Singapore
Asked — Edited
Resolved Resolved by thetechguru!

Servo Burnt Out?

I was running JD successfully last night, up until the time I had to put him on charge. Charged up the battery, disconnected after the charge. Tonight after powering him up as usual, I found one of the servo's (Port D5, his left 'forearm') was unresponsive.

Symptoms are: the servo doesn't move when instructed. The servo is able to be rotated by hand even when it should be 'locked' in place (similar to it's powered down state in battery saver mode). There's no change if I place him in the 'calibrate' position. All other servo's are in their normal calibrated position.

There was no smoke and no buzzing from the servo last night prior to powering JD down. I have disconnected the servo from the port and reconnected it as well. There are no obvious signs of the wires at either end being frayed or unusual.

I am wondering if there's anything else I can try, or if this is an indication of a dead servo? I received JD just over 2 weeks ago.

Any advice would be appreciated. Thanks.


ARC Pro

Upgrade to ARC Pro

Synthiam ARC Pro is a new tool that will help unleash your creativity with programming robots in just seconds!

#1  

Try the servo on a different port - instead of running the JD project, just start a new project and add a single servo control. Define it for the port you are plugged into and set the minimum and maximum values. If the servo is working it should move as you change the values.

Servos are wear items and can burn out, particularly if they try to move beyond a physical limit or otherwise have their movement hampered (cable pulled too tight... A common mistake is threading the servo cable through the rectangular hole in the end of the lever. That will cause the servo to bind and burn out).

I had one EZ-B digital port fail, but it was completely my fault and would not occur with a servo plugged in (shorted out the pins).

Alan

Singapore
#2  

Thanks Alan,

Good suggestions, unfortunately no luck. Changed it to Port D23, opened a new project as suggested, added a horizontal servo control, tied that to port 23 and set min to 45 and max to 113. No movment at all when running the control through the allowed values.

I can see what you'd mean by threading the servo lead through the rectangular hole, but that's not the case here either.

JD occasionally emits a low hum when coming to rest between autopositions, but I believe that's normal, he certainly never emitted a loud buzzing, and each servo had gone through both tuning, and then fine-tuning originally with no adjustment values > 10. There's no visual discoloration on the outside of the casing to indicate a burn-out either (I'm not techy enough with hardware to feel confident of checking the inside of the servo casing).

Any other thoughts or do I need contact the robot hospital and order another servo do you think?

Singapore
#3  

Edit for clarity: There really is no such place as robot hospital, it's just what I tell JD to make him feel better.

#4  

Unfortunately it sounds like you need a new servo. It is possible that it is just a loose connection inside, but if you aren't confident with checking inside the housing, then it may not be worth it (although if you are going to be doing a lot with robotics, getting handy with a soldering iron would not be a bad skill to learn).

You could try clicking the Contact Us link and reporting as a warranty issue (reference this thread). I am not sure how EZ-Robot is handling warranty issues on servos these days since it is so hard to determine if there was abuse, but since yours is only a couple weeks old, they may take pity on you.

If you do need to order, I would get several spares since the shipping costs for just one servo are more than the servo itself costs, and because they do occasionally burn out. Robots like JD and Six, where all of their weight is held up by the servo's active location holding are more likely to burn out than a Roli or Adventurebot too. I would not expect to lose a lot of them. I have had Revolution robots since they first started shipping and only had one servo burnout and it was completely my fault (claw too tight for too long when I got distracted), but I still have several spares just in case.

Alan

Hong Kong
#5  

Roughly 2 weeks ago, I got a JD too. After the excitement of watching JD several dances (several charges, quite continuously during a day) upon finished building, both shoulders "died", one freely moved but no response, and the other locked up but didn't moved nor responded, and ended the cheers and happiness of all the watchers old and young.

Also I got a Six a month ago, until now 2 servos died too.

Upon opening the servos, I can see that one of the FET, if not both, on the little circuit board have the usual sign of burnt-out, and smelt. (Wonder if ezrobot has any replacement of this little circuit board for the servo?)

Singapore
#6  

Thx Alan, I'll contact support but also order some spares as you suggest at the same time, since shipping is (as you say) relatively costly.

#7  

@DJ

Have you recently examined the currently shipping servos to make sure your manufacturer is accurately following the BOM? If you recall this thread from Giskard https://synthiam.com/Community/Questions/9282 she was trying to point out that it looked like the servos were using cheap knockoff Mosfets rather than a branded Mosfet 9926a (I think you missed that point and went on to explain what a Mosfet was, which she obviously knew).

If that is what you are paying for, and you are satisfied with the performance, that is fine, but if you are paying for 9926a's and not getting them, it might explain the slight uptick in forum complaints of early servo burnouts over the past couple of months.

Alan

PRO
Synthiam
#8  

The servos are great still - only people with an issue will post on the forum. So, the one user per month with issues is pretty good :)

also, if you look at giskard's custom builds. There are some questionable builds with incredibly long arms being lifted by a single servo - which disregards all servo warnings and usage.

Servos are a wear and tear item - and will damage over time. Specifically when used in a robot that is carrying a significant weight.

You can revisit that thread to read my reply regarding that you cross-your-fingers that the mosfet burns out rather than a fire. Be happy the mosfet is damaged instead of the wires catching on fire, the battery exploding or worse...

#9  

ok, just checking to make sure you are on top of your manufacturer and they aren't slipping in sub-par parts. There are two users in this thread alone, and at least one more this month who claim normal usage (of course, impossible to prove).

When I worked for a hardware phone system vendor, we had issues when we moved some of our manufacturing from Canada to Mexico and China. Mexican company could not get the injection molding right, and the Chinese company left parts off some boards (which we did catch in fixture testing, but it delayed some deliveries while we got them straightened out). Because our boards cost between $1500 and $10,000 depending on the function, we would randomly pull one out of every 100 and check that every component was exactly what we ordered. Other than the missing parts in the first run, we never had them cheat us, but some of our employees had experience from previous jobs where it did happen.

Alan

PRO
Synthiam
#10  

EZ-Robot pays for the components that go to our manufacturer - the components are ordered and shipped to their warehouse. They don't source components, that's not how it works. EZ-Robot also has our own exclusive manufacturing facility, it's not that we outsource or contract a company to manufacturer. EZ-Robot has a facility exclusive to our manufacturing needs. Nearly a million EZ-Robot servos have been shipped - and 1 servo a month is damaged on this forum. Not bad :)

Although, it's the very bad for the 1 user per month who experiences a bad servo. I, like you, emphasize... but it's impossible to rectifiy

#11  

For me, so far with all the servos I've use on CY, for its arms and head. I have only had one servo ever burn up. Sadly, it was a new servo from Brookstone.

The arm moved a couple times, then stopped (was not even holding anything). Next thing I smelled a little smoke, looked over to see a red flame shooting out if the servo followed by part of the servo wire beginning to melt away.

What a site to see!:) never seen a flame shoot a servo before.

Was able to turn the robot off and eventually the red flame went away. I took the servo apart, pretty much the entire pcb was a burned out mess.

I think the servo simply just got stuck in a position during a movement, and that over-heated everything. All the other servos are still working fine too.

#12  

Thanks for taking the time to give a detailed answer, and great to know more about your operating model.

I have posted many times about the huge number of out of box failures in the cell phone and other consumer electronics industries. The fact that you have so few failures really is amazing and a sign you are doing things right.

Alan

PRO
Synthiam
#13  

Thanks! It really is rewarding to experience low number of issues.

Ironically, Jeremie and I were having a lengthy conversation today. Sure, it was great to sit in the sofa chairs at the new facility watching construction - man it's going to be a sexy place. But the conversation was great because it touched on this subject.

Quality has always been really important to me - and I have always disliked feeling ripped off with purchases. So I push the QA pretty hard. But the conversation with Jeremie was interesting because it touched on quality and development time. Ezrobot has been pushing hard to release new products... More than the ones you are aware of. And these products are built using components. The challenge of development time is actually a question about component quality review and assisting third parties to better their product.

What I mean from this is that product development uses third part components. And if those third party components do not meet our requirements (or their specification), we have to work with their engineering team to fix their product.

In short, ezrobot product development actually takes longer than it "should" because we work with other component manufacturers to make their stuff better so we can use it. It's a bizarre scenario, but as Jeremie put it "ezrobot is advancing the industry by pushing component manufacturers harder".

its interesting that if the "robot industry" is supposed to be taken seriously... That ezrobot is really one of the only companies actually "consumerising" it. Meaning all these diy component manufacturers get away with murder by not meeting their specifications or just downright don't work at all!

I can't tell you how many servo manufacturers we have gone through to get ours dialed in. And the number of servo revisions? Don't get me started!

Battery chargers alone... We've reviewed and destructive tested dozens - and even when we find one that "works", we still have to work with the manufacturer to fix/tweak features or components.

Starting last month, ezrobot even has our own lipo battery brand now! The new upcoming battery charger is another ezrobot one-off with revision qualities that aren't in their regular version.

If you had an opprotunity to tour the ezrobot facility, we would show you the shelves packed with containers from each component review and revision. In servos alone, there's close to 75 revisions - of just OUR servo! That's how many attempts it takes to get it right. And that's not 1 servo per revision... It's dozens per revision for proper testing!

It isn't easy to make "a product"... Imagine making 30! Ezrobot has dozens of products! We have cameras, ezb, display, plastics, servos, software with hundreds of controls. Man, it's a lot!

Even the ultrasonic distance sensor is changing this summer. We are fed up with these hobby sensors so we designed a piggyback board that converts the sensor into a single wire and filters out more false positives.

I don't get it, that some people are okay with half-working products - not me!

In short, if a servo burns up - sure there's a very slight chance of manufacturing defect. But in more than 99.9% of the time, it's wear and tear or usage related. And when that mosfet pops, be happy it did!:) because it could have been worse. The number of tests to identify correct components that acted as fail safe took a lot of work.

Hong Kong
#14  

@DJ,

It seems you missed this:

<Quote>

Roughly 2 weeks ago, I got a JD too. After the excitement of watching JD several dances (several charges, quite continuously during a day) upon finished building, both shoulders "died", one freely moved but no response, and the other locked up but didn't moved nor responded, and ended the cheers and happiness of all the watchers old and young.

Also I got a Six a month ago, until now 2 servos died too.

Upon opening the servos, I can see that one of the FET, if not both, on the little circuit board have the usual sign of burnt-out, and smelt. (Wonder if ezrobot has any replacement of this little circuit board for the servo?)

</Quote>

JD has 16 servos (big and small), and Six has 12 servos, altogether I have 4 servos died (within a month). It is 14+%. (For Six alone it is 33+%.) As I'm "designing" the many different ways of "walking" around with the Six, I would expect more and more servos will die with the FET burnt, during testing and watching the way Six walks. Actually I have order more than 10 spares, it used up quite easily. In your way I have to order more, again, then. (And I restricted my kids to play around with JD so often, only once in a while without re-charge.)

NOT to see those dead servos sitting there without any help or repair, me asking if there is any replacement for the little circuit board without buying the whole servo?

If I were any other ordinary people who bought ezrobot for their kids, very likely it is boxed and put away after a servo or two died; in this way you would never know.

Regards, KC.

PRO
Synthiam
#16  

@kc, sounds like you're a pretty experienced power user! Keep up the exploring and learning.

Thanks @dave :)

#17  

@DJ,

I don't want to keep poking at this can of worms, but I have some ideas to consider....

And when you get more distribution channels set up worldwide so the shipping cost of small parts is not prohibitive, this may become a non issue.

  1. You could slightly increase the cost of Revolution robots but include a spare of each servo kind used in the robot, so if someone does have a burnout while learning, they don't need to spend more on shipping then the replacement servo costs and wait a week or more before getting back into it.

  2. Instead of automatically including it, pop up a suggestion in the online store when purchasing a revolution robot that the user might want to add a couple of spare servos as they are wear items and therefore not warrantied (or short warranty, not sure what your current policy is... I should go read it again).

2a) The idea of suggested purchases in the pop up actually made me think of a couple more. Things like when ordering a developer kit or a standalone EZ-B, suggest a battery, charger, and barrel to deans adapter cable. When buying a camera, suggest an extension cable. Maybe like Amazon does where you show "others who bought this also bought xxx"

Alan

PRO
Canada
#18  

I don't have the time to respond but I'd like to weigh in on this topic as well as I have a little more insight to add to the conversation. I'll try to respond tonight. The summary of it is that there's more than one way to damage an servo, stall conditions are the most prevalent but there are few other ways. This information may be quite helpful for KC.

Thanks Alan for your suggestions, keep those ideas coming :)

Hong Kong
#19  

@Jeremie,

I knew the ways which will blow a servo. (Please post them for others anyway.) I never had blown a servo before but NOT until recently. (I knew where is the panic switch.)

They blew just too easily. They are all dead in action (normal freely running action.) After the battery warning message, I can feel the heat from the servo. (Might be they need a heat sink.)

What my "concern" is I have 4 dead servos, and expecting a few more to come, what to do with them? Throwing them away is a waste, a set of shining metal gears, a powerful motor. Knowing that the FET blown in the little circuit, so I have to put them in a "dead" box and wait for some rescue.

Honestly speaking, the duty cycle of the servo need "some" improvement.

regards, KC.

#20  

Following up on KC's thoughts with what is probably a dumb idea.

If the servos are designed that the mosfet is the weak point, is there enough room (and appropriate parts available) that in a future hardware revision they could be a plug in rather than surface mount part so they can be easily replaced? Basically treat them like a smart replaceable fuse....

Alan

PRO
Canada
#21  

Hey guys, just hold out a few more hours and I'll post some info you may not have heard before. The answer is neither changing the duty cycle or making the servo electronics more modular. There are some best practices that I'll mention that may not be apparent when running metal-gear analog servo robots.

PRO
Synthiam
#22  

No, because the brushes on the motor short which is why the mosfet's blow.

If anyone feels they need more information to understanding the amount of testing on product development, please revisit my previous post.

Lastly, it's important to understand how a servo works, which can be found here: https://synthiam.com/Tutorials/Lesson/48?courseId=6

PRO
Canada
#23  

@DJ is correct that when the motor brushes short out the MOSFETs they see a zero ohm condition which damages them. This condition doesn't just damage the fets in our servos, for those that work with fets know, it would damage any fet. Now, you might be asking yourself "how do the motor brushes short out in the first place?".

We had a reviewer a while back that was unfamiliar with robots that used servos for locomotion. They were used to wheeled, H-Bridge driven robots. This person did not realize that they needed some pauses between their servo positions. The EZ-Script that they wrote caused the servos to vibrate extremely quickly back and forth, this caused damage to the motor brushes and in turn shorted out the MOSFETs. It didn't matter what kind of servos were in that robot that day, any servo with a brushed DC motor would not have lasted in that scenario. Brushed DC motors cannot keep instantly changing directions as it puts huge strain on the motor brushes and they can curl or break off. The best practice is to move a direction, pause, then move the other way. This limits the huge current spikes and crazy amounts of heat that can happen when instantly changing servo directions.

This is the reason that the ez-robot revolution example projects include many pauses, some may be small but any pause (sleep) helps. When building custom actions, creating a custom build, or doing any DIY design with servo motors it is important to keep this in mind.

That was pretty much all I really wanted to mention, but I had a couple more things "just in case".

I believe the majority of us here have probably been using servos for a long time. The difference though is that the majority of our experience has likely been with using plastic gear servos rather than metal gears. With plastic gears, it is the gears that fail first when placed in a high torque situation. This is not the case with metal gears. Metal gear servos will damage their own electronics (usually the MOSFETs go first) before the gears ever coming close to failing themselves.

Some conditions in which metal gear servos could damage the servo electronics are as follows:

  • Servos may not be calibrated [see here] or fine tuned [see here] and this could lead to servos running into each other during some actions

  • Servo output gear could be seeing some off-axis torque which could bind the gears. If a servo output gear is ever placed in a situation where it's being twisted in a direction that's different to it's regular plane of travel

  • If a robot is caught on an obstacle or object and is not freed from it in a timely fashion

  • Servo could be trying to lift too much weight, a custom build could be much heavier than a classic revolution robot

  • A foreign particle is introduced to the servo gearbox

Hope that helps guys!

#24  

Great info Jeremie. Thanks for the lesson. I bet I've destroyed a servo in each one of those examples you listed. LOL. eek

Hong Kong
#25  

@Jeremie, @DJ,

I don't mean to be rude. Don't think that the aforementioned servos, that I have, are died of "abuse". I have been using high torque servos with high stress usage in other applications. They won't die like these so easily. (I did see others burnt some of them, though, but NOT me.) Never mind, those servos are very expensive, compared to those sold by ezrobot. Say for example, a comparable Futaba High-voltage (7.4v) High Torque (~15kg/cm) costs around USD150, even a cheaper Hitec has a price tag of USD140. (Not to mention JR which has a price tag of more than USD200.)

One more thing I wanted to comment, as to "pause" and "delay". If "pause" and "delay" are NOT added to programming the robot servo movement, what would the robot behave?

My Six behaves like a lady and dances like breeze.

I could only accept the servos are died of my "heavy" use. (I would be very careful in the future. One charge in a while without re-charge.)

regards, KC.

Hong Kong
#26  

The kind of servos that I used to work with are those heavy duty and expensive ones. It's unfair to compare the ezrobot servo as such. If I were to do some "serious" work I'd better source some not-too- expensive but heavy duty ones to play with.

regards, KC.

Hong Kong
#27  

A bad news. I just burnt one more (5th) servo which is used to pan the camera from left to right. I can hear a very light "pop" sound during panning and then it stopped further response.

Upon opening the servo, one of the FET burnt. Please see the burnt FET from the jpeg file uploaded. User-inserted image

PRO
Synthiam
#28  

I would recommend revisiting the use of servos in your application to ensure they provide the correct requirements. Also, visit the software configuration to ensure the max and min limits are correct understood and configured. Lastly, ensure the servos are not being used wildly between directions back and forth, etc.

If you, specifically, are experiencing issues with servos, visit the application requirements.

Hong Kong
#29  

Thanks for your advice.

I think I have enough experience in using the servos, especially after I have "burnt" five. After replaced a new one, it works fine again. But I don't know when I will burn the sixth one on the other one, or the same one.

#30  

@DJ "You can revisit that thread to read my reply regarding that you cross-your-fingers that the mosfet burns out rather than a fire. Be happy the mosfet is damaged instead of the wires catching on fire, the battery exploding or worse..."

I use these robots with kids and am just learning about robotic- background more in software than hardware. I tried to follow this thread but some of it was way over my head. The thought of exploding servos makes me VERY nervous so I guess I need to go through the tutorials again.

I've purchased 3 robots in 2016 and had this happen to 3 servos, one pretty much out of the box. I thought it was just me but wanted to bring this to your attention since it may be an issue.

Anywho, One of the servos that died was the neck servo and I couldn't find that part on your website. The head is fine so how do I replace that servo on the neck?

Thanks a bunch for all the guidance! ;)

#31  

Never mind- found the microserver- amazing how much I accomplish when I actually read! Might want to put that in the body parts section for newbies like me!

PRO
USA
#32  

@kaycekwan I have read thru this thread and think the answer to the problem is to keep the servo from drawing to much current (DAH). If we put a polyfuse inline with the power lead to each servo, you could protect the servo from drawing to much current and burning out. Here is a link I found on the subject. I have had this happen once with a continuous rotation servo, so I plan to add these little resettable fuses to each of my servos. Just as a precaution.

I hope this helps.

RichardZ

PRO
USA
#33  

Sorry forgot to add the link. http://www.arxterra.com/servo-protection-methods-trade-off-study/

Hong Kong
#34  

@Richard,

I don't think it is a good enough fail-safe. (Have you tried that?) As all my servos are burnt under &quot;normal&quot; condition. I'm very careful in preventing the servos to stall or to take up too much load. The ezrobot servo is a &quot;heavy duty&quot; high torque (15kg/cm) and high voltage (7.4v) one. Nominally it will take up a high current say 2.5A when moving. Whatever the protect-current rating is used, either the FET in the servo burnt before the poly-fuse trip, or the poly-fuse do its job always.

To prevent the servo from burning, I'm now running an init (or startup) script setting all servos to a slower speed, when connecting to ezb before any &quot;Action&quot;. Delaying as much as possible, so the robot moves as &quot;smooth&quot; and as &quot;human&quot; as possible. Also preventing any &quot;too heavy&quot; load from occurring for too long at the shoulders of jd, the long arms of roli, etc. (Actually I did most of them already, but still they burnt one by one.)

For the long run, I'll source other better (more tolerant to inductance surge) servos, either core-less or brush-less. There are NOT too expensive digital servos, with similar specification, in the market that I can try, the only problem left is the dimension onto adapting the clip-n-play bracket. Anyway I'll get one or two to try out. (I bought more than 10 spare ezrobot servos which originally planned to add more &quot;feature&quot; to my robots, but now I have to keep them as really spare.)

regards, KC.

#35  

Along with setting the servo speeds in the INT script, also set your Min, Max positions for the servos. Foe example:


##Specify the ranges for the servos
##this prevents the servos from ever going past these limits

# Rt side_to_side Wrist
SetServoMin(2.D0, 120)
SetServoMax(2.D0, 165)

# Right Claw
SetServoMin(2.D3, 74)
SetServoMax(2.D3, 102)

# Rt Up_Dn Wrist Servo
SetServoMin(2.D1, 20)
SetServoMax(2.D1, 170)

# Bubble Blocks
SetServoMin(D15, 85)
SetServoMax(D15, 177)

# Bubble Servo
SetServoMin(D1, 45)
SetServoMax(D1, 97)

# Rt Wrist Rest Servo
#SetServoMin(2.D2, 93)
#SetServoMax(2.D2, 139)

Hong Kong
#36  

@Dave,

Thanks.

I'll adopt these min/max setting.

regards, KC.

PRO
USA
#37  

Good info, I think I will make the same program changes as well.

I 3D printed a JD and used some Tower Pro MG996R servos I had used for a hexapod robot I built in the past. Only to find out that the servos turn the opposite direction as the EZ-Robot servos. Just a heads up when using other manufactures servos.

RichardZ

#38  

There's also a box inside of the servo control that you can check that will reverse the circles easily and simply.

PRO
USA
#39  

@Dave Thanks Dave, I looked for a check box that would reverse it in the JD Project but could not find it. I assume I need to setup a servo profile in the project . Is that correct?

#40  

No, all you need to do is install a servo control. Then for the servo you want to reverse just open the control by clicking on the gear symbol and assign the port and ezb board to that servo. In that same window you should see a box to check for reversing this servo. I'm not sure if you can do the same thing from a servo profile. Good luck and have fun.

Australia
#41   — Edited

@ DJSures - The ez-robot servos have unbranded Mosfets in them.

And this.. "The servos are great still - only people with an issue will post on the forum. So, the one user per month with issues is pretty good" is not true. Unbranded Mosfets are a curse in the electronics and robotics realm and I am extremely disappointed that you would support them as there are branded Mosfet servos available for the same price as yours.

Since two of my servos have died, I replaced them with branded Mosfet servos and regardless of my extra long arm have had no problems what so ever even though V9 has been rocking me out to one of her favorite trance tracks for the last month -video upon request! There's no excuse for a servo to freak out over a three servo stack or a three-second jam.

Dealing with China successfully is a road filled with landmines and I implore ezrobot to seek out a better and branded Mosfet deal as soon as you can. Seems to me you have a big reputation to keep and that will be hard to do with crappy servos.