PDA

View Full Version : [FTC]: TETRIX DC Motor failure analysis


PhilBot
12-15-2009, 12:34 PM
We just came back from the MD regional, and while doing some Autonomous programming, I noticed that YET ANOTHER DC Motor was acting up.

A quick observation indicated that it was a gearbox failure (We have yet to actually smoke a motor).

When I got home, I added it to my growing collection:
Three from last year, and two from this year.

I decided to pull them all appart to see if there was a common fault.

I'm mention here that all of these motors were used to drive mechanisims OTHER than the drive system. Last year it was the Puck Bucket (used to collect and score) and this year it's the Ball Chute Trigger arm.

Here's the final analysis:

3 x Second to last gear in drive chain shattered.
1 x Final Drive Shaft main brass bushing come loose
1 x Final Drive Shaft small secondary brass bushing broken

I've attached a pic of the drive gear that failed three times. Clearly this gear cannot deal with the force/shock of a mechanism hitting something. It looks like a sintered gear (not brass or steel).

I guess if there was a published spec for MAX load on this gearbox I could make sure we don't exceed it, but there isn't one that I can find. I guess I can add some padding for impact, but that won't help if we get stuck on something.

I'd really like to find a replacement for this gear, becasue it's an expensive motor to throw away. Luckilly I was able to canibalize the two gearboxes with the busted bushings and make two good motors out of the 5 bad ones.

Phil

JesseK
12-15-2009, 12:49 PM
Indeed, we noticed a large difference in the speed of the motors from the start to the end of competition this past Saturday. I'm not sure whether or not the gearboxes are the cause, but they're definitely suspect after seeing this.

I would recommend teams add "spin up" and "spin down" code to their shooters so that the instant impact of torque is smoothed out over a couple of seconds and hopefully prevents this during a single competition.

You would expect that for $28 per motor/gearbox combo, they would use cheap gears (gearboxes alone are typically $25+). Yet being that Pitsco's website is terrible for engineers who want actual information, I don't ever believe we'll get the types of information we need to keep anything running reliably.

Rick TYler
12-15-2009, 01:35 PM
We've had four gear boxes fail so far this year -- two in our holonomic drive train and two on the launcher. Including our collection of 12 bad motors from last year, we've been able to patch together a few that work. Last year we had three FTC teams and this year we have one.

JesseK
12-15-2009, 03:11 PM
I just got some info back from yesterday's build session. The students discovered that the slowdowns were due to the excessive use of rubber bands that over stiffened the contact between the shooter wheels and the ball. They were adding more rubber bands throughout the day in order to increase consistency of the shots. Everything's back up to normal speeds now.

4 teams with a total of 22 motors on the bots and 0 broken after 1 competition. Just a different environment and different luck I guess. Last year I think our 4 teams burned through 12 or so.

aklego
12-15-2009, 08:36 PM
For those of you who have lost a motor due to the failure of the main bronze bushing, how many had an encoder installed? The encoder collar set screw, installed per the instructions, seems tailor made to crush the bushing.

PhilBot
12-21-2009, 11:24 AM
For those of you who have lost a motor due to the failure of the main bronze bushing, how many had an encoder installed? The encoder collar set screw, installed per the instructions, seems tailor made to crush the bushing.

We had the main bushing fail, and it did have an encoder mounted, but I don't think the set screw was the problem. The bushing pulled free of the motor housing.

It's likely that the motor shaft had excess side load because the encoder-mount forces you to have the drive gear out at the far end of the shaft.

It seems like the bushing quickly develops some slop where it mounts to the end of the gearbox. (It's only crimp fitted to the gearbox). It's possible that since the encoder mounting ring is wider, it converts the side-to-side torque into a lifting motion on the bushing. Thus pulling it out of the housing.

Even on our wheel drives, the losening up of the bushing enabled the encoder mount to spin freely, with only the wires holding it into place. It seems that the support ring really needs to be attached to the gearbox housing, rather than the bushing.

Note to FIRST... it would be great to be able to epoxy the encoder ring to the end of the gearbox. This would have the dual effect of protecting the bushing AND making the encoder more secure. It wouldn't have any impact on robot performance other than to prevent common failures.

aklego
12-21-2009, 05:17 PM
One of the (many) failure modes I've seen among our local teams stems from the encoder set screw. When this screw is tightened into the soft bushing, the bushing deforms and binds on the output shaft. This binding causes a torque to be transmitted to the bushing every time the motor is rotated. At best there will be accelerated wear and the bushing will survive. At worst, we send another $30 to Pitsco. The set screw torque needed to cause this failure is quite low and I consider it to be a fault of the design rather than of the installer.