FRONT-WHEEL SENSOR PAIR

Last modified:

INTRODUCTION

Inspection of the Front Wheel Assembly has revealed a tricky mechanism that is capable of actuating both, a wheel-down sensing system, and a wheel-revolution sensing system with a single, reciprocating rod. In this section, we look at what has likely turned out to be an electro-optic assembly which transforms the up / down motions of the wheel's push-rod into electrical signals. To be more precise, we won't actually "look" at the dual-sensor's ingredients, but rather deduce what they are by using physical evidence, experience, and common sense.

Dual-Sensors Assembly

Figure 1 illustrates the forward (1-a) and aft (1-b) faces of the subject sensor module. In (1-a), the black Cap hides the parts we need to see; but, by looking at the aft face, we can see that the Cap has been fusion-staked to the PWB (see three pads on the foil-side, each about 3-mm dia.). It would, therefore, require a bit of destructive work to remove the Cap, and a little more work to renew the staking with an equivalent retention method. Hence, that effort gave sufficient deterrent to ward off further investigation -- until a closer look at the silk-screened letters on the PWB revealed the information-path!

    Figure 1-a. Wheel's Push-Rod Actuates Sensors Behind Black Cap           Figure 1-b. Sensor-Parts Under Cap, Mount on this PWB

Fortuitously, the PWB-engineer labeled the group of eight solder pads near the upper-end of the PWB. Readers will have a hard time resolving the letters in the provided, low-resolution, image, but take our word, the pads at the left edge of the PWB are labeled (upper to lower): "E","C", "E", "C"; and the right-column of pads are labeled (upper to lower), "A", "K", "A", "K". And that tells us, as any 'elektroniker' would, there is a diode connected between each A-K (Anode & "K"athode) pad-pair; and since we are studying a sensor, those diodes are likely Light-Emitting Diodes, (LED)! Knowing that much, we say "there ought to be a pair of photo-transistors, (PT), opposite those LEDs, and looking right at them". So, we now recognize the pair of E-C letters as marking the Emitter and Collector, lead positions, for our two PTs! There you have it! We have a stimulus and a detector (slotted, optical-switch) combination, and we deduce these four devices must be packaged as side-lookers -- just like the LED & PT packages seen in the Tachometer Assembly section (Figs.3 & 4). Now, all that's needed is a translatable beam-shutter; to block / un-block one IR-beam, and to modulate the other IR-beam -- depending on what the caster-wheel is doing.

The Sliding Shutter

We must now envision what the beam-shutter might look like. For 'starters', we may say it is a rectangular-block of opaque plastic; and with a hole, about two- to three-millimeters in diameter, passing through the block, from one narrow side-face to the opposite side. Of course it must be sized to fit between the EO-devices, and to slide up and down, as the Wheel-Assembly's push-rod position dictates. It must also be thin enough to fit under the Cap, and a spring must be included to force the block down as the push-rod lowers. If we study Figure 1-b some more, the following information may be deduced:

1. Block-guide / Lens' Guard: Notice the circular, black dots interior to the two columns of solder-pads. Those are the ends of "pins", cast as part of the black Cap, passing through registration-holes in the PWB. It is reasonable to assume that those four pins also relate to surfaces within the Cap, that serve to guide the sliding-shutter, and to prevent it rubbing on the PT and LED lenses. With those assumptions in mind, we can claim the the left-right width of the block will later be discovered to be a millimeter, or two, less than the horizontal span between left and right pin-pairs.
2. Scaling-Data: Fairly accurate PWB-scaling information may be obtained by use of the pin-separations of the side-looking EO-devices. Looking at their data sheet(s), we find their leads are spaced 2.52-mm. Further, but without actually measuring, the vertical spacing between adjacent PTs' leads, and LEDs' leads, is such that those neighboring devices are also, approximately, 2.5-mm apart. So, that all goes together to allow us to say the vertical separation between the lower light-beam-axis, and the upper one, is on the order of 5-mm.
3. Upper & Lower Beam Functions: BTW, we are adopting the notion that an open light-path equates to the logic HI condition (The inverse logic may be used. Which logic is actually in use, is not important to this discussion.). IOW, when the shutter-hole is aligned with the lower beam, the PT's signal would alert the Controller about a dropped Front Wheel. Then, as the "knee-action" travel of the caster-wheel is pushed up to its limit, the push-rod elevates the shutter-block, thus moving the cross-hole off axis to block the lower, light-beam -- which signals the Controller: "the front-wheel is UP".
   When the wheel is elevated to its running height, the push-rod and shutter-block have moved upward about five-millimeters, with respect to the wheel-down position. At that point, the cross-hole may be aligned to the upper-beam's axis or not. However, as Scooba starts traveling, the wheel's rotation will crank the push-rod up and down, and (presumably) translate the cross-hole in and out of the upper beam-path. The upper PT's signal, HI, LO, HI... with each revolution, can then be counted by the Controller.
4. Cross-hole's Location: If everyone buys into that story, we can claim that the cross-hole must be located very close to the top of the shutter-block. The reason for that is: The upper face of the block can't go much higher than the upper pair of "pins", else it will hit the top wall of the Cap. Therefore, the cross-hole has to be near that top-face, so the hole is in a position to modulate the upper-beam.

The above description ought to be reasonably accurate. If any owner has cause to peel the Cap off his Scooba's Wheel-Sensor Module, we would be interested in seeing photographs of its interior. Now let us talk about replacing those EO- components.

Replacement Electro-Optics

Let us be clear at the start of this paragraph -- we do not know exactly what LED and PT are built into this sensor module. Thus, we cannot say exactly what parts may be used to replace a duff LED or PT. However, it seems reasonable to make the following statements, which will then form a basis for offering suggested replacement-parts:

• These same, side-looking, EO-device configurations are used numerous places in Scooba (three-places, anyway) and in all Roomba models. If you were manufacturing these robots, would you not use the very same part-numbers for every application (barring some [unknown] specific deterrent)? Of course you would. It makes economic sense to do that.
• Over the last couple years, enough reports have trickled in (to the several Roomba-boards) regarding weak or failed LEDs, perhaps PTs too, used in drive-wheels, and bumper-switches, to demonstrate that these are not six-sigma hardware items (nor should we expect them to be!). Since they are not high-reliability components, and they do exhibit a modest failure-rate, it would be wise to be prepared with replacement data.
• Based on the preceding two elements, Scooba owners must expect a similar incidence rate, possibly a lower rate -- just because Scooba runs cooler. But stuff happens.

The nice thing about failures of these devices is this is one of the few items that an owner stands to be able to repair! Here are some EVERLIGHT part-numbers that have been found to fit and work for replacing similar Roomba parts:

Everlight, Side-looking IR-LED: P/N = EL-IR928-6C; in USA proc. from www.mouser.com .

Everlight, Side-looking Photo-transistor: P/N = EL-PT928-6C; in USA proc. from www.mouser.com .

The PT is the more expensive part, of the two. Its unit price, in 2005-2006, was-is $0.22! :-)

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