ON- GOING-OBSERVATIONS AND INVESTIGATIONS

Last modified:

INTRODUCTION

The principal thrust of this document is to diagnose the cause of a catastrophic failure of a transient-voltage suppression (TVS) device. Destruction of the TVS-component was so dynamic that no portion of the device remained on the board, its connection pads were vaporized, and the G-10 board material was charred through approximately one third its thickness. Adjacent connector-jacks suffered fused inserts, resulting in welding the plug-cap to the jack in both cases. In addition to studying that problem, there is a desire to understand the operations of several other sub-circuits, e.g., the current- sensing of all motor-currents and their assigned protection thresholds; and the on-board battery-charging control operation (affiliated with the blown component). It is expected that this duff PWB-assembly will come close to supplying circuit-data that will give some understanding of those functions.

FAILURE BACKGROUND INFORMATION

Raul's main PWB was taken out of operation when a protective device absorbed more energy that it was designed to absorb. The owner had the impression that "...there was a power surge during a monumental thunderstorm and the charging brick was incapable of handling the additional load and blew up". Raul's charging brick was obviously duff, flashing its green pilot LED. Following the electrical fault, Raul would light the green Power button, but none of the other buttons worked at all. There was no change after doing a battery-out, system reset. Various charging attempts failed.

INITIAL FINDINGS

Meta-Data

Disco "Raul", with SERNO = 4210 041113 E xxxx, is having its PWB rev-engineered in certain areas. When necessary (i.e., data can't be obtained from Raul), data from Noll's Disco (SERNO = 4210-040910-xxxx) will be used as a backup source. Noll's Disco is older.

Identifying numbers on Raul's PWB are: 2004.09.11 | 74328-003-0032(07-03).

Identifying numbers on Noll's PWB are: 2004.06.26 | 74328-003-0030(05-06)

Failure Indications on the Hardware

A "30 | CA" transient-voltage suppression (TVS) device was so energetically destroyed that no portion of the device remained on the board. Its connection pads were vaporized, and the G-10 board material was impacted over a 5 x 10 mm zone, and charred through approximately 1/3rd its thickness, over a 5 x 5 mm zone, and down to the underlying 'V_batt_pos' plane of copper. Adjacent connector-jacks, just left and right, suffered fused plastic bodies, resulting in welding the cables' plug-cap to its jack in both cases. The following image shows the damaged area and connectors.

Identification of Failed Part

In the preceding image, three other 30|CA TVS can be seen. They have been given the location marker of a 'diode' -- which is appropriate once the construction of these TVS are known. The 30|CA happens to be non-polar because it is formed by placing two (special) break-down diodes in aniti-polar series connection. Four of these 30|CAs (a.k.a. "CCD|CK", on other Disco PWBs) are used:

  1. The one that got blown out, normally shunts the two pins of J15 ( charging-power from the rear, co-ax jack). The locator identification of this TVS has not yet been determined, so it will be called "D##(198,6)" (named by its location coordinates) until its true label is determined. Even looking at Noll's PWB was no help -- the 'Dnn' marking must be under the device (a practice used in tight spots, elsewhere on the board)!
  2. D32 has been found wired across the top and bottom pins of J7, the power connections, 'V_batt_pos' & 'V_batt_neg', from the battery.
  3. D35 is seen to be wired from the center pin of J7, to its bottom pin. That connection shunts the temperature-sensor line coming from the battery.
  4. The fourth TVS, D12, shunts the two pins of J25 (charging power from the Home-Base contacts.

REFERENCE INFORMATION

In this section, all components which seem to be affiliated with the topics of interest, will be listed. Their known meta-data, along with location coordinates will be provided; and some interconnection data may also be listed.

Coordinates to Components Centers

Dimensional data is used herein to roughly record the location of every feature of interest on both sides of the PWB. Measurements are to the center of the feature, and are referenced to the PWB's far-left (Roomba's LEFT), upper-corner (a virtual-corner, since there is a small, physical-notch in the card's corner) as the coordinate origin, x=0,y=0, for both sides. The positive x-direction is from left to right; the positive y-direction is from top to bottom; then, when viewing the front face (side nearest the battery) of the PWB, the positive x-direction reverses (in the viewer's mind) to go from right to left.

Method for Numbering Pin-Arrays

The Hirose Electric Company's connector bodies are so small, it is not feasible to apply contact numbers to them; and, similarly, integrated circuit pin-outs are at times unknown until the device has been identified. For such components using a two-dimensional array of pins, a pin-numbering scheme based on matrix-element numbering will be used. The number-sets will be applied by viewing the array of solder-pads from the rear-face of the PWB.

Here is an example of a six-pin connector that is oriented with two pairs of three pins lying on x-directed lines:
|1,1;1,2;1,3|
|2,1;2,2;2,3|
In other words, the top row is number one, and column-numbers march off, increasing, to the right (three columns in this case).

There are two, six-pin ICs that are oriented like this. When referring to pin-#1 on IC U11, it will be stated as "U11[1,1]", showing an "element" of the matrix in [ ], square-brackets. U7, in another example, does not use one of its six pins, there will never be any mention of that pin in my own continuity-tables, but it can be 'named' via this process -- even without knowing what the device is -- as U7[2,2].

A single row, or column of pins may also use the matrix-element numbering to not only identify the contacts, but to also provide explicit orientation knowledge, when you know the total number of pins. For example, the two-pin jack J3, at the PWB's left end, has pins: J3[1,1] and J3[1,2].

Here are some 'constants' that may help keep us oriented:

Main-PWB Jacks and Cable Destinations

By looking at several PWB-asms from the Disco series, it has become apparent that connector-numbering is not consistently maintained. To avoid confusion, I will list the connector labels and functions found on Raul's PWB-asm. Anticipate that the data may differ for other PWB versions.

Rear-Face-Mounted Jacks and Their Plug's-Harness Data

In the following list, jacks are listed in the order left to right, as seen by viewing the PWB's REAR-face, specifically viewing the pads into which the pin-leads solder.

Forward-Face-Mounted Jacks and Their Plug's-Harness Data

In the following list, jacks are listed in the order right to left, as seen by viewing the PWB's FRONT-face. Remember, dimensional data uses the same coordinate-origin defined for the rear-face.

Identified Components

The following elex-components have been identified. Data sheets have been collected as possible. Locating coordinates are referred to Roomba's-left, upper-edge corner of the PWB (whether viewing the FWD face, or AFT-face), and measured to the center of the pad-array; units are "mm".

Capacitors

Diodes, Including TVS (varistors)

Fuses

Transistors

Resistors

Integrated Circuits (caution, some FETs are labelled U...)

Jacks

Miscellaneous

Unidentified Components

text-TBD

This page is currently maintained by G. Plews