. @alankerlin Hydraulics are usually closed, but that adds mass vs short acting open systems. F9 fins only work for 4 mins. We were ~10% off.
They dump the fluid. Saves weight and reduces complexity.Other rockets which use open hydraulics:1) Delta III2) ConestogaI'm certain there are many, many others. I believe the majority of hydraulic systems for rockets are open, though verifying that is going to be difficult.
They dump the fluid. Saves weight and reduces complexity.
Quote from: Robotbeat on 01/11/2015 12:56 amThey dump the fluid. Saves weight and reduces complexity.That was my initial guess, but are you sure? Source?
Quote from: Kabloona on 01/11/2015 01:21 amQuote from: Robotbeat on 01/11/2015 12:56 amThey dump the fluid. Saves weight and reduces complexity.That was my initial guess, but are you sure? Source?Is Elon good enough?
Quote from: Excession on 01/11/2015 01:33 amQuote from: Kabloona on 01/11/2015 01:21 amQuote from: Robotbeat on 01/11/2015 12:56 amThey dump the fluid. Saves weight and reduces complexity.That was my initial guess, but are you sure? Source?Is Elon good enough?I don't know how they get away with that. If we spill a pint of hydraulic fluid at work, the next thing is a full-up environmental remediation process.Maybe they do use one of the bio fluids.
I don't know how they get away with that. If we spill a pint of hydraulic fluid at work, the next thing is a full-up environmental remediation process.
As to this talk of open vs. closed hydraulics:There are actually many variants, so it's important to get the context correct.Open 'loop' (hydraulics) is the fluid coming from a reservoir (typically vented to atmosphere, but it could be pressurized, typically with air for low pressures or Nitrogen for higher pressures and to keep the fluid 'dry') and is then delivered via a pumping device to deliver flow to the control valves and/or actuators. The returning fluid is fed back to the reservoir to be re-used.Closed 'loop' (hydraulics) is the fluid kept in the loop, typically between a pump & actuator (typically a motor). It is only when the actuator is moved that the fluid in the loop moves within that loop. Of course typical hydraulic closed loop systems utilize a small pump to dump oil into the lower pressure of the two sides of the loop (after the work is done, from the motor for instance). And for what is allowed into the loop has to have an equal amount dumped out, and is usually handled by a valve, and is typically cooled & filtered. Properly designed, just the cooler itself could be used to hold the fluid required due to expansion & contraction.Now the rocket would have to have a closed 'system' (so no fluid escapes), but it also must have either a pump driven by some means to deliver this flow, or it has an accumulator to store the fluid under pressure (think of a water pump & tank on a well system in a house). The tank (accumulator) has to have a pressurant gas (Nitrogen typically for hydraulics to prevent explosion due to hydrocarbons present) at a pressure just under that required to operate the system's actuators. Now if there is no pump in the system (external turbopump or similar driven off a shaft or via hot gases, or an electric pump with a battery), you have to rely totally on the size of the accumulator (tank) to supply that fluid. This fluid typically passes through a pressure reducing valve to drop the pressure to somthing more usable, plus it allows a higher charge pressure in the accumulator to lengthen the amount of time to operate the hydraulics. Once the pressure drops off below the setting of the reducing valve & the minimum that is required to run the actuators, they cease to work effectively.Although it adds weight, the simplest way to boost the amount of time available to operate hydraulics in this manner (if it is indeed how the system works) is to add more accumulators, or make the existing one(s) larger. Increasing the pressure in the accumulator is possible (gas side) but it requires a stronger tank, and likely a new pressure reducing valve that can handle the higher initial pressure.Now they can also have the accumulators charged with Helium, and with the propellant tanks also using this gas, if they had an auxilliary bottle on board, they could supplement it with that.
Quote from: Lee Jay on 01/11/2015 01:36 amQuote from: Excession on 01/11/2015 01:33 amQuote from: Kabloona on 01/11/2015 01:21 amQuote from: Robotbeat on 01/11/2015 12:56 amThey dump the fluid. Saves weight and reduces complexity.That was my initial guess, but are you sure? Source?Is Elon good enough?I don't know how they get away with that. If we spill a pint of hydraulic fluid at work, the next thing is a full-up environmental remediation process.Maybe they do use one of the bio fluids.Toxic Substances Portal - Hydraulic Fluids Issue is usually organophosphates - often additives. Concentration and substantial contact matters.You can use a variety of common substances too. Pure mineral oil would be a likely one.So not necessarily that its the hydraulic fluid itself, but an additive that allows for longer term use in a closed system, to decrease the breakdown of the working fluid. If its an open system, there's no reuse, thus breakdown doesn't matter, thus no need for additives.
I don't know how they get away with that. If we spill a pint of hydraulic fluid at work, the next thing is a full-up environmental remediation process.Maybe they do use one of the bio fluids.
The main reason mineral-based hydraulic fluids are toxic are the heavy metals that are added (zinc, among others) to aid in anti-foaming, anti-oxidation, viscosity improver...;hence the term 'additives'.If the system is essentially one-time use, or controlled (temperature, ETC) you could easily get away with an hydraulic fluid that is environmentally 'sensitive'. It's not 100% biodegradable, but does meet many environmental regulations. Our company uses a product called Environ (made by Petro Canada).
I'm still surprised the grid fins aren't electromechanically actuated.
So, why on earth would they vent it overboard when they can contain it all within the system at effectively zero added weight?
You may have missed the extensive discussion we've had on this earlier in another thread.
Quote from: CJ on 01/11/2015 03:25 am So, why on earth would they vent it overboard when they can contain it all within the system at effectively zero added weight? There is not a single thing on a rocket that has zero weight, zero cost, zero assembly time, zero paperwork, zero anything. Every screw, nut, bolt, wire tie, etc, adds to all those categories.So a different question is: Why would you spend money and time and add weight and complexity in order to catch spent hydraulic fluid that you can easily and safely just dump overboard? Since the system is at the top of a stage flying backwards, the slipstream is carrying the fluid up and away from the stage, so it won't be oozing down the sides. (And maybe they do catch it on F9R tests so it doesn't mess the vehicle/pad up in low-speed tests.)SpaceX has an admirable design philosophy: KISS. Dumping spent hydraulics fluid is totally consistent with that philosophy.
There is not a single thing on a rocket that has zero weight, zero cost, zero assembly time, zero paperwork, zero anything. Every screw, nut, bolt, wire tie, etc, adds to all those categories.
And maybe they do catch it on F9R tests so it doesn't mess the vehicle/pad up in low-speed tests.
What I do see as an issue is the mess the fluid would make on the F9; do they really want to have to have it all over the skin and getting into exposed parts? It just seems like an added headache to no gain to me. (and if they have to spend extra time cleaning the F9, that's money too).
Quote from: Robotbeat on 01/11/2015 12:56 amOther rockets which use open hydraulics:1) Delta III2) ConestogaDelta IV SRM's, Athena…..
Other rockets which use open hydraulics:1) Delta III2) Conestoga
I just saw some interesting speculation on the SpaceX subreddit that, if they used RP-1 for the hydraulic fluid, they could just release it into the main fuel tank when done. Could add a bit of margin for that landing burn...
I'm a bit puzzled by a few of the comments in this thread, so in the interest of clarity, I thought it might be helpful to envision roughly how much hydraulic fluid we're talking about here? Is it a gallon or so? Or is it (to go to an absurd level) closer to 100? If there's a lot of it, overboard dumping may make more sense. If it's in the range of a gallon or so, a small, very light, catch tank could be the simpler, cheaper, easier option. So, I'm asking; how much fluid volume do people have in mind? My own guess, somewhere between one and two gallons (based on the assumption that this is a pressure-fed one-way system, one actuator per grid fin, so 4 actuators.)
If the grid fins work independently, then a dual quadrant cylinder arrangement style (assume 2" bore, 1" rod, 6" stroke cylinders) would require 33 cu-in for each full turn. I'm not sure how many of these it would perform during its flight down, but if you assume 10 full turns: 33 x 4 quadrants x 10 turns = 1321 cu-in = 5.7 gal
Quote from: robertross on 01/11/2015 08:54 pmIf the grid fins work independently, then a dual quadrant cylinder arrangement style (assume 2" bore, 1" rod, 6" stroke cylinders) would require 33 cu-in for each full turn. I'm not sure how many of these it would perform during its flight down, but if you assume 10 full turns: 33 x 4 quadrants x 10 turns = 1321 cu-in = 5.7 galAnd, again: this was the root of the issue: SpaceX modelling indicated that they would do no more than (say) 10 full turns on the way down, and instead they needed to do 11. Modelling a control algorithm for hypersonic reentry is hard. It's rather surprising their estimates were as close as they turned out to be.
Quote from: cscott on 01/11/2015 10:48 pmAnd, again: this was the root of the issue: SpaceX modelling indicated that they would do no more than (say) 10 full turns on the way down, and instead they needed to do 11. Modelling a control algorithm for hypersonic reentry is hard. It's rather surprising their estimates were as close as they turned out to be.I'm surprised by this. Either they didn't budget sufficient margin, which is shortsighted, or in fact they did, which means their estimates were off by far more than 10%.Why not play it extra safe and double the amount of fluid until you know how much you don't need? Overbuild, then optimize the design.
And, again: this was the root of the issue: SpaceX modelling indicated that they would do no more than (say) 10 full turns on the way down, and instead they needed to do 11. Modelling a control algorithm for hypersonic reentry is hard. It's rather surprising their estimates were as close as they turned out to be.
Without knowing how the actuators work it is also quite feasible to assume that the fluid is expended just holding the fins in a specific position.
Seeing the launch video I had a question about grid fins. This thread is the nearest about grid fins.Belong these rods to the grid fins?P.S. Maybe it helps also for this thread.The whole video found here:http://forum.nasaspaceflight.com/index.php?topic=35853.msg1314858#msg1314858
Not really. If you have a finite quantity of hydraulic fluid, there's no way you'd design a system like that. All the linear and rotary hydraulic actuators that I am aware of do not expend fluid in holding a position. The speculation was not about how hydraulic actuators work, but about which type was used.And Elon's reference to time was likely based on Monte-Carlo simulations by which they determined that the hydraulic fluid would last for about 4 minutes of controlled flight.And yes, the speculation about "number of turns" is some people just pulling numbers out of a hat. But they do know how hydraulic actuators work.
Here's an updated version with what I believe to be the essential hardware, including proportional valves & safety valves.
I've come to this discussion late and I'm not clear where the "evidence" for this being a blow down hydraulic system.
Obviously the accuracy of that modelling becomes critical to avoid running out of charge/gas/fluid/whatever.Somewhere along the way their modelling didn't work out.
The reason I ask is that in earlier days SpaceX made much of analyzing launch failures and explained how their architecture countered the major failure modes. One of which was the use (in some LVs) of such systems and the way they could run out of actuator fluid.
AFAIk the SoA in actuator systems seems to be individual electrohydraulic units, replacing centralized pumps and networks of high pressure tubing with small separate packages to deliver high pressure fluid on demand.
Any system that can't draw on main engine power or tap the main propellant tanks (RP1 APU using atmospheric air perhaps?) will have severe limits on maneuvering. This would normally be countered by careful modelling and essentially choreographing the whole landing process very carefully. Obviously the accuracy of that modelling becomes critical to avoid running out of charge/gas/fluid/whatever.Somewhere along the way their modelling didn't work out.
Quote from: robertross on 01/12/2015 09:37 pmHere's an updated version with what I believe to be the essential hardware, including proportional valves & safety valves.I would put the pressure regulator behind the nitrogen valves, then the accumulators don't have to hold the full pressure of the nitrogen in the beginning.Springs in the mode shown (fins retracted) would have there lowest force (as the springs are at their highest length). I wonder if we couldn't pressurize it from the beginning. To avoid the oil draining over the servovalve pilotstage there could be a very small switching valve, pressuring the cylinders into retracted position. Would be (I guess) lighter than springs.
The springs are shown with the fins in the horizontal position - retracted or deployed. Actuating the cylinders in either direction causes the fin mechnism to rotate outside the tank.
As I had noted earlier on servovalves: they are inherently inefficient. I would suspect they use proportional directional valves instead (though possibly 2-stage, it doesn't really matter for this notional concept).
Quote from: robertross on 01/12/2015 11:41 pmThe springs are shown with the fins in the horizontal position - retracted or deployed. Actuating the cylinders in either direction causes the fin mechnism to rotate outside the tank.I was never good in reading mechanics. I thought if two cylinders extract, then fin goes out, if cylinders move differentialy rotation occurs. I guess I just wait for the 3D-Animation.
QuoteAs I had noted earlier on servovalves: they are inherently inefficient. I would suspect they use proportional directional valves instead (though possibly 2-stage, it doesn't really matter for this notional concept).www.mylesgroupcompanies.com/moog_pdfs/Moog%2030%20Series%20Catalog.pdfIt says on page 11 that leakage is <4% of rated flow. Might be worth taking a proportional valve, albeit I never saw it used in aerospace applications. I'm thinking of this grid fin at super sonic speeds, where flow changes very rapidly and the control system must act very fast in order to compensate disturbances. A servovalve has an insane dynamic, which is not found (afaik) in other valve types.
Elon's comment about running out of oil (Twitter, after he was asked, how the fins worked).
If we talk about modelling, we talk about modelling aerodynamics of 9 engine nozzles directed into the wind at super/trans/subsonic speeds. I think that's really a tough thing to model/predict (Edit:no steady flow state). It was a test flight, now they have their data.
But another question: Do we have evidence, that loosing the fins is the cause to the crash? It is a valid assumption, but it could also be unrelated.
Elon tweeted it was an "open" system and they ran out of fluid.
Yes, and they avoided that problem by using a fueldraulic system for the TVC actuators. But in this case, the grid fins were an "add-on" package on the wrong side of the fuel tank, and they evidently chose a standalone fluid tank.
Yes, there's a good summary here:http://www.mobilehydraulictips.com/end-hydraulics-aerospace-fast/
That's what test flights are for. Won't happen again.
I suppose I should have added an on/off solenoid valve inline of the fluid stream after the pressure reducing valve, to enable the flow of oil.The problem is: once the fins deploy, that leakage is continuous. You are discharging up to 4% of the rated oil flow whether you use the fins or not.
I do agree that servovalves are second to none in performance of frequency response, but there are some phenominal proportional valves made these days with much less leakage. Parker makes an industrial proportional valve that has servo valve performance that operates much like the voice coil of a speaker. They actually use that in their promotional material.
QuoteI do agree that servovalves are second to none in performance of frequency response, but there are some phenominal proportional valves made these days with much less leakage. Parker makes an industrial proportional valve that has servo valve performance that operates much like the voice coil of a speaker. They actually use that in their promotional material.From memory proportional valves are a lot more expensive than simple on/off types. My favorite hack in this area was the Aerojet tank pressurization system for the Kistler stage. IIRC they use 4 on/off valves with different flows to in effect implement an electro-pneumatic DAC. The flows (IIRC) were sized to the tank emptying rates so the scale was pretty non linear.
I have a few additional questions:(1) when the fluid ends, do the grids go to a "parking" position (e.g. is spring loaded or just driven there by the aerodynamic friction) or not? This is relevant because if yes, the engine can still try to correct the trajectory (at least at high speed), if not it may create a problem. I'm afraid the answer is NO, and that's may have played a role in the accident.(2) when the ACS find a problem with sensors or actuators, does it trigger an operation mode where it is trying to miss the target to save the ASDS or not? I"m quite surprised they didn't protected the equipment on the bridge from a crash.
I do think it is interesting that the onboard control didn't try to vector the stage away from the barge, especially if they had a fluid pressure sensor and knew roll was building up.
The first stage made a valiant effort to hit its mark. However, the return was deemed to be too fast for a secure landing, resulting in the loss of the stage.
I never said anything to imply anything. You are reading way too much into that.
Quote from: Chris Bergin on 01/13/2015 03:36 pmI never said anything to imply anything. You are reading way too much into that.Thanks, Chris. So you didn't mean to imply the stage may have attempted to divert. Sorry for putting words in your mouth.
Chris, can you tell us:(1) is "too fast" meaning... coming down too fast? (someone pointed out this is not obvious).
Quote from: pagheca on 01/13/2015 03:45 pmChris, can you tell us:(1) is "too fast" meaning... coming down too fast? (someone pointed out this is not obvious).Do you mean as opposed to going sideways too fast?
Elon elaborated on the grid fin issue at a recent Q & A, where he states that SpaceX needs "to make sure our hydraulic actuators don't run out of fluid and go hard over".
Because the valves are rarely rated to full pressure on the tank port. If it were, it would be a massive valve. In addition, the solenoid tube has a slug of metal inside that rides on a film of oil and that slug has to displace the oil in front of it to move.
I do agree that servovalves are second to none in performance of frequency response, but there are some phenominal proportional valves made these days with much less leakage. Parker makes an industrial proportional valve that has servo valve performance that operates much like the voice coil of a speaker. They actually use that in their promotional material.A proportional valve might be 4x the cost of an on/off valve, but a servo valve can be more than 5x the price of a proportional valve.
And there is no way you can get good control response (and accuracy) from an on/off system like you can with a proportional or servo system. If you overshoot your mark, you have to activate the opposite valve, and hunt back and forth, because the flow of oil is constant to the actuator in an on/off system. IN a proportional or servo system, your control logic determines how much power to send to the valve, which sends a specific flow rate of oil to the actuator.The response time for the valves will be dictated by how fast the grid fins need to move to compensate for a vehicle's movements. That there determines the overall cost of the system: response time. That's also a great benefit of returning the first stage: all those servo actuators & control computers used to gimbal the engines (not to mention the engines themselves).
I had a schematic in the other thread, but it really belonged here.Here's an updated version with what I believe to be the essential hardware, including proportional valves & safety valves.The drain system is still the biggest unknown.If it were a lead screw or rotary motion actuator, not much would change.
An interesting thought I wonder if something similar to Xcor's piston pumps could be a dirt cheap APU if the ever decide to go with a closed loop hydraulic system on one of their rockets.
I don't think we need to speculate about whether lack of fluid pressure got the fins "stuck"---a full minute of flight with no roll control authority means that residual roll is bound to have built up. Spinning stage centrifuges the fuel/ox away from the inlet, so the engine shuts down. Result: "hard" landing.
I would put the pressure regulator behind the nitrogen valves, then the accumulators don't have to hold the full pressure of the nitrogen in the beginning.
Quote from: cscott on 01/13/2015 01:51 pmI don't think we need to speculate about whether lack of fluid pressure got the fins "stuck"---a full minute of flight with no roll control authority means that residual roll is bound to have built up. Spinning stage centrifuges the fuel/ox away from the inlet, so the engine shuts down. Result: "hard" landing.Elon says the fins were hard over so yes they were stuck. There are different ways that they can be designed, and the way it was done as seen by the evidence we have before us is that opposite pairs were designed to get stuck in the same plane when actuating fluid went missing. Simple FMEA driven decision to avoid the possibility of spin up.
I'm guessing that the grid fins are deployed via a simple mechanical system based on the motion of the interstage away from the second stage and stay in the 90 degree position until folded in preparation for re-launch (not sure if I missed anything, do we have any indication that the fins ever fold back toward the stowed position in operation?)
every fin seems to be controlled in both axis.
I think too they have 2 degrees of freedom (tilting on the rotation axis, and up/down). I also think that they are interconnected, so there is only one valve or mechanism controlling all of them. Something like wires or rails.
Quote from: pagheca on 01/20/2015 08:19 amI think too they have 2 degrees of freedom (tilting on the rotation axis, and up/down). I also think that they are interconnected, so there is only one valve or mechanism controlling all of them. Something like wires or rails.Not likely.To keep it simple, it's better to have two actuator for two degrees of freedom.This is also the strong point of hydraulics, actuators are light, cheap and simple (compared to electromechanic).
I suggest that a single mechanism could have been used to control the four grid fins through a system of wires or common rails, so that the movement of the 4 grid fins is coordinated at all time.
Ok, let me elaborate a little bit on this suggestion...If the 4 fins are interconnected, they could be used to (1) counteract spinning, (2) change the deceleration rate(3) act as the feathers of an arrow, to realign the longitudinal axis and damp vertical oscillations.
IMHO a system with 4 grid fins interconnected, for example, by a rotating crown could be as solid, light and easy to design as a system with 4 independent servos.
Apologises for resurrecting this thread, but was there ever any confirmation on what was used as fluid, and if RP1, was it vented back in to the main tank or overboard?
Sorry, I was a bit unclear, I meant venting back in the main RP-1 fuel tank to be used as fuel, not in to a holding tank.But if it was definitely vented overboard that's a moot point, and there is no point in using RP-1 either.Thanks.
Not RP1, and vented overboard. They had separate fluid tank for the fins, with no recovery because the fins are needed only for a short time (a fluid recovery system that would pump it back to a tank would weigh extra, not worth it vs. having a tank big enough to just have enough while dumping used fluid overboard)
Quote from: Jarnis on 06/23/2015 12:20 pmNot RP1, and vented overboard. They had separate fluid tank for the fins, with no recovery because the fins are needed only for a short time (a fluid recovery system that would pump it back to a tank would weigh extra, not worth it vs. having a tank big enough to just have enough while dumping used fluid overboard)Do we have confirmation that it was vented overboard? I thought that the consensus here was that the used fluid would most likely have been dumped into a lightweight low-pressure container (something like an empty 2L pop bottle) near the fins, rather than expending the engineering resources to design the modifications to the interstage to allow for dumping the working fluid overboard.
I don't recall we ever got definitive word. But I don't believe there was ever "consensus" here either that the spent fluid was captured. Overboard dump of hydraulic fluid is common aerospace practice (see Conestoga rocket TVC system) and "modifications" to the interstage for such a system would be minimal. The outlet would be just a tube that vents spent fluid into the atmosphere. It's about the simplest plumbing system you can design on a rocket.
Quote from: Remes on 01/12/2015 11:28 pmI would put the pressure regulator behind the nitrogen valves, then the accumulators don't have to hold the full pressure of the nitrogen in the beginning.- Not likely that the liquid fluid is in an unpressurized (or low pressurized) tank initially and gas pressure is applied shortly before fluid pressure is needed. Reasons: a) no benefit to the tank from the standpoint of time exposure, time doesn’t count unless you’re dealing with material creep (which isn’t the case), b) no benefit in number of pressure cycles tank will see (one per flight either way) c) a closed valve will be more reliably successful in leaklessly closing against a liquid (the “hydraulic fluid”, whatever it is) than a gas (presumably N, but even more the case if He), d) energy in the form of gas P*V would be lost at the moment that you opened the valve to pressurize the hydraulic fluid tank in flight, e) The gas pressure tank size would need to be increased (vs. pre-pressurizing the volume above the fluid.
Also, there would seem to be some probability that there would be pistons (if cylindrical tanks) or bladders (if spherical tanks) separating the gas volume from the liquid volume so that the gas didn’t escape through the actuators in the early going before air drag has a chance to orient gravity relative to the stage.