After half a year of duty, the main brush did not spin. On the main PCB one of the parts (U11) did burn down. Because of the not responding European Roomba-support (I tried is several time without success, the US send me an answer within 45 minutes, but redirected me to the European support), I tried to repair it myself. .


Gordon Plews from Diese E-Mail-Adresse ist vor Spambots geschützt! Zur Anzeige muss JavaScript eingeschaltet sein! did a great job in finding out that this part most likely is a N-MOS-FET. (
Gordon, Special thanks to you!

When disassembling the Roomba memorize the position of the different screws and the cables!

Remove the side-brush-motor (1) to remove the bumper

Remove the cables (1) to remove the bumper (2)



First, the dead part was removed and replaced with a N-MOS-FET of the type BSP297 in a SOT223 package. Main reason for this type was because I found it on a old PCB :-). I used the IRF 5802 to get an idea of the pin out. With some short wires I attached it to the pins.

A first test was disappointing, the brush still did not spin.
The BSP297 looked fine, but a function check showed up that it was dead to.

A stronger type in a much bigger package, a STP60NE, was used as a second try. This time the brush started spinning, but stopped after a few moment and then started again, just like when Roomba grabs a carpet fringe. So it looked like there was something wrong with the current of the brush motor or its sensing.

A check of the motor-current at 12V supply voltage discovered that it took about 2.2 A with a free brush wheel.

A crosscheck with a comparable motor resulted, that this one runs with 0.3 A @ 12V. Fortunately I found a motor in my personal replacement storage.
The bush motor looks like a motor for (small) RC-model-cars Type: RS-385SH-2270
The motor should be for long duration, so a RC-model-car motor would be not the best solution. Thanks to Gordon for this hint.

After replacing the motor and reassembling the main-brush-unit, the test showed up that everything works fine now.
With the new motor, the N-MOS-FET first selected, should also work fine. The STP60NE is much to strong. Every N-MOSFET able to deal with about 1 A current and more then 20 V should be fine. The RDSOn should be smaller than 0.3 Ohm.
Attention: Take good care of the polarity of the motor. If it is wrong, the spin direction is incorrect, which will result in a 'hopping' Roomba. The wind-up-mechanism will not work too and Roomba does not clean very well.

Be careful when removing the cogwheel, it is very easy to damage it.

Take good care of the polarity of the motor. Its best to make a dry run and compare the result with the old motor.

When the gearing it fully assembled, make a dry run by hand and test if it smooth running.

Fixing the cables with hot-melt adhesive

The STP60NE N-MOSFET was fix with hot-melt adhesive, but it was not a good idea to attach to the case. It is better to attach it on the PCB, for example on the four diodes.

Insert the mandrel in the gearing before inserting the motor-unit


In overall. it took afternoon to repair the Roomba!

The next 'project' will be a replacement of the side-brush. It looks like I won't get a new one in Europe...

Thanks for all replys on this picture:

It is above the main brush and is a utrasonic dirt detection, based on piezoelectronic.

If there are any questions or comments, please send a email to hubert ät