Soft Hydraulic Muscles Lift Weights As A Team

Hydraulic artificial muscles

Working with hydraulics usually means having a fluid tank and valves. [consciousflesh] does away with both those for his DIY hydraulic artificial muscles. Instead, he uses a pair of muscles, both preloaded with fluid. To contract one, he pumps the fluid into the other, expanding that one, and vice versa. A bidirectional gear pump moves the fluid while also acting as a valve. And flexible materials replace heavy metal cylinders.

As we said, this is a DIY project. He made the muscles by surrounding silicone tubes with aramid fiber sleeves, giving added strength. The blocks at either end are also custom-made. The gear pump is one he purchased and made substantial modifications to, including removing the tank and fixing a brushless DC motor to one end. The final custom piece was a controller board for the motor. A Gerber file, schematic, and technical drawings, along with further details are all on his page. Meanwhile, check out the load test in the video below as the muscles lift and lower 5 kg (11 lbs) each.

A search of Hackaday shows hydraulic artificial muscles may be rare, so perhaps this will be the first of many. For example, how about replicating how human arm muscles work together, one contracting while the other expands? We’ve seen that done already using pneumatics with [James Hobson’s] exoskeleton arms. Perhaps someone should do it with these pairs of flexible hydraulic muscles?

Our thanks to [starhawk] for tipping us off about this one.

17 thoughts on “Soft Hydraulic Muscles Lift Weights As A Team

  1. Let mankind become divided into a human revolution lol
    But on a serious note, this seems a low powered option to create exo skeletons, if someone could create a device to recover a lot of the used energy in this system, it could result in Elysium, edge of tomorrow etc style powered exo suits becoming a reality on current battery tech.

  2. After reading the description, I was kind of surprised that the demonstration did NOT use opposing “muscles” acting on a lever.

    Something to note here, is that these will not act well in extension, since there’s nothing to constrain the motion to linear. In fact, in the demonstration, all of the work is being done in retraction, when it’s lifting the weight. But in retraction, it is limited to atmospheric pressure, since you can’t draw negative pressure. However, this can be improved by operating the muscles in a high-pressure environment, for example by enclosing the whole mechanism in a sort of sausage skin (also wrapped in a fiber mesh) that holds pressurized fluid. This provides higher “atmospheric” pressure, giving more force in retraction.

    1. This will probably come off a little more aggressive than intended, but…have you ever seen hydraulic/pneumatic muscles before? It’s not an issue of them being “weak” in terms of producing force on extension, by their design, they can’t exert _any_ force when extending, regardless of what the atmospheric pressure may be. They’re totally…flaccid…when being drained

      I’m assuming when you wrote “But in retraction, it is limited to atmospheric pressure […]” you meant “in extension”. If you didn’t, you need to re-read the write-up itself. They aren’t limited by ambient pressure in any way; rather, their limitation is almost exclusively the tensile strength of the sleeving material.

      1. No, I mean in retraction. This is because you can’t have pressure lower than zero, so when you’re pumping fluid out of the muscle, atmospheric pressure is what is actually doing all of the work. This is hydraulics 101. You can’t suck hydraulics.

      2. Oh, but in in answer to your question, no, I haven’t seen hydraulic muscles before, and that’s pretty much because of the limitations I spell out.

        Also, they are not flaccid when being drained, because if you actively draw fluid out of the muscle, i.e., apply a pressure less than ambient, the atmospheric pressure compresses the tubing.

        And while we’re at it, “they can’t exert _any_ force when extending” is nonsense. If you pressurize a flexible tube, this will counter any external force that reduces its volume, so it WILL tend to straighten out as much as it can, since its maximum volume is when it is straight. This is how Bourdon tube pressure gauges work: there’s a bent tube, and as pressure increases, this applies force that tends to straighten it out, and this motion is used to turn a gear to rotate a pointer.

        1. These “muscles” differ because of the aramid (kevlar) sheathing. When you apply pressure, they are forced to expand laterally and the cross-woven fabric is forced to contract in axial direction. Therefore the pressurized state is the contracted and there is no limitation by surrounding (atmospheric) pressure.

        2. It seems you are, indeed, completely mistaken about how these work, then. They do not retract (produce tension) when being evacuated. They retract when being PRESSURIZED; the weave of the sleeving, when forced to increase its diameter, reacts by decreasing its length. The idea isn’t to inflate them to increase their length and draw vacuum to decrease it, the idea is to cause the tube inside the sleeve to balloon out radially, pushing the sleeve out with it.

          It cannot produce any pushing force because there’s no way for it to do so. If you pressurize the tube,the assembly retracts. If you then start to evacuate the tube, all it does is relieve the tension on the sleeve and the assembly just stops producing tension. It will only return to its extended state when pulled by an external force. The atmosphere cannot push on the sleeve because it’s just a permeable weave of fibers. Even if it could, it wouldn’t matter; as they are both highly flexible, the tube and sleeve would just bend under any kind of resistance, unable to hold even its own weight.

          1. Okay, thank you for clearing this up. I hadn’t read the page for this, just this article, where the author says, “To contract one, he pumps the fluid into the other, expanding that one, and vice versa.” I interpredted “contract” to mean reducing in length, not in diameter.

  3. Why bother with hydraulics if you are going to move an identical weight an identical distance? You could just use a balanced lever. It would be cool seeing something done with servo hydraulics but the control valves tend to be pricey and as are the proper actuators. Cool stuff to play with though.

    1. Running the muscle with symmetric load was only a first test . Please check my youtube channel or visit project’s page at – I’ve added a video of the test with asymmetric and much heavier load.

  4. Biomimetically cool as these are, do they make sense as anything more than a Cronenbergesque art piece?

    I mean, if you’re going to have a motor, controller, and power supply anyway, why burden the whole thing with a pump, plumbing, fluid, ‘muscles’ and all the associated nonlinearity, lag, inefficiency and control loop issues? Why not just use the motor directly?

    If you want it to “behave” like a spring-like muscle, or have some desired force-direction characteristic, then use the force feedback and encoder position reporting (that you want to have anyway) and just program the desired response. Then you can even change it as the task requires.

    Remote hydraulic motors can make sense if your power pack is heavy or expensive (say, a diesel engine) and you have multiple motors, like a backhoe. It makes a lot less sense when you can have multiple cheap, light, efficient high-performance motors at the point of use.

    1. My idea behind the hydraulic muscles was to replace the heavy and expensive gearing needed with all the direct motor drives. Making a reliable , high torque harmonic drive at home is impossible , while the hydraulic muscle can be made with some basic tools. Even the gear pump can be made if one has the access to the lathe. The hydraulic muscles have one more advantage over direct motor drive : they can be installed along the “bones” of the robot or exoskeleton , while the pump and motors can be installed in a more convenient place , for example on the torso .

  5. I wonder, has anyone tried to use the “self extending” irrigation hoses that you see on shopping channels as hydraulic muscles? They are usually biased to retract with a spring, so the strength would be limited, but they would be fairly easy to source.

  6. pneumatic muscles are cool and the company I work for has produces them since about 20 years(that company with the flying penguin, smartbird, air ray, aqua ray, …). But pneumatic muscles are nothing more than a nice toy. The problem that they have is, that the rubber will crack after some time. I they are in permanent use it is not such a big problem, since the rubber stays warm and flexible. If they are only used occasionally they will crack much faster.
    A hydraulic version seems like a very bad idea.
    1. you can not use the high pressure advantage, that hydraulic has over pneumatic
    2. the oil will probably cool the muscle, so it will break faster
    3. the muscle will crack and oil will spill out

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.