Society of Robots - Robot Forum

General Misc => Misc => Topic started by: Admin on June 12, 2007, 01:55:06 PM

Title: Servo Cycle Lifetime
Post by: Admin on June 12, 2007, 01:55:06 PM
I plan to do some genetic algorithms on a robot that uses servos to actuate.

Here is the problem:
Genetic algorithms require days and months of continuous actuation, and in my case I calculated 3.5 straight days is enough to get the data I need.

So the question:
Does anyone know the lifetime of a servo, ie number of times it can reverse direction (medium load) till it breaks? (Im using the Futaba S3110)

Anyone had a servo break from continuous use and can tell me the servo name and an estimate number of cycles before failure? (rough treatment doesnt count)

(I guess I will need to run a break test on a few servos and see how they fare)
Title: Re: Servo Cycle Lifetime
Post by: hgordon on June 12, 2007, 02:02:00 PM
It's probably been 5 years since we ran these kinds of tests, but we were using Hitek servos for a webcam pan/tilt, and found that the resistive film on the feedback potentiometer would wear out very quickly on the HS311, but units with the HS422 servo stayed in service for years.  The durability of the resistive film was a much more significant factor than metal gear vs bushings.  You might want to take apart some servos to inspect this feature.
Title: Re: Servo Cycle Lifetime
Post by: maverick monk on June 12, 2007, 02:36:57 PM
my rc car has had its front wheels going back and forth for 2 years, ill figure i drove it about 2 hours a week and thats with the weight of a 1.5 lbs car pushing the wheels at about 20mph... so, with medium resistance on a nice servo like that futaba id guess, 300+ hours of continual function, assuming the motor dose not overheat and grit dose not cause the gears to wear
Title: Re: Servo Cycle Lifetime
Post by: Admin on February 03, 2008, 11:47:43 AM
A quick report of a servo failure:

HS-225MG . . . supposedly it was supposed to have all metal gears, but then the servo started making clicking noises.

Opening it up, I found that all but one gear was metal - its the second gear after the motor shaft gear. Apparently its plastic, and literally half the teeth have been stripped from it.

I was using it as a wheel motor, so no high forces were ever applied to it. I'd say the servo has experienced a subjective 'medium' amount of usage.
Title: Re: Servo Cycle Lifetime
Post by: Admin on March 31, 2008, 08:20:02 PM
Another servo failure report . . . the HS-225MG again!

This time its the little metal pin that somehow snapped off on its own. It appears they redesigned the metal pin to have a much smaller diameter where it inserts into the gear . . . I was lucky to have an extra gear set for the last time a gear broke. And I'm not abusing this servo either! not acceptable . . . I think I'm gonna write Hitec, they've answered previous emails of mine . . .

And as I mentioned in another post, another HS-225MG went up in smoke literally when accidently reversing battery polarity for a second . . .
Title: Re: Servo Cycle Lifetime
Post by: Admin on April 06, 2008, 10:53:45 AM
Man, yet another HS-225MG failed :'(

My replacement gear broke in the servo that just recently broke.

Hitec never replied to my email about this problem . . .

I'm no longer buying this servo again, and will try other brands from now on.

bad product + bad service = lost customer :-X
Title: Re: Servo Cycle Lifetime
Post by: Admin on April 17, 2008, 04:12:04 PM
Another HS-225MG just went up in smoke today (literally), and I wasn't abusing it or anything! :'(

I got four left, lets see how long it takes for them to fail . . .
Title: Re: Servo Cycle Lifetime
Post by: Admin on April 17, 2008, 04:50:27 PM
Hmmmm I believe I'm cursed.

Only 30 minutes since the last failure, and I already have another broken servo :'(

This time the HS-645MG. Funny thing tho . . . the gears didn't break, but the plastic casing where the stopper is broke!

I'm really starting to think Hitec servos are low quality . . .

With the exception of one servo, all these failures are occuring on my robot arms (that are only supporting its own weight) . . . I'm thinking Hitec servos aren't design for sudden electrical or mechanical shocks . . .
Title: Re: Servo Cycle Lifetime
Post by: Spoil9 on April 17, 2008, 04:55:56 PM
Admin,
Is Hitec servo's all you have bought in the past? If so, is this a new thing that they are failing cause it might just be a bad batch or something.
Title: Re: Servo Cycle Lifetime
Post by: Admin on April 17, 2008, 05:27:45 PM
I also buy futabas and they break too - but I abuse and mod my futaba servos so its understandable.

There is always a chance Hitec has changed their manufacturing methods or swapped to shoddy components/manufacturers . . . in either case, they never replied back to me.

I was hoping I could avoid this, but I guess I need to rewrite my servo control algorithms to smooth it from shocks . . . a lesson for all of us - don't do ninja moves with robot arms! Should be more Yoga like . . .
Title: Re: Servo Cycle Lifetime
Post by: AndrewM on April 17, 2008, 06:00:56 PM
a lesson for all of us - don't do ninja moves with robot arms! Should be more Yoga like . . .

When it comes to taking over the world, Yoga moves are just not going to get the job done
Title: Re: Servo Cycle Lifetime
Post by: Asellith on April 18, 2008, 06:28:15 AM
You know what this means admin. After you complete your Axion run its time to design and build your own super servo!
Title: Re: Servo Cycle Lifetime
Post by: ALZ on May 03, 2008, 12:29:15 AM
Hi: I think maybe you want to much from them. Look at all they have to do in that small space. The gears are so thin and the chips really don't have any place to get rid of the heat. No wonder they don't last long. For the money you spent on blowing out servos, you could have use a gearbox motor and built a bridge driver. I know a servo is taking the easy path.   

Hmmmm I believe I'm cursed.

Only 30 minutes since the last failure, and I already have another broken servo :'(

This time the HS-645MG. Funny thing tho . . . the gears didn't break, but the plastic casing where the stopper is broke!

I'm really starting to think Hitec servos are low quality . . .

With the exception of one servo, all these failures are occuring on my robot arms (that are only supporting its own weight) . . . I'm thinking Hitec servos aren't design for sudden electrical or mechanical shocks . . .
Title: Re: Servo Cycle Lifetime
Post by: Admin on May 03, 2008, 05:57:25 PM
Quote
I think maybe you want to much from them.
I used them within the specs on the datasheet. I didn't abuse them. If I abused them, which sometimes I do, then it'd be ok if they failed . . .

Quote
For the money you spent on blowing out servos, you could have use a gearbox motor and built a bridge driver. I know a servo is taking the easy path.
You forgot that I also need a feedback position sensor, control electronics, and a mount for it. Not so simple . . . I'm building a robot here, not a servo :P

Anyway, I figured I'd upload what a fried servo looks like. Not too exciting . . . the top servo is undamaged, the bottom one smelled all burnt like . . .
Title: Re: Servo Cycle Lifetime
Post by: ALZ on May 04, 2008, 12:07:41 AM
No, I didn't forget. Who ever said making a robot was easy? not me! At some point if you want to make a robotic arm that can lift more than a small ball  you have to get away from useing RC servos.


"You forgot that I also need a feedback position sensor, control electronics, and a mount for it. Not so simple . . . I'm building a robot here, not a servo "

Title: Re: Servo Cycle Lifetime
Post by: Admin on May 04, 2008, 07:16:24 AM
Quote
At some point if you want to make a robotic arm that can lift more than a small ball  you have to get away from useing RC servos.
Actually, based on my calculations and the specs on the servo datasheets, the two robot arms on my ERP can allow it to do pullups - serious. But, given the failure rate and cost of the servos, I wouldn't actually try that . . . There are some fairly beefy servos out there, I could have my bot do some heavy lifting with those before I'd need to move to stronger servo-fied DC motors. But I have a personal preference towards small robots that can work well with hobby servos - cheaper and easier.

Oh, and my ERP has 11 servos on it, and will have 3 more soon - very expensive if I used servo-fied motors!

(servo-fied should be a word)
Title: Re: Servo Cycle Lifetime
Post by: Metal Slug 2 on February 15, 2009, 12:12:42 PM
. . . a lesson for all of us - don't do ninja moves with robot arms! Should be more Yoga like . . .

I agree.  After building my $50 robot i noticed that when i turned it on the bot would surely by photovoric (is that even a word?) and everything, but when it would turn to avoid a shadow either servo would suddenly jolt in the opposite direction.  when i had modded the HS-311's i noticed how fragile the nylon gears were and after seeing the sudden forces applied to them i decided to modify the original photovore code so increase the servos life.  I just changed the servo speed values so that when turning only one servo moves to turn the bot and the other remains still, so the bot basically pivots on the non-moving wheel.