I am curious about hydraulics to replace the 32 Hitec HS-7950TH servos in my full scale humanoid robot. The servos are way too loud, catch fire and have absolutely no holding brakes despite my best efforts to invent them.
I have some questions.
What mechanism is used to propel the Dodeca 12 in 1 Solar Hydraulic Robot?
Does Hydraulic fluid increase the holding force against massive weight strain to reduce jitter and servos catching fire?
Are worm drives required at the power bank of servos that push the syringes in and out to act as holding brakes or would another (much faster) system work well enough such as racks and pinions?
How long can tubes be before they lose their effectiveness? Do they collapse/burst too easily or kink or have other issues, and how are they attached securely to syringes? And where does one buy such supplies?
What can I buy that hydraulically pan-rotates in each direction (infinitely or at least 180 degrees) based on fluid direction flow? Tilt seems easier than pan by a long shot. Iâd at least expect to find a round plunger of some sort, or a clever linear bevel system.
Hydraulics, even with all that water in tubes seems a lot lighter than servo motors and wires. Is it?
If a servo motor has specific speed/torque specs, will these remain the same at the end of the line where the syringe pushes and pulls?
8.If my servo rotation is limited to 180 degrees, I can push something hydraulically 360 degrees at half the speed of the servo. Is my maths right?
Itâs the equivalent of syringes filled with fluid, and as one empties, the other fills.
Hydraulics is a completely different system than servos and itâs hard to compare the two. The Owi kit uses microhydraulics, whereas more powerful robots use (for lack of a better explanation) ârealâ hydraulics because of the forces involved. Hydraulic force is linear whereas a servoâs is rotational. To compare the two you need to understand the concept of torque (force times distance). Each hydraulic actuator would still need an actuator. Small hydraulics = small forces. Note that the force you get out will be slightly less than the force you put in.
Owi makes a âhydraulicâ robot arm which you can play with to see the results: https://www.robotshop.com/en/owi-hydraulic-arm-edge.html
Worm drives are not required. If you look at Owiâs system on the arm, you can push fluid into the cylinders however you want - even servos would work well. However, whatever you use should ideally have position feedback to prevent the motor from âover-pushingâ or âover-pullingâ.
Hard question to answer - in theory, the tube can be quite long without any losses since youâre using force to fill and empty cylinders.
RC servos seem easiest, but youâll really need to compare the rotational torque which the servo can provide to using it to actuate a hydraulic cylinder. You may end up getting significantly less force out of a hydraulic system which uses the RC servo than the RC servo produces by itself.
Normally not, but that doesnât mean itâs not possible to create a system which is lighter. See the section below which better explains âproperâ hydraulics.
Too many factors at play to give an easy answer here. You will really need to understand the concepts of torque and converting pressure to a force. Do the exercise on ONE joint.
Not sure what you mean given that the hydraulic force you are referring to is linear rather than rotational.
So, a bit about hydraulics: there are many reasons why you donât see hydraulics being used in almost any small to medium sized robot, let alone industrial robot arms - they are complex, expensive, prone to leaking and very heavy. A basic hydraulic system would consist of:
Reservoir (contains enough fluid to fill all pumps, and a bit to spare_
DC pump (powerful enough to fill several hydraulic cylinders at once without significant losses in force to each)
Hydraulic cylinders, each with their own actuated valves
Battery pack to power the pump
The pump alone is normally really heavy, making a hydraulic system quite impractical for a robot which needs to be lightweight. The requirements of adding a liquid reservoir and a separate battery make it even less practical. Since each cylinder still needs either a solenoid driven valve or its own DC type actuator, the complexity of the electronics just gets worse.
A full scale exoskeleton system which is used to provide very high forces may have to use a hydraulic system, but most still use DC motors, and the more expensive ones use âtorque motorsâ.
Thanks for all your time and effort, but given youâve seen my movie and know I make full scale humanoids using 32 Hitec 7950TH servos that cost $180 each and still catch fire from the strain, its strange that you would make reference to $50 785HB servos with vastly inferior power. You can immediately assume each of my robots cost at least $10,000 in parts based on the sample information in my first sentence, so Iâm CLEARLY not a beginner. Also, your first answer to what âpropelsâ the hydraulic Dodeca robot was that syringes send water to one another. Its insulting to be told that I need to learn the basics at this stage, given my achievements. Your answer to question three about worm drives not being required because you can power a robots joints with your hands in inappropriate, and you totally missed the point on my line of questions relating to pan rotation using hydraulics - seemingly because youâve never considered it before. It seems that you donât understand the tens of thousands of ways that fail in the construction of full size humanoids - the reason Iâm considering hydraulic tubes stemming from a bank of servos. I can see that you went to a lot of effort in trying to help, so for beginners there may be some useful information there. But for me, Iâm about ten years ahead of this information.
Apologies - the response was only intended to help, not to insult. No, had not in fact caught that you were also the one who had posted the movie. The HS-7950TH can provide up to ~30Kg-cm of torque (stall) whereas a 5:1 geared 785HB can provide ~66Kg-cm. Nevertheless, if the answers provided have brought forth this reaction, I am happy to step back and see what the community suggests; there are many here who have successfully built full-size humanoid torsos as well as large scale (though not 1:1 humanoid robots). Best of success.
Hi Zxen, I am new to this forum, and my interest is in applying hydraulics to robots. For a customer, I developed a light-weight, low power draw, high pressure (1000 PSI) positive displacement, zero leakage, precision control hydraulic power unit (HPU). Size, weight, speed and pressure could all be modified for any application. I think my design would work nicely in robots, and I donât see anything like it anywhere. I am looking for applications, but need to see if the economics would be feasible. If the numbers work out, I think the technology would do the trick.
@ctyankee Do you have an idea of the overall size? Are you thinking about mobile robots or stationary? What parts are new? Normally hydraulics are used in robotic applications where very high force (or torque) is needed, which outweighs the added complexity. Consider the robots which took part in the megabot challenge were hydraulic, and were absolutely riddled with problems, required constant maintenance etc. etc.
Hi Cbenson, the design I made for the application I had turned out to be 18" long and 2.5 inch in diameter (roughly). This is including a reservoir, and my design was a master/slave arrangement. I can make it smaller depending on what is needed, very small if needed, but along with that goes volume and pressure - so it is a juggling act. It could be any kind of robot, it is very low power draw compared to ânormalâ HPUâs, but it does involve sensors, CPUâs and other controlâs to make it work. I am not into robots, so I am not familiar with what they use. Pointing me to a link with other hydraulic would be helpful. I looked at the Megabot site, but I want to see the guts of other hydraulic robots. Thanks.
It looks and sounds expensive and appears to be a prototype since it says it has a 3D printed titanium housing. I still havenât found anything like my design, or at least my concept (redesigned for specific robot needs).
