I have cooked one of these guys before running it off 7.2V volts, which according to Hitec, it fine. Not even close.
So this time I was using them on a small arm type setup and only gave them 6 volts. Thought they should be fine. Nope.
I had no load on the arm and was just testing some basic movements with 3 brand new HSR-5995’s and after 5 minutes, poof, smoke and sizzle from the elbow servo.
I know Ive heard alot of complaints abut this servo not living up to Hitec’s published specs on running it on 7.2 Volts. But 6 Volts? Come on. I can’t believe this.
I send an email to Hitec, like many others have, asking that they replace or repair my servo for free since it was being used within spec and under a minimal load. I have a feeling my requests that they rethink their published specs for this servo, will fall on deaf ears. But for $140 a pop, I want a servo that can actually work without starting on fire.
Can anyone think of a comparable servo in a standard case? I will try the HS-5955TG’s and hopefully I have better luck. I was really looking for something in the 400+ oz/in range though.
I would swear that Hitec just jacked up the voltage rating on these as an excuse to post higher specs. Because these servo do not work on 7.0+ V. And barely work on 6V.
Have you done any test to see what the actual current draw is at stall and the current draw at the rated torque? The servo maker probably needs to publish a duty rating if the servos will fail if operated at max current for any length of time. At $140 a servo, I’d be fusing the power to the servos to see if that would protect them for the short term.
I just noticed lynxmotion has them as well. For cheaper too.
Which do you think would be better, the 5990TG or the 5955TG. They look to be the same other than the feedback thing with the 5990. I don’t really need that.
I don’t know if the below chip is the same as the one that smoked, but if it is, then the chip appeares to be under rated for its application in the servo. Might be worth while to see if you can identify the smoked chip and replace it in the servo to save that $140.
I’ve already bought 2 of the 5990TG, and plan to use around 20 of them… I hope they don’t have that problem, or I’m going to have a $3000 pile of junk.
Another data sheet. Looks like the chip can be pulsed at 14A but continous current max rating is 3.5A. Given the servo data below, under continous heavy loads the chips might well eventually overheat.
Current Drain (6.0V): 300mA/idle and 4.2 amps at lock/stall
Current Drain (7.4V): 380mA/idle and 5.2 amps at lock/stall
I recently had an issue with Kondo 10.8V servo’s blowing FET’s. On one of the servo’s changing the FET’s cured the issue and the servo is stilling running happily. One of the others though didnt like the fix and just blew the replacement.
When I investigated further I found that the motor itself was at fault. The higher current being drawn had somehow damaged insulation inside the motor causing a dead short.
It would seem that the motor wasn’t upgraded when the circuitry was. Might this be a similar issue?
The numbers on the chips are 9958, AC, w341.
The only mosfet I find searching with these numbers is the Vishay mosfet but I don’t know if this is the same chip. datasheetcatalog.com/datashe … 06DY.shtml
I’ve read on other forums of lots of people having trouble with the HSR-5995TG’s burning up on 7.2Volts. I simply don’t believe that they can actually handle it. And it seems they can barely handle 6 volts. I think I’m going to try the HS-5990 out. I just hope they did more than add a heatsink to the case. I hope they actually upgraded the chips to something that can actually handle the amp draw they claim the old ones could.
The story on the 5995 is, it was designed as a Robo-One servo, and it was never intended to be run for longer than 3 minutes at a time, before a cool down was required. I want to let you all know that Lynxmotion was the only sales channel of the HSR-5995TG to actually mention this duty cycle limitation in the first product information screen. (This is where you usually talk the product up, not mention it’s drawbacks.) I have always put the people and the robots above the bottom line… If I were you, I would request that they allow you to upgrade to the HSR-5990TG and offer to pay the difference. I do not know if this will be accepted… The HSR-5990TG has one feature no one has discussed here. It will shut down if the heat builds up too high. Can you say “sentient servo”? I only mean it will protect it’s own life if doing too much work. I don’t want to turn this into a discussion of what sentient really means.
In addition to having more protection for overheating the 5990’s case is a heatsink. The 5955 has the aluminum case too but no overheating protection. Continuous use is certainly possible, but it depends on how much work the servo is doing. Unfortunately I don’t think ANY hobby servo can do continuous operation at anywhere close to stall conditions.
Thanks for clearing that up Jim. I too have other alternatives for servos. Which I guess is one of the drawbacks to building something around the most powerful hobby servos available. There is no upgrade.
I’ll put an order in for 1 or 2 of the HS-5990tg’s soon then. I just need to be sure there are no other alternatives.
I still have a question about the 5990TG which I think is important. Did Hitec actually replace the mosfet that was cooking with one that can handle the current draw at stall torque? I understand that it will shut down if the heat gets too high, but can it actually handle more load before shutting down?
There is actually more to this than just the Rds-on rating of the MOSFET. A lot of power can be dissipated by the rate at which the MOSFET is turned on and off. Inadequate gate drive voltage will prevent the MOSFET from reaching the specified Rds-on value. Lastly the thermal coeff of the MOSFET package to the PCB will determine what type of de-rated specs the MOSFET really will find itself operating under. In short it is possible to increase the power handling capabilities of a servo at a simplistic level by upgrading the MOSFETs to a lower Rds-on part, however in doing so you change the gate load characteristics seen by the driver which will result in less of a performance gain than anticipated. Adjusting the driving circuit allows you to get more of that performance. Also decreasing the thermal junction of the package to ambient will increase the MOSFET survivability. I guess my point is that just looking at the MOSFETs will not guarantee the electronics can tolerate abuse better.
A minor hijack I think but here’s a thought I had about all this. With the bottom cover of the servo removed, if the MOSFETs are on the visible side of the PCBA then it might be possible to machine a replacement cover from say aluminum. You could then leave material on the inside of the cover to come within a couple ten-thousandths over the MOSFETs and use a conforming thermal pad material to bridge the gap between the two surfaces. We do something similar on a small brushless motor drive and we can run Si4559 MOSFETs at 12-15W just because we can get the power off of the parts (sealed motor is immersed in a moving fluid, ie an awesome heatsink.) I am not suggesting this is a solution for any normal situation with these servos but for the guys with access to CNC machines or a machine shop it might allow for much more robust performance, especially if you can put a threaded pivot point on the outside of that same cover… similar to how the new servos are arranged.
From experience of a number of users of the HSR-8498HB on the RoboNova, where MOSFETs have burnt out and been replaced. The failures on these servos were initally motor failures which then popped the MOSFET. The motor was measured to be a dead short in most cases, so replacing the MOSFET did not help. Thankfully Hitec have been very good at replacement, if a little slow
The MOSFET do get very hot very quickly, but appear to be amazingly resilient.
Try measuring the motor resistance as the first step.
The HSR-5990-TG is running with IRF7389 Mosfet. Also if the primary failure is in the motor, then the heatsink should be major improvement too for the new servo.
Also do get in the habit of lifting the robot to check if any of the servos are stalled. This may not be just due to vertical load, but also due to friction or blockage. I did not check the 5990 yet, but on the 8489, when outside the deadband (+/- 4 microsec), then the PWM and hence motor current rises from zero to maximum over a range of about 100microsec, which I think equates to about 10 degrees.
Over HMI on the 5990 you can read both the applied current and also check the actual versus desired position yourself, before the protection cuts in.
I’m still not completely clear on the capabilities of the servo. I had kinda assumed that since it’s $125, has titanium gears, gobs of torque, rated for 7.4V, heat sink, robot designation, spline on bottom, etc… that it is the most robust servo on the market. And to me, “robust” means you can actually use it for 15 - 20 minutes without worrying about it shutting down or catching on fire.
Basically, what I really need to know is whether a large SES robot like this (or slightly bigger):
…using these 5990TG servos, and run from a 7.4V lipo battery, can actually be played with for the duration of a full charge on a pack of about 4000mah. If this is not the case, then my assumptions, which were based on the specs, are invalid, and I should probably not be making a multi-thousand dollar investment in such a project.
Powerful, expensive, looks real good on paper and achieves some spectacular results but after the first few minutes it has to coast for a couple of million miles.
The Kondo servo’s come with the same recommendation, 20 mins max which is about the battery life then cool them down while the batteries recharge…
If 3 minutes is the number… not so much. 