Brush Bearing

bobnielson · Joined: 12 Jun 2007 Posts: 279

This is the second year of owning a roomba. I have several that have lasted as long as the beater bar and bursh hubs are not worn out. I always inspect the hubs to see how they are doing and if they need shims. Today I took a good look at how the bearings work and fit on the bars and into the deck. The bearing have never given me any problem so they don't get much attention. But, today I just took a good look. There are 2 ways for them to fit into the deck - I had never noticed that. The way I had just put the bearing in was so the 2 butterfly flanges are visible and horizontal to the deck. I was kind of fighting with a bearing and noticed that it could also fit with either of the butterfly flanges at an angle. Then I saw this was a better fit and that the brush end fits snuggly into the hub with no play in the hub or the bar and bearing. I suggest you do the same and see if putting the bearings at an angle in the deck will take away some or most of the end play in the bars. A much better fit and should make the hubs last much longer.

Gordon · Posted: June 6, 2008, 10:17 pm Post subject:

Bob, while I fully appreciate the utility in maximizing engagement of the square-drive members, the tactic that you are advising for accomplishing such engagement(s) is ill advised. IMO, you are trading one failure mode for another. One the one hand, you are preserving the geometry of the square-coupling features -- a useful endeavor, but, on the other hand you are ignoring heating of the brush-gears and gear-case cover-plate, which is followed by the well-known subsequent cracking/, then breaking of the Cover. Frankly, I'm not sure which failure mode is to be preferred, since they are both equally debilitating to the Roomba (assuming Roomba's non-operation is not solely due to rounding of the brush-end square).

Perhaps owner selection of the robocide process should be based on the values of "running-time-to-failure" for each process. Even with that as the criterion, I am unable to say for certain whether square-drive failure will occur sooner than cracking the transmission-gears case-Cover. The latter one has a short, built-in delay feature which must first wear away, but once that is done, the Cover heats very quickly.

Here's what happens. By clocking a yellow-bearing 60º in its mount (so its tang butts against the adjacent end-wall, rather than falling into its planned slot-receiver) free end-play of the brush is consumed (as you have indicated) before Roomba is put to work. If every bit of the bristle-brush's end-play has then been overcome, it means the largest gear ('gear#6', for short) is being pushed against some travel limit within the gear-case. When Roomba is new, travel will be limited as gear#6's central (plastic) hub (on the Cover-side) presses against a bronze bearing flange, which is good and is intended. But after some period (tens of hours -- don't have an exact number) of use that hub becomes worn by rubbing the bearing, because the hub presents only rather a small surface area to the bearing. When that wear has progressed sufficiently (on the order of a small fraction of a millimeter) the rim of gear#6 will start to rub against the inside surface of the case Cover-plate.

Each time Roomba is placed into a cleaning mode, the gears spin up to speed and gear#6 continues to rub the Cover. Any grease that may have been between the gear and Cover is quickly squeezed off to the side. The gear-rim and Cover get hotter and hotter as the mission progresses. The small-diameter, gear-hub continues to wear away, allowing even more thrust-force (side-force), then friction-heating at the gear rim. Hairs & fibers may be wrapping around the brush-ends, wedging themselves into whatever clearance-gap remains, and further increasing the thrust force pushing gear#6 against the case-Cover. Gear#4 (Flexible-Brush) is also aiding in this heating process, but to a lesser extent.

The end result (a case-Cover with the most severe heat-distortion I have seen) is shown in a video:

...which Ventura (VFM) originally linked to this forum. The video shows a severely heat-distorted case-Cover. More gears than just the two brush-drive gears put bulges in the case-Cover. If you watch that video, don't think that the fault shown is limited just to the 5XX Roomba, no, it has been with Roomba since its beginning, and it seems to be the older Roomba models that suffer Cover breakage.

There must always be some end-play to help prevent excess heating of the gear assembly. Some other way must be found to improve the square-drive couplings.

bobnielson · Joined: 12 Jun 2007 Posts: 279

Your explanation and the vidio go hand in hand, but what does it have to do with the 2 ways the bearing fits in the deck. The failure you explained and the video is all about is the great build up of dirt and hair inside the gear box of the 5xxx. A completely different subject for us 4xxx users. To take out as much play and therefore the wear from the hubs and gears is what I was trying to show. I can see no pressure put on the gears and the gear bearings to make any more heat. To have the gears with less play and less wear is a good thing and that is what I saw on my roomba with the bearings at an angle - in fact the deck and the wire thing you close down on the brushes and deck have little slotts just for this purpose. The odds of after you clean and put the brushes and bearing back will just fit at an angle anyway if you didn't know.

You comments and insight to what we need to do to keep our "Little Guys" running at best effeciency and length of service are greatly appreciated. I, for one, look for your comments and valuable advice. I will watch with great care how my roombas are running and see that what I think is a better way of keeping hubs and even gears serviceable longer.

Gordon · Posted: June 7, 2008, 12:10 pm Post subject:

bobnielson wrote: The failure you explained and the video is all about is the great build up of dirt and hair inside the gear box of the 5xxx. You are correct about that emphasis, and I neglected to foresee that incomplete interpretation. The fellow that produced the video was ignorant of the importance of maintaining free play, end-wise, of the main brushes (I say that because he said nothing about it). I'm sure his main concern was the loss of his 5XX machine in just a few cleaning applications. Quote: ... A completely different subject for us 4xxx users. Not at all! 2XXX, 3XXX, and 4XXX model-series owners have all battled gear and gear-case damage (but, most were unaware of the specific failure mode) from self-heating of the gears as they get pushed hard against the case-Cover when fibers wrap into brush-ends' clearance gaps, thus consuming all free-play. Quote: ... I can see no pressure put on the gears and the gear bearings to make any more heat. ... Bob, I would like to recommend a thermal experiment for you to try, it involves use of your finger-tip (whichever one is small enough to fit into the large square-socket (bristle-brush's socket).

The idea is to start with a room-temperature Roomba for two timed, trial runs of cleaning over the same flooring-type. 30-minutes of running (each trial) on pet-furry-carpet should be adequate (hard surfaces may require a longer run-time).

Run the first trial after positioning the brush-end bearings in the deck as prescribed by iRobot. After 30-mins of run-time, halt Roomba, quickly remove the bin and bristle-brush and poke your pinkie into its square-socket, right up against the metal spot you can see in there.

CAUTION: It is likely to be hot enough to burn skin -- only 1st degree burn!!

If you can maintain contact with the bottom of the socket, count the number of seconds of contact, up to ten or more. If you have to let go at 10-seconds, the gear's temperature is on the order of 50ºC.

After Roomba's internals cool to ambient temperature, say a couple hours later, re-mount the brushes, but with their yellow-bearings clocked as described in your top post, and repeat the 30-mins cleaning run on a different area of the same flooring-type. I expect you will not maintain finger contact with the hot-socket, nearly as long as after trial #1.

bobnielson · Joined: 12 Jun 2007 Posts: 279

I usually run my roombas at night and on a schedule but 3 times a week I take my 1 month old 400 into the bedrooms and let it roam around. So this was the first time for me to "Stick" my finger in the socket as you suggest. Only one part of what you say needs to be done - the bearings at an angle as I am now doing. But it might just be fun to run it with the bearing on the brush to sit on with the bearing flanges horizintal on the brush deck. By the way, as I remember and have done and everyone that takes the brush out to clean it (I am ONLY talking about the 4xxx) and puts it back stands a 50-50 chance of putting the bearing in at an angle or not. Since I want to get the longest use of my roombas and the hub ware has always been a limiting factor I was open to this experiment and what were the results. I set the timer to 30 minutes so I would be right on the "Touch Test". I first touched the deck and the gear housing and felt no heat - hardly even warm. I then put my little finder on the brush hub expecting it to be hot as you say. BUT - nothing - no heat, the same temperature as the rest of the area around it. I set the timer for another 30 minutes (I have a new battery that goes for almost 1 1/2 hours) with the same results.

I am NOT trying to convince anyone to change how they put the brush and bearings back after cleaning. What I am doing is trying to pay back all the GOOD tips and information I have gotten from this site. I would suggest anyone wanting to see if they can decrease the wear on the hubs to take a look at my suggestion and see if it works for you - It does for me. I don't have pets (and long haired daughter might count) and I don't have an extreme environment as some have so my roombas are not tried to the max as some do. But anything I can do to make the roomba last longer I do.