" -- its axis is offset from the wheel's axis by a greater amount than indicated in the picture. Wrapped around the crank-pin, there is what looks like a piece of wire. That's the lower end of the Push-Rod! It is a wire-form (i.e., a piece of single-strand, steel-wire that is bent up in a specific manner) that is straight, for the most part, but has a looped-hook at its lower end that wraps around the crank-pin. So, again we see a bit of clever engineering, that transforms the rotary motion of the crank-pin into linear, reciprocating motion at very little cost. This is an adaptation of the classical "Scotch Yoke" -- which harks back to the days of steam-engines.
The next figure illustrates the dynamic range of the push-rod as the wheel rotates (the wheel is 'down'):
Figure 4-a. Push-Rod Fully Lowered Figure 4-b. Push-Rod Fully Raised
Of course, when the wheel is UP, the rod would protrude farther, but the same peak-to-peak modulation (4-mm, P-P, has been measured) would be superimposed on that static stick-out.
We would now like to talk about the pair (presumably) of sensors which are mounted above the wheel, and which transduce either a wheel-drop condition, or a repetitive up / down motion of the push-rod, into pips. Well, yes, we would like to, but, since that discussion is taken care of here, we will just flash some images of the normally hidden apparatus which the removable wheel plugs into. Figure 5 shows the front and rear sides of the structure above the wheel, and reveals only the rear face of the PWB
Figure 5-a. Push-Rod Actuates 'Parts' Behind Black Cap Figure 5-b. 'Parts' Under Cap Mount on this PWB
The front-side of this apparatus is partially enclosed by the black-plastic cap seen in the figure. Considered removal of the cap was thwarted by the cap being held in place via three-each, fusion-staked studs, so it was left alone -- for another time, when egress may be more driven.
There is one other mechanical feature which can be pointed out, and that is the detent-spring that engages the wheel's swivel-shank. See 5-b. Notice the spring-form that is centered in the square cut-out at the lower edge of the PWB; that is the detent's spring.
Should any owner see that Scooba's front-wheel, its axles, or its crank-pin is getting hair-wrapped to the extent that the wheel no longer functions correctly, the wheel can be dismounted for very fast clean-up; which means the tweezers may remain wherever they are stored. Roger perfected the process and provided a well-written procedure describing how it may be done totally without tools. Strong finger-nails are necessary.
The essence of the process is to use fingers and nails to spread the side-walls of the wheel-shroud, while also applying a dislodging force to the crankshaft by pushing the Push-Rod into its concentric swivel-shaft. One axle-end is jockeyed out of its hole, then the opposite end is walked out of its hole; now, with the two ends free of radial restraint, the wheel / crankshaft can be pushed outward while its shaft-ends get guided along a pair of installation channels (end-views of those channels can be seen in Figure 3, near finger tips).
Re-assembly follows a reversal of the above process. Pay attention that the U-bend of the Push-Rod is to be near the normally trailing edge of the shroud / fender.
Users have reported the first evidence of a Scooba part wearing excessively; fortunately, the item IS replaceable, and iRobot is issuing replacement wheel-assemblies as well as stuffing a spare wheel-assembly in with Scoobas being shipped. The worn areas are in the shroud.
Here is some insight into the nature and source of the wear problem. The wheel's steel axle uses cast-in holes in the thin walls of the shroud (which functions as a combination fender and wheel-fork) as non-lubricated sleeve-bearings. Someone might point out that Roomba Discovery models use a similar axle-bearing arrangement, which is true, but there are important differences:
When known elements (1) through (3) are considered together, it is clear that Scooba's wheel bearings are severely overtaxed. Figure 6 [an owner-supplied image] shows the portion of wear that can easily be photographed:
Figure 6. Shroud-Walls Prove to be Inadequate Shaft-Bearings -- One Month of Use
Notice the enlargement around the axle, and particularly observe that the axle is in the recessed position, with a portion of its end-face vignetted by the smaller part of the hole! If you were to slice through the bearing hole and look at the cross-sectional shape of the hole, you would see there is a step increase in diameter right about where the end of that axle is shown; and that interior enlargement would be the result of a greater rate of wear due to the reduction in bearing area whenever that shaft ducked back into the hole.
When those axle-holes elongate by slightly more than 1.07 mm, the tires will begin rubbing against the adjacent fender/shroud. Ultimately enough friction will occur to halt the wheel. iRobot is issuing a notice with replacement wheels being sent to owners. The notice advises that Scooba will detect the slow, or stopped wheel (since no revolutions data are being received) and commence a series of fast spins (presumably to clear away some debris caught around the wheel); and that is the owner's clue to renew the wheel.
A modicum of wear reduction: Roger suggests that owners can reduce the rate-of-wear by frequently forcing dabs of auto-grease into the clearance gaps at axle-ends. "Frequently" means: as often as needed to keep a visible supply of grease present.
We can be certain that iRobot is working rapidly to overcome this weakness. One might expect the re-design, and re-tooling efforts to occupy most of 2006-Q1; and a good part of Q2 may go by before the new design has been field tested and reports of success come forward.
If you wish to return to the main (Scooba Technical Comments) document, click Main
Otherwise you may return by pressing the 'Backspace' key.
This page is currently maintained by G. Plews