Last modified:
By paraphrasing descriptive material provided by Roger, and working the resultant text around several of his pump pictures, a fairly adequate introduction to Scooba's Pump Assembly is herein provided. From this material, owners may become intimately familiar with the pump that Scooba uses to transfer cleaning solution from its Clean Tank to a pair of jets, which are located just ahead of the Cleaning Head. A careful worker could use this section as a guide, to do this sort of disassembly work, say, for the purpose of clearing out some particle(s) that had gotten stuck in a pump-valve. Speaking of "stuck", reports have come into one of the Roomba/Scooba boards about Scoobas that can't wet the floor on their maiden cleaning mission -- right out of the box! The cause of that malady seems to be stuck-valves, which the pump cannot get working without help. Some words are given in regard to applying such 'help'.
In Figure 1, two plates of hard plastic (colored dark turquoise) sandwich a flexible plate (off-white in color -- only the pump-diaphragms formed in that material can be seen in this first figure) containing valve and diaphragm features. A rotating cam moves the arms of a yoke that then push and pull the half toroidal-dome portions of the flexi-layer's diaphragms (tied to the yoke via barbed-tips on those two domes). This action pulls cleaning-solution from the tank and moves it through the pump and out to a pair of jets under the robot.
Figure 1. Forward Side of Pump -- Blower Assembly Removed
Looking at the aft-face of the rear turquoise plate in Figure 2, the pump's inlet is via the lower / central tubing, and the outlets are via the left and right lengths of tubing. Each pump-diaphragm squirts on being depressed and reloads on being pulled back out.
Figure 2. Aft Side of Pump
The double-action pump is mechanically actuated by a motor-driven cam, a uniquely shaped cam; sort of an asymmetrical cardioid, the profile of which can be seen in Figure 3. The black-part, of course.
Figure 3. Plan-View of Pump Shows Cam's Profile
No attempt has been put forth to decipher the utility of that particular cam-shape in this application. Then, regarding that splotch of white-paint on the cam, it is assumed to be a strobe-marker.
Roger has paved the way toward 'pump-maintenance' by dismantling his pump assembly. Referring to his photo in Figure 4, he points out that vias through, and channels in the turquoise plates, in conjunction with the four circular diaphragm portions of the flexi layer, provide the valve actions needed in a reciprocating pump. When a valve-diaphragm gets flexed off its seat there is a flat portion on the rigid part behind it that stops the flex going too far.
Figure 4. Pump Dismantled
Looking at the upper-right turquoise plate, the inlet is the hole in the portion that looks like a cross section of an old film cassette (or an odd eyeglasses style), and the outlets are the holes in the centres of the circular areas annexed inward off the main dome seats. As already said, each pump squirts on being depressed and reloads on being pulled back out; so compression springs the flexible valve-disc away from an outlet-seat, permitting solution to exit. The distended disc is backed by the simple, circular feature(s) seen in the plate at lower left.
Here, lets do a mark-up of that back plate. We'll name a few features, and draw a double-dashed line that shows the path solution takes as it wends its way through the LHS of the pump, from the central INLET, to the left-OUTLET. See Figure 5.
Figure 5. Solution-Path Demarcated
Retraction of a pump-diaphragm reduces pressure in the individual output-chamber to cause the outlet's valve-disc to seal against its seat. At the same instant, the suction that occurs lifts the inlet valve-disc off its seat in the 'eye-glasses' cavity (note, you see no central-hole in those 'inlet'-seats, because the...) liquid can then flow over the rim of the seat and go through a hole in the flexible disc. Solution travels along a groove in the forward side of the front-plate for a short distance, where it then diverts aft, back through the flexi-plate, and into the 'compression-chamber' -- ready for another pressure pulse.
Perhaps a schematic of liquid-flowing through the pump would clarify its operation for some. The sketch in Figure 6 converts the kinked, double-dashed trail shown in Figure 5 into a planar, cross-section through the pump-assembly. Unfortunately, the reader will have to imagine the required movements of the (red) flexi-plate's valves and pump regions, as the rocker-arm oscillates, since only a static scene is presented.
Figure 6. Cross-section Through the Liquid-Path
Now! Isn't that perfectly clear? Cleaning solution is constrained to the centres (that's Roger speaking) of any channel system, the rest of the grooves are to form gasket sealing surfaces, and to help the flexi-plate register properly with yoke and ports.
Yes. What can be done if an un-used Scooba is turned loose on its first cleaning mission, but no cleaning-solution reaches the floor? Its Clean Tank has been checked for contents, and the tank's drain / feed connection has been verified to provide reasonable gravity-flow. Scooba runs OK after pressing Power & Clean buttons, the pump-motor can be heard to start before any other motor, but the robot scoots around on a dry floor! Hold it! Scooba won't even begin to move if clean-solution is not provided to the pump! When told to clean, Scooba first switches ON the pump motor and runs it ahead of starting any other motor. The pump for three to four seconds, and if solution is not detected in the inlet-elbow (just below the lance portion of the Inlet Connector), Scooba quits and sounds / indicates (via a blue Check Tank display) an error! This is a new machine! Don't they verify operation at the factory?
Well, we don't know what is done at the factory in the way of performance testing, but we do have a theory regarding this problem. We believe both inlet-valve-discs (diaphragms) are stuck to their seats, and that condition is a result of shipping / storing Scoobas dry! Certain freshly cast elastomers tend to self-adhere to smooth, conforming surfaces.
We think the flexi-plate's material is one of those elastomers, and its valve-discs must be developing this sort of 'adhesion' to the seats which they normally contact. Not much adhesion is needed to inhibit pumping action. When the dry-pump is operated, the suction stroke does not supply sufficient force to lift a stuck inlet valve.
We might point out that this sticking is not restricted to new machines, since it has been found to repeat in Scoobas that are not used every few days.
Owners that have experienced these sticky valves either exchange Scooba, or follow a process of pressurizing the solution inlet-connector to the robot. When adequate pressure is applied, one or both jets will eject liquid (or air -- whichever pressurization tactic is being used). If only one side is freed, that side's jet must be blocked with a finger tip as applied pressure is increased. Roger posted a good process for doing the pressurization / clearing, followed by operating Scooba on a stationary test-bed that would allow visual verification that Scooba is then able to eject solution from both jets. Read about it at Roomba-Review 060327.0622.
Figure 4 also gives a view of the pump's motor. At present, not much can be said about the motor, other than it is probably a Mabuchi motor design, looking very much like their FK-260SA motors. This style motor is found used in Roombas to operate the edge-brush and 'vacuum' impeller. In this Scooba application, it can be seen that motor-speed reduction is accomplished by worm-drive.
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