Built in tests - wheel tests

arnard · Joined: 04 Sep 2007 Posts: 36

I am getting a red status light for both tests 7 and 8. I went so far as to run without belts and also without the right motor even plugged in. I still get red lights indicating a current overdraw. Can this also mean an underdraw as well or is there a problem with the board triggering the overdraw indicator?

Any other ideas?

arnard · Joined: 04 Sep 2007 Posts: 36

To answer my own question for future search results, yes the main board can tell. Either it monitors current draw or it also is monitoring the tachometer readings, but as soon as I cleaned both wheels and cleaned and regreased everything was back up and running and test 7 through 9 went perfectly.

RoombaExchange · Joined: 21 Aug 2006 Posts: 1187 Location: Florida USA

Arnard, In tests 7-8 the power should be illuminated red and the status should be green unless there is an over current condition. You do need both wheels connected to the PCB to complete the wheel tests since the LED/PTs share the same power supply and if one is missing it results in an open circuit.

Here is a data dump from the Diagnostic routine for TEST 7;
factory-test 7 left-wheel
not left-wheel-stall PASS
(baseline-current-ok?) mA -91 min -202 max -13 mV 15558 degrees-C 26 PASS
(drive-speed-ok? left forward) PASS
(drive-speed-ok? right stopped) PASS
(left-drive-current-ok?) mA -215 min -326 max -117 mV 15363 degrees-C 26 PASS
not left-wheel-stall PASS
wait: left-wheel-stall

There is a MIN and MAX current draw recorded by the MCU so an over current, below -326mA will report a failures as a red status light, although I've never run into an under current, above -117mA demand it would likely report the same way as an over current. Possibly Gordon or Vic have seen it before.

[Arnard, looks like you beat me to a reply]

arnard · Joined: 04 Sep 2007 Posts: 36

Barely Chris.

P.S. Go to bed

RoombaExchange · Joined: 21 Aug 2006 Posts: 1187 Location: Florida USA

Thanks, I'll take that advice.

Gordon · Posted: May 5, 2008, 10:35 pm Post subject:

RoombaExchange wrote: ...Here is a data dump from the Diagnostic routine for TEST 7;
factory-test 7 left-wheel
not left-wheel-stall PASS
(baseline-current-ok?) mA -91 min -202 max -13 ...
...
...
(left-drive-current-ok?) mA -215 min -326 max -117 ...
...
... Chris, I can't make any sense out of the extra values in those two reports. Viz., the "13" as related to baseline-current, and the "117" related to drive-current. The max & min limits are all one needs, I would think.

Those two values can't even be the 'actual' currents that were measured, since they are low enough to have caused FAIL to be displayed two places.

RoombaExchange · Joined: 21 Aug 2006 Posts: 1187 Location: Florida USA

Gordon, I agree that the MAX and MIN are related to the same need and is the MIN/MAX set by the manufacturer and what the MCU uses as a basis to compares the motor current demands for rating as PASS/FAIL.

Here is a set of readings from a brush motor from 3 different roombas for comparison;
Quote: (brush-current-ok?) mA -377 min -566 max -208 mV 15251 degrees-C 20 PASS In this roomba the motor ran within the current demand and the status light would show a green after test 11

Quote: (brush-current-ok?) mA -638 min -566 max -208 mV 15251 degrees-C 27 FAIL This roomba motor was drawing too much current and reported an OC condition with a red status light for test 11.

Quote: (brush-current-ok?) mA -6 min -566 max -208 mV 15251 degrees-C 26 FAIL This roomba had either a siezed motor or a failed field coil, I dont recall which it was but both conditions have always reported a -6mA

As a load is applied to any brush motor in this example the first -mA value (based on 1A) will continue to decrease (as demand increases) until it reaches or surpasses the -566mA at which time the report by the MCU would change from a PASS to FAIL.

Gordon · Posted: May 6, 2008, 7:53 am Post subject: Rules are rules!

Chris, thanks very much for taking time to assemble those three examples for me.

My problem is now clear to me: I was attempting to use rules developed in science and engineering fields to interpret data records which had been left to a non-engineering programmer to format!

I had already become well aware (from vic7767's published charging dumps) of iRbt's ass-backward placement of units before value. But, last night, I failed to also apply that non-orthodoxy to the "max" and "min" flags; hence, I was assigning the wrong values to max and min.

The issue is further complicated by the programmer calling a more negative number (higher-value preceded by minus sign) "min", when, in fact, it is the maximum! The programmer is failing to note that his/her use of the minus sign (with electrical-current values) in these records is simply a device to show whether charge is leaving (-) or entering (+) the battery; NOT where the value reside on a +/- number-scale. That sign should be ignored, since even when a Roomba wheel-motor is run in reverse, the reported current will be preceded by a minus sign (charge is still being taken FROM the battery, even though we know current through the motor has been reversed!).

Of course there is little use in whining about this formatting--it was not done for us! I will just have to learn to use it. I'll just have to paste the following crutch up on the wall:

Code: (brush-current-ok?) mA -{ACTUAL} min -{MAX} max -{MIN} mV 15251 degrees-C 27 {STATUS}
---

vic7767 · Joined: 14 Jan 2006 Posts: 3099 Location: Louisiana

Thanks for that memory jogger Gordon. That does help plug in the values correctly when attempting to interpret the SCI output.

RoombaExchange · Joined: 21 Aug 2006 Posts: 1187 Location: Florida USA

Gordon wrote: The issue is further complicated by the programmer calling a more negative number (higher-value preceded by minus sign) "min", when, in fact, it is the maximum!

No argument here.... I fight the urge to explain this every time it comes up in the shop because logic dictates that the values are being stated in reverse. :shock: It takes a strong cup of java just to see straight afterwards.