Scooba's CONTROL AND DATA PANEL

Last modified:

INTRODUCTION

Control-Panel's User Interface

When a new Scooba owner first looks at Scooba's Control-Panel, s/he may notice only the centered Power and Clean push-buttons; and, that is likely what Scooba's designers intended -- 'suppress unneeded functions until they are required'! Figure 1 presents a top-side view of the panel, and emphasizes (due to the photo not being shot to expose the suppressed features!) that viewpoint. BTW, the Bumper-Shell is not shown, hence the entire upper profile of the panel is seen.

In fact, if the new owner jumps right into operating Scooba by referring only to the Quick Start Guide, no information about additional features will be seen. It is not until the owner reads pp. 6 and 7 in the Owner's-Manual that a special light-sensor and four status display zones get introduced! Owner's who know what to look for may see the light-sensor and the service-code demarkations in the Figure-1 image. Let's list them, from left to right, including the two buttons:

1. "Service" code-character display:
   When no character is displayed, this rectangular area (barely discernible in the figure, gives no hint about the characters: 1 through 9 plus b,H,d,E,F, and P, which can be displayed in response to expected trouble conditions.
2. "Power" switch / indicator:
   The primary function of this switch is to apply power to the robot, however, with red and green LEDs behind its button, this device can also indicate four battery states.
3. "Clean" switch / indicator:
   The primary function of this switch is to start Scooba cleaning the floor, but with blue and green LEDs behind its button, this device can also report the state of cleaning, as determined by the robot.
4. "Check-Tank" indicator:
   The letters in the first two words become back-lit when either tank needs servicing.
5. "Check-Brush" indicator:
   The letters in those two words become back-lit when the brush is poorly seated.
6. "I'm-Stuck" indicator:
   The letters in those two words become back-lit if Scooba's travel is halted.
7. "Light-Sensor" window:
   In Figure 1 this sensor's view-port can be seen as a dark-gray circle just to the left, and forward from the serial-connector. There are only two mentions of the light-sensor in the O-M pages. Its purpose is to help Scooba avoid stopping underneath an item of furniture.

In a prior section regarding the serial connection, the point was made that the connector is physically attached to the Control-Panel Assembly, rather than to the robot's upper chassis structure. The images in Figures 2 and 3 verify direct mounting of the connector to the underside of the Control Panel's PWB assembly.

Internal View

After the robot's upper-chassis has been separated from its lower-chassis, removal of four TFS3x1.6,8LG fasteners will dismount the Control-Panel Assembly from the chassis, and the Panel's bottom-plate from the Control Panel. Figure 3 illustrates the bottom face of the Control-Panel's PWB assembly.

As shown here, the Control Panel is still connected to the main PWB by the white ribbon cable. That cable has connectors at both ends, hence it is possible to disconnect the Control Panel from it.

On the other hand, the Control-Panel PWB is also connected to the panel switches and LEDs, by a stiff, solder-connected, four-conductor ribbon cable of short length. (Figure 3 does not show this cable, however, it is revealed in Figure 4). Manipulation of that cable is best avoided, unless necessary work is to be performed in that area.

Just in, Q4,2007: Not only is manipulation of that four-conductor ribbon-cable "best avoided", it should not be manipulated at all while it is still functional! The reason for that statement is: Cable-conductors are not metal strips, and the plastic-film insulation that separates conductors tears very easily! Once the cable is damaged, there is no practical cable-repair method. Additional information follows, to relate this wimpy cable construction to Power-switch failures which have been reported during the 5900's first year-plus in homes.

User-Interface Switches Failure Mode

Reports of Scooba's Power-switch becoming inoperable have come in to the RoombaReview forums. The owner simply can't switch ON the floor-cleaner! This has nothing to do with the owner getting inside Scooba and harming the four-conductor, flat-cable. To get Scooba working, the owner's only choice is to try use of a remote-control device (e.g., Roomba's Remote-Control Unit) to enable Power-ON.

Destructive examination of the flat-cable, and consideration of the switch-element's configuration, has given us clues as to why pressing the Power button may fail to accomplish a switch-closure. Descriptions and pix of observed features, which may participate in switch-failure are these:

1. The four conductors are not metal, but appear to be a polymer-type conductive material!
2. Dome-shaped, spring-type, tactile-feel switching elements are used to momentarily connect together pairs of the polymeric-conductors. While these dome springs are very similar to those used in Roomba's user-interface switches, one important difference has been noted: Only pressure contacting occurs in the Scooba switches. Roomba's dome-springs have spurs on their outer legs which can achieve metal-to-metal contact by digging into the copper-foil on a PCB, while a Scooba dome-spring has curved toes which only rub against the polymer. Take a look at them (one is flipped over):
   

   Figure 4. Dome-shaped Spring is the Conductive Switch Element

3. The suspicion is this: Consider the conductive polymer to be something like a narrow stripe of polyethylene film, hardly thicker than plastic-film used to make grocery bags, and highly impregnated with conductive particles, carbon-black for example, and in such high particulate density that neighbor particles are in fair electrical contact with each other. This is not a low-resistance conductor, but that's OK, its not really necessary because the conductors will carry only an extremely low current to a high-impedance input of the Micro Controller Unit. Oh yes--the "suspicion". The dome-springs appear to be formed out of 'stainless-steel', and each have four curved toes which are curved away from their contact-pads. Each toe imposes a 3-mm long line-contact with the polymer. The suspicion is that repeated depressing of the dome causes the toes to slide radially outward (a very tiny excursion) and then to return inward upon release. There is no possibility for the domes to rotate, such that toes ever contact fresh polymer, therefore, it is sort of a localized wearing out of the conductive material that finally opens the switch, forever. Of course this is speculation! An electron-beam microscope might be used to confirm this wear-out concept, but the point is it has happened to certain Scoobas.

Whatever the wear-out mode is, we can only hope that iRobot has revised this design in newer model Scoobas by substituting metal conductors and switch-pads in this user-interface assembly. Figure 5 shows the undersides of the ribbon-cable's conductors, which may be seen only after de-laminating the blue trim layer from the supporting casting (to which the PWB is screwed).

There are several features of interest in Figure 5. Notice only three conductors extend into the switches area. What is now the center conductor, is seen to be the common connection to both switches. It loops around each switch position, bulges out at eight points to provide circular contact pads for toes of the two dome-springs. Conductors are metallized full-length on this side, yet, dark-gray borders of polymer can be seen peeking out beyond the metallized areas. On the other side of the cable, metallization does not cover the ten switch-pads, and does not touch the flea-clip solder-terminals shown in Figures 6 & 7. Notice, in Fig. 5, the convex sides of the two domes may be seen through the transparent, thick-film substrate to which the ribbon-cable adheres. As a matter of explanation, the dark, sloping bar, and its shadow, at right, is one of the metal mini-spatulas used to aid de-lamination of the blue and clear thick-film trim-pieces.

These pix reveal some interesting characteristics or weaknesses in that four/three-conductor ribbon-cable:

1. Figure 7 (left): A short span of dark-gray conductive polymer is visible between the flea-clip and end of metallized span (that gray material goes all the way to the opposite cable-end and becomes the switch-poles under the dome-spring).
2. Fig. 7 (right): Flea-clip transition to solder-termination is an electrical connection of questionable reliability (see four claws on each flea-clip which are pushed through the cable, then mashed over to stake the parts together).
3. A rip from handling, shows how tender this cable is--cable thickness is only 0.15-mm (0.006-in.)!

Anyone attempting to de-solder the entire cable from the PCB, had better use exceptional care in doing so if the cable is known-good, and is expected to be re-used! Of course, if you have confirmed one duff switch, the cable can be snipped and you won't care; since de-soldering its terminals will be easy.

Figure 8 simply illustrates the adhesive-backed trim plates after de-laminating them from the UIP-structure. The entire white surface is sticky, like adhesive tape. If the trim parts are to be re-used, that sticky surface must be protected from gathering debris.

One Way to Repair Duff User-Interface Switches

Looking ahead to some point where my own 5900 suffers a Power or Clean switch failure, I became interested in determining a method of substituting some other switches for what's in there. Since I have some of the little buggers, shown next, I can salvage them from this old C-D-player's PWB...:

...(they are only 6mm SQ x 4mm tall), I feel it is possible to cut holes right through the old bulkhead (Figure 10) which has been back-stopping presses on the Power & Clean buttons--and bond a pair of these mechanical SWs into the drilled, milled, whatever, vias. Those holes would have to be placed near the centers of the curved edges of the lenticular light-pipes shown in this final image:

No engineering work, such as a scaled mechanical layout has been done to confirm fit of the planned substitutes, but enough consideration has been given that it seems quite feasible to me. Once the switches have been mounted, its will be simple to connect a three-wire harness between the two switches and their respective PWB-pads.

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