Fried Servo
Fried Servo
by subpilot
Wed Jul 12, 2006 2:09 am
I had one of my servos go up in smoke and found that it had blown one of the mosfet ICs. The rev 0.1 servo spec has them listed as a Vishay Si9958 but my servo had IRF7389s in it. Although they have better specs than the Vishay part, I wonder if the blown servos others have seen used the IR parts?
The servo wasn't under load when it failed but I had been running it for awhile. It's typical for a Mosfet to fail shorted and burn things up and there is no protection to keep it from happening. Wonder if it is a problem with the drive circuitry or lack of sufficient heatsinking? Would probably be a good idea to add a fuse to the board to prevent fires etc.
The servo wasn't under load when it failed but I had been running it for awhile. It's typical for a Mosfet to fail shorted and burn things up and there is no protection to keep it from happening. Wonder if it is a problem with the drive circuitry or lack of sufficient heatsinking? Would probably be a good idea to add a fuse to the board to prevent fires etc.
by Bullit
Thu Jul 27, 2006 11:58 pm
I've had 5 servo's fail so far. All of them one of the h-bridge's failed. I tried replacing a couple but they didn't fix it. I think the motors cooked. They nolonger rotated easily. Once the h-bridge was replaced the same part heated up once power was applied again. I thought these things were suppossed to be current limited anyway. All of my servos failed under normal operation with the stock battery and they were not under excessive load. I think Hitec must have a quiality control problem on the motors.
by bauermech
Fri Jul 28, 2006 11:05 pm
The rev 0.1 servo spec has them listed as a Vishay Si9958 but my servo had IRF7389s in it.
Fortunately, I have yet to burn up a servo... stripped a few gears, but that's about it so far.
However, I looked under the hood of some of the servos I've been using...
All the HSR-8498HB's I've checked sure enough have the 7389's as you said.
The HSR-5995TG has the 9958's.
Hmmmm. Makes me wonder. I could be way off track, but I heard that the reason the 5995's don't support feedback/catch-n-play, was mainly due to problems in their H-bridge. ...sorry, just thinking out loud
by subpilot
Sat Jul 29, 2006 2:52 am
Looking at the design I don't see anything wrong with the drivers. Seems that the motor is the weak link. If it weren't for that, running at 7.5V would be no problem. Wonder of there's a better replacement motor? But then we would probably be seeing more stripped gears.
by i-Bot
Sat Jul 29, 2006 12:10 pm
First, you are correct the IRF7389, seems to be the MOSFET part normally used in the HSR8489. The Si9958 is used in all the other Hitec digitals I have seen, including the HSR5995.
Sorry for the confusion, I will update the spec, and add the IR data sheet.
Second, the reason the catch and play does not work on the 5995 is due to the firmware. The processor does not turn off the H bridge, nor does the position feedback protocol work in the firmware. The h-bridge appears identical configuration except for the devices used.
The motor on the 5995 has much lower DC resistance at 1.6 ohm compared to the 8498 at 4.5 ohm, also the 5995 is spec. at 7.4 V compared to 6. The 5995 motor is coreless, and probably of higher quality.
I did look at the heatsinking of the devices, and at first sight the look very similar, neither is very good. Also notice that the major heatsinking is to the motor terminals, which may put more heat in than out !
Since the motor is not moving smoothly, I agree the pointer is to the motor as the source. Did you measure the DC resistance of the failed motor ?
On the topic of current limiting, there is no current limiting in the 8498. They do say current feedback on the box, but there is not circuit in the servo to do this, and Hitec have not responded to requests to confirm that this too does not work as advertised. Current feedback, or maybe better temperature monitoring may be able to save the motor. It may be worth trying a fuse or polyfuse, but this may save the h-bridge, but maybe not protect the motor.
I feel it would be sensible for Hitec to quickly and quietly replace all failed servos. The HSR are robotic servos, and Robonova is a biped, and these are tough on servos, then Hitec know this, so they must be fit for purpose.
Given the poor technical support, and the non arrival of promised product upgrades, Hitec need to maintain or recover their reputation in the robotics area. This reputation had been built on the quality and reliability of their servos. Please keep us all updated on their responses.
Sorry for the confusion, I will update the spec, and add the IR data sheet.
Second, the reason the catch and play does not work on the 5995 is due to the firmware. The processor does not turn off the H bridge, nor does the position feedback protocol work in the firmware. The h-bridge appears identical configuration except for the devices used.
The motor on the 5995 has much lower DC resistance at 1.6 ohm compared to the 8498 at 4.5 ohm, also the 5995 is spec. at 7.4 V compared to 6. The 5995 motor is coreless, and probably of higher quality.
I did look at the heatsinking of the devices, and at first sight the look very similar, neither is very good. Also notice that the major heatsinking is to the motor terminals, which may put more heat in than out !
Since the motor is not moving smoothly, I agree the pointer is to the motor as the source. Did you measure the DC resistance of the failed motor ?
On the topic of current limiting, there is no current limiting in the 8498. They do say current feedback on the box, but there is not circuit in the servo to do this, and Hitec have not responded to requests to confirm that this too does not work as advertised. Current feedback, or maybe better temperature monitoring may be able to save the motor. It may be worth trying a fuse or polyfuse, but this may save the h-bridge, but maybe not protect the motor.
I feel it would be sensible for Hitec to quickly and quietly replace all failed servos. The HSR are robotic servos, and Robonova is a biped, and these are tough on servos, then Hitec know this, so they must be fit for purpose.
Given the poor technical support, and the non arrival of promised product upgrades, Hitec need to maintain or recover their reputation in the robotics area. This reputation had been built on the quality and reliability of their servos. Please keep us all updated on their responses.
Twice Blown Servo Left Leg Port 2 servo below kneecap ...
by CaptKill4Fun
Sat Aug 05, 2006 4:04 pm
...
Last edited by CaptKill4Fun on Sun Aug 06, 2006 7:24 am, edited 1 time in total.
by beermat
Sun Aug 06, 2006 5:50 am
Hi,I've built a prototype replacement motor for the HSR-8498HB servo, which in theory should ease some of the issues users have been facing. It can, in theory, detect overload situtations, reduce the current draw on the servo and activate the MR-C3024, in theory, to initiate compensatory motion on opposing servos to further reduce the load on that servo in much the same way, in theory, that a gyro can do.
I have currently assembled the prototype motor out of the brown paper wrapping my Playboy magazine is delivered in, and it *looks* great. If I can get enough orders, and find a manufacturer willing to fabricate these for around half the cost of the standard RN servo, I will produce them en masse.
______________
C U L8r, Beermat ...
"Let me design the motors
.....that will enhance your products ..."
by subpilot
Sun Aug 06, 2006 6:34 am
I suspect the motors short out due to an internal problem not related to load. A few of the motors were under no load when they fail. If the motor shorts they draw lots of current which = smoke & heat till they pop a driver FET. You have to open the back side of the servo and measure across the motor leads to see if it's shorted. I would have sent my failed servo back by now but I'm currently a bit far away from a mailbox.
Speaking of smoke, I detect a flame war smoldering.
Speaking of smoke, I detect a flame war smoldering.