-
Posts
80 -
Joined
-
Last visited
Content Type
Profiles
Forums
Developer Articles
KSP2 Release Notes
Everything posted by A Fuzzy Velociraptor
-
KSP inspired me to design a liquid-fueled rocket engine
A Fuzzy Velociraptor replied to ap0r's topic in Science & Spaceflight
Im not sure if you are, but take note that material strength will change significantly depending on your operating temperature. Im not certain what your anticipated equilibrium temperature of your walls are. -
Red Dragon confirmed!!
A Fuzzy Velociraptor replied to MajorLeaugeRocketScience's topic in Science & Spaceflight
Kerosene itself is considered a storable propellant. Depending on the thermal properties of the system, which I do not know as well as what sort of internal energy they would output which would help determine what the equilibrium temperature of the craft would be. Certainly with a kapton coated aluminum would likely result in the propellant freezing though most nuetral or abrobant coatings kerosene would be quite fine and would not suffer boil-off. It is the LOx which suffers from boil off though that is a different discussion.- 453 replies
-
- spacex
- red dragon
-
(and 1 more)
Tagged with:
-
Simple Merlin-derived SSTO
A Fuzzy Velociraptor replied to sevenperforce's topic in Science & Spaceflight
I think you have some misunderstandings about getting to orbit and the relevance of thrust to weight. The thrust produced is pretty much not relevant to the discussion provided that you are somewhere between a T/W of 1.2-1.6. What is more important is your total change in velocity and herein lies the difficulty of SSTO concepts. From the graph you can see that the available non-propellant mass fraction increases almost linearly with the increased specific impulse. The Merline 1D engine on its own has an ISP over the course of a whole flight of probably about 3000m/s. This means you are looking at a 4% non-propellant fraction when generating 9500m/s delta V. Now with a 100000kg vehicle that results in only 4000kg to play around in between payload and structure which really isn't feasible for most systems where you might be able to maintain a 500-1000kg payload. For a hydrolox system which may have an average of 4200m/s or so effective velocity which would allow you 10%-11% non-propellant fraction. However, for a hydrolox system about 10% of your vehicle will be structural mass. The Aquarius launch vehicle was proposed for SSTO using hydrolox and would have had a wet mass of 130000kg, a dry mass of 10000kg and only maintained a payload of 1000kg. Of course you can use techniques to try to reduce the structural mass which would allow a higher payload mass like higher density propellant or composite materials, but there are many circumstances where that may not be available such as with the cryosystems used on many launch vehicles. You can of course try to increase the specific impulse which air augmentation like you are describing may be able to or through other means but ultimately you are limited. The large ducts for air augmentation are also quite heavy which will further ear into your payload fraction. At this point, your idea needs more development and numbers associated with it before any analysis of the system can really be done. -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
ULA needs to lay-off employees not fire them. There is a large difference and that is an effect of changes in government spending on a defense contractor, and effect seen by a very large number of defense contractors. It isn't a matter that they are just wildly inefficient and didn't plan or know how to use budgets or business plans. Also experience is not a payment method not has spacex demonstrated significant stability and actually may be seen to be the opposite given the very high turnover rate of their workforce. A cost which they claim with a system that is currently not proven with many unknowns with a frequency claim the market does not support. Ariane is heavy subsidized by the governments of France and Italy to keep costs low so not really relevant. Also if you are referring to Ariane 5, the cost is 200M per launch. They claim values but have yet to make accomplishments. Honestly I would love to see them accomplish what they claim and I think the excitement they have generated towards space is fantastic, but until the data speaks it is just hype. -
Wait how did I get pulled into this? @Nibb31 since he's the only one I can think of maybe?
-
How pointy Minuteman III does not burn
A Fuzzy Velociraptor replied to kerbarara's topic in Science & Spaceflight
I don't know much about the minuteman systems, however for high speed vehicles which do not use a bluntbody, vehicles may be passively or actively cooled. Most ICBM systems will be a passive system which will either use heat resistant materials like tungstun or graphite or may use a ablative material like graphite or cork. -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
Please forgive me, but I don't actually understand what you are trying to say. -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
I wish I knew how to break up quotes on this forum. The problem is that he doesn't actually have a huge amount of money. He may be worth a good amount but a large amount of that is based on how much he could sell his assets for and not his liquid assets. They need money from selling services or subsidization. They cannot survive on their own just doing cool things. It is actually quite easy to call what they have done not revolutionary. They haven't done anything significantly revolutionary, they have made some evolutionary improvements like the supercooling for cryo systems but that is not revolutionary. Landing a rocket vertically is not revolutionary, we've done that before and even so it doesn't make a bit of difference right now. If they can demonstrate extended re-usability and significantly reduced costs for the the system then there would be a discussion. At this point they have not achieved anything revolutionary. Maybe they will at some point and when that happens that'll be a different discussion. Even the lower cost isn't a revolutionary aspect, its amazing how much cost you can cut by overworking people and using low-paid intern hoards. Without a significant profit a company cannot serve, regardless of how noble their goals may be. Yes, SpaceX employees have a choice. This is why they leave. They only survive right now because they are popular. That won't continue forever and when it does they will either have to radically alter their corporate culture or die. There is a reason traditional business thinking is traditional business thinking, it is a timeproven way business works. It isn't how Musk thinks about it, its about what works and what doesn't. No good idea that requires a lot of money no matter how cool it may be will survive without customers and a market. They are the only group pursuing a Mars initiative because everyone else realizes there is not a market nor money available. Also we have no idea if their idea is reasonable or even at least feasible at this point. We know literally almost nothing about it. Speculation is ultimately useless. Until the data speaks it is nothing but hype. -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
Hush you. I know a fellow called Nipp on another forum, must have thought it was the same name. It is fixed now. Also I'm fairly sure we agree. -
Optimal shared-engine TSTO configuration
A Fuzzy Velociraptor replied to sevenperforce's topic in Science & Spaceflight
Short answer: It depends Long answer: It depends on a large number of factors which can't easily and fully be broken down into an explanation over the course of a post in a thread. One of the first factors being the vehicle mission. A vehicle designed to be reusable will have different parameters than one that will be entirely disposable. For instance a reusable vehicle that intends to propulsively land will require an amount of delta V to perform the landing that is no longer used towards the orbit system. Different missions will also have different trajectories that may require more or less thrust on the upper stage compared to lower stage. Different feed systems will vary the mass ratios as will different geometries. Upper stage engines may also have extended nozzles and even without them will have substantially different effective velocities than those used in the low atmosphere. There are a lot of different other factors as well. You can compare what various companies use but ultimately it will be a case by case basis what will work best and it may simply come down to what is convenient or good enough rather than what is optimal. At least for a disposable system you can estimate stage mass ratios with the following. x=[1:100]; y=zeros(1,100); c=1; % Mass Fractions nuL=0.07; nuU=0.07; % Effective Velocity EVL=3400; EVU=3550; % Total Rocket Mass Mt = 1920; % Payload Mass Mp = 80; for c=1:100 % Lower Stage Mass Ml=Mt*x(c)/100; % Upper Stage Mass Mu=Mt*(100-x(c))/100; % Delta U estimation delU = EVL*log((Ml+Mu+Mp)/(nuL*Ml+Mu+Mp))+EVU*log((Mu+Mp)/(Mu*nuU+Mp)); y(c) = delU; end plot(x,y) title('Delta U Estimation'); xlabel('Percentage First Stage of Total Vehicle Mass'); ylabel('Delta U') grid on -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
I think you might be misunderstand what @Nibb31 is saying, though I may be as well. That fact that younger people tend to have less experience is just a fact of being young and many people here may have very little experience with the industry. There is a difference between what is realistic, and pessimism and optimism. Understanding what can be done and what the market can support and understanding those things probably won't live up to the revolutionary promises we've had isn't pessimism. Unfortunately, space systems aren't really a revolutionary field, they are an evolutionary field. It's not bad to be excited, space is exciting, there is so many cool and interesting things that are being done everyday. Just because an idea may not be able to transform the market, or something is extremely difficult doesn't mean it isn't exciting or shouldn't be worked on. These systems are extremely complicated and difficult and will take time to develop. More importantly a new product has to have a customer and a market, and for many of the ideas people propose (young and old) it simply isnt feasible. Many new companies rise and fall. SpaceX has done some cool things certainly and they have done some things that other groups before them haven't and the story of an eccentric billionaire that wants spaceships that have come up lately is certainly a story which helps the excitement. Ultimately though SpaceX hasn't really proven anything. They have demonstrated that they can change a lower launch price off the backs of their employees and intern hoards, though the very high turnover rate of employees should be testament to the treatment of their employees. It will work for now and while they can keep the hype and the starry eyes of the younger people, but eventually they won't be able to anymore and they won't survive as they are now. They have made claims of re-usability and costs as low as 11MUSD, and yes they have landed a booster on land and barge, but until the data speaks it is nothing more than empty claims. Even to get to the costs they are claiming they require a market which simply doesn't exist. They make further claims about plans to colonize beyond Earth, and so do some of the other people of their grande plans, but without a market or a need, or the money available to do it they have no chance of succeeding. It is possible they could build these grande plans of theirs investing their fortunes to do it, but ultimately a product without a customer is just a money sink and will either be cancelled or force the company into bankruptcy. This does not mean that revolutions can't happen or that they won't be exciting as they happen, but until the data speaks it is hype, nothing more. -
What do you think about SpaceX?
A Fuzzy Velociraptor replied to freakazoid13's topic in Science & Spaceflight
I think they have done some cool things and I think some of the change they have brought through to the industry will ultimately have a generally positive effect in both costs and availability of investments. I think some of their plans are more hype and less actually fully thought through that I do not believe there will actually be an effective market for things that are planned like their colonization. Right now, they are new and different and proposing revolutionary over evolutionary systems. They can get people really excited in ways that other companies dont care to or need to. Ultimately this excitement is their survival. If you talk to many aerospace engineering students you may find that almost 80% of them are really excited about spacex and very few that get excited and want to work for ULA, Orbital, Ariancespace, Airbus Defense, Aerojet, Mitsubishi Heavy Industry etc. They have a young population and a very high turnover rate. They can only maintain their process as it stands while the starry eyes remain. Eventually, they won't be new, their cult following and the stars with fade and they won't have the masses of interns they do now. They will have to make some tough decisions about their culture and how they treat their employees that I don't think they will be able to make. They may make some lasting impact, but ultimately, they won't survive. -
Apparently the actual question proposed was 1.5 times Earth radius not 1.5 Earth mass. It is easy enough to fix, the calculations weren't difficult. If we assume that this new planet maintains a similar density to earth then the mass of the planet will be 1.5^3 or 3.38x Earth mass. If we assume a 500km orbit (though at this point the atmosphere may be starting to get thicker than this then using the simplified vis-viva equation and our starting assumptions to be sqrt(3.375*3.986E5/(1.5*6371+500)) then we find we need a horizontal velocity of 11.56km/s. Gravity drag and aerodynamic drag will certainly be higher than on Earth. Even then vehicles with 15-16km/s available delta V are available. Certainly they would struggle miniaturizing their satellites and manned space systems would be quite expensive, though it would not be impossible. Though actually I have a fairly decent comparison for what 10km/s vs 15km/s does to an available load, our Dyeus H system is expected to carry 150kg to LEO but would likely only be able to carry 5kg to a C3 of 1.78 (which would be 15km/s, about 1km/s more than LLO). Probably the largest issue they would struggle with is the atmospheric pressure at low altitudes which would either cripple efficiency or require massively increasing chamber pressures in order for rockets to actually be effective. They would likely need to use some other system for the first stage.
-
Actually they would probably be smaller because of the square-cube rule. A smaller creature would have an easier time supporting itself. Also there isn't any reason they couldnt develop in a different manner maybe looking almost drawven, to quadrupedal (or more feet), having some sort of prehensile tail or tentacles or even being aquatic. There are any number of ways that a creature could develop given its circumstances. Anyway, I don't see any reason why it wouldn't be possible for them to develop space capability. Certainly some things like SSTO would be nearly impossible but just getting to orbit likely wouldn't be much more difficult at all. Assuming the planet in question has a density comparable to earth and a mass 1.5x that of earth and that earth's radius is 6371km using R=(1.5*re)^(1/3) (this can be found by setting the two densities equal and using the volume of a sphere) the new planet's radius is 7293km. Using a simplified vis-viva equation for a circular orbit (v=sqrt(1.5mu/R), assuming that the gravitational constant is the same everywhere in the universe and a 400km orbit the orbital velocity may be found to be about 8.8km/s. This is higher than earth of course and we can probably expect higher gravity drag and higher aerodynamic losses. Even if those loses totaled 4k it would require 12.8km/s of delta V. It is quite a lot, but it isn't impossible given that that number is a little less than what we require to launch something to GEO. Of course there would be engineering challenges to address but even most of those could be dealt with. At launch engines would likely be less efficient than on earth and would either require a higher chamber pressure, earlier staging, or expanded use of altitude adapting nozzles. Though it really depends on how much their standard atmospheric pressure is. If they were only a 2atm, it likely wouldn't pose too much of a challenge except that engines would either have to run on a higher pressure or run less efficiently. If it were in the range of 20 then it would be quite a challenge though that could likely be overcome through launching from a high altitude, airlaunching, or accelerating through some other method. One thing that would be interesting, given that they would likely resort to higher chamber pressures they might never encounter engine coking below 5MPa.
-
I believe seven was referring to engines of similar thrust levels. If you take a look through Aerojet Rocketdyne's capabilities page most of the monopropellant engines are lighter than the bipropellant engines of similar thrust levels. Monopropellant engines tend not to be used for high thrust applications so the thrust to weight scaling is a little screwy.
- 66 replies
-
- monopropellant
- engine design
-
(and 1 more)
Tagged with:
-
The first part isn't really a logical argument especially when talking about rocket propellants. These things have to have a lot a chemical energy which can be released from a reaction, they aren't necessarily going to be extremely human friendly. Also as the other fellow said, peroxide is actually quite stable at high concentrations and actually gets more stable on its own as the concentrations rise. Yes HTP can start decomposing which causes a run away reaction rather than a detonation. HTP itself is actually decently stable and requires a significant amount of energy to begin to thermally decompose. Of course is certain impurities get into the system the impurity can act as a catalyst which can have non-optimal consequences. HTP is often used in some capacity as a propellant though often in monopropellant and normally not main engines for space launch vehicles. It can be bought in 50, 70, 90, 98 and higher concentrations from various suppliers though you may have to buy a large quantity of the stuff depending on the supplier. Also the safety concerns of HTP and Hydrazine can outweigh the storage concerns of hydrolox. For one safety concerns can easily be managed and mitigated where storability may not. There are a number of other reasons and variations why some entities pursue hydrolox vs kerolox vs keroperox vs UDMH/NTO etc and it can't really be simplified down to a couple sentence discussion.
- 66 replies
-
- monopropellant
- engine design
-
(and 1 more)
Tagged with:
-
Attached shock is the main thing there vs the detached shock since air itself is an insulator. The oblique shock doesn't slow the air as much but the temperature effect is not as dramatic as a normal shock. Though the effect of the oblique shock with such a ridiculously large Mach value means the air itself should be glowing rather than the body (assuming no convective heat transfer to the body).
-
Even then you would generate an oblique shock which would have a significant temperature change. At Mach 100 even at a turn angle of only 5 degrees the temperature after the shock is 22 times the temperature before the shock. Even 2 degrees of turn angle and the temperature will still be almost 5 times the temperature before the shock.
-
Anyway, sorry I haven't responded to this in like a really long time. I've been meaning to and have been extremely busy and distracted with my own work. So I think what have a couple holes in your understanding of how rockets work and mistaking some causes. We've already talked about the energy and thermodynamic misunderstandings associated with this design so I'll leave that out. I think you are misunderstanding the reason monopropellant engines tend to have a high thrust to weight ratio. While they generally don't produce large thrust, they are very light. They can be this light because in general monopropellant engines are very simple. Often times they will be pressure fed and of course only require a single inlet. Often times the same can be said for the small bi-prop engines (though the monoprop is often still lighter).If you compare the MR-104A/C 440N monopropellant thruster and the Hi-Pat 440N High Altitude thuster both made by Aerojet Rocketdyne, the difference in mass is that the MR-104A/C is about about 60% lighter than the Hi-Pat. Monoprop engines don't release nearly the chemical energy most bi-propellant reactions used in rocket propulsion. As such they may not require the cooling other systems may need, but this also means they don't have nearly the energy the other systems do. If you compare the nozzles of the MR-104A/C and the Hi-Pat, the Hi-Pat has a significant expansion added to it which about doubles the nozzle length and before the expansion the two engines are about the same size. I don't know how much that expansion weighs, but by dropping it the mass of the engine would likely be significantly reduced bringing the two mass figures significantly closer together. A monoprop engine on its own scaled up would likely be slightly lighter than a biprop engine of the same thrust but at significantly lower performance. Also consider that, while thrust to weight is an important factor, the variation of the thrust to weight of just the engine masses among almost any chemical system is going to be almost irrelevant. For example consider that you desire an initial acceleration of somewhere in the range of 1.25G and both engines listed earlier produce 100lbf of thrust. So vehicle mass will likely come in around to about 80lbm to start. The MR-104A/C weighs 4.11lbm and the Hi-Pat 11.5lbm which is about a 9% difference in the total mass ratio compared to the 60% in the engines. This is a fairly extreme example, the R-4D is a 4% difference. The point here being that seemingly large differences in thrust to weight ratios of engines aren't nearly as large an effect on the whole vehicle as may appear. We do talk about thrust to weight ratios and why they are important but the main reason we talk about them especially with engines is when talking about mass limited vs energy limited propulsion systems. Note the difference in thrust to weight of BPT-200 Hall Effect Thruster vs the chemical example of .0024 and 24.33 which is a difference by about a factor of 10000 and that only factors in the engine and not the masses of all the systems needed to make the engine work which will only hit the energy limited system harder. While you talk about thrust to weight quite a lot and keeping the mass of the system low, the engine you are proposing is actually quite complex. The reason monoprop engines are so light is largely because of that simplicity. Aerospike/single expansion nozzles are not light apparatus to start with and the air intakes associated with an air augmentation system are significantly more massive. In addition by using two separate combustion chambers you are essentially doubling the weight of the combustion chambers for a given mass flow rate which significantly more hampers your thrust to weight. The last thing is in regards to the air augmentation system. Air augmentation is a very complex problem of aerodynamics that honestly I'm nowhere the level of education required to design of fully comment on one. I can tell you that they are very heavy and difficult to design, they can allow for significantly higher specific impulse but are expensive, complex, difficult to design, and will only work in the lower parts of the atmosphere as I understand them. The weight of the system is large enough that it will significantly affect your system rather than the weight of monoprop vs biprop engines. Also given the time and expense of manufacture it would likely be unadvised for a disposable vehicle and would likely have limited advantage is a vertical take off vehicle. Also normally you increase the working mass after combusting the propellant rather than before. Nitrogen especially in certain reactions can act as almost an anti-catalyst and can significantly slow down your reaction. This is already non optimal in a combustion chamber where the fluid may be moving at 50m/s, it is significantly worse when flying down a nozzle moving at 500+m/s. You will likely wind up with some nitrous oxide as well from oxygen colliding with the nitrogen though that shouldn't have too large of an effect except a bit of pollution. At least with your crude drawing the air would likely significantly impede the combustion of your propellants.
- 66 replies
-
- monopropellant
- engine design
-
(and 1 more)
Tagged with:
-
Rocket math for custom fuel tanks in RSS
A Fuzzy Velociraptor replied to Halo_003's topic in Science & Spaceflight
Looking at your original problem you have about a 12ish percent mass ratio on your first stage and about a 6 percent mass ratio on your second. If you were to plot that you would probably find that your graph would be left shifted quite a bit wanting a substantially smaller first stage. A bunch of variables were changed when I wrote the program so a direct comparison isnt particularly valid. Thing wouldn't let me edit the previous post. -
Rocket math for custom fuel tanks in RSS
A Fuzzy Velociraptor replied to Halo_003's topic in Science & Spaceflight
Open matlab and you can either just paste it in the command window or create a function file, save and hit F5 or the run button. If you wanted it in excel create a column from 1-99 and then another column from 99-1 using a desired starting mass determine the mass of the stages for each case. Add in the other factors like isp and structural mass ratio then just drop in the delU equation (though change variables for the respective excel boxes) and drag it down. Then you can either plot it or search for the highest delta V value. If you want to determine rough ratios for different propellants try to find the stoichiometric mass ratio. If you want more specific numbers use a program called NASA CEA though it isnt very user friendly if you dont know what youre doing. CEA will also give you your ISP values. Your structural mass ratio will vary as you change in size and as your propellant densities change. Denser propellants will have lower structural mass ratios as will larger vehicles. While it is possible to get your structural mass ratios quite low I would keep them around 10 percent for more things you will be looking at and especially at the level you are looking at. As for why the graph behaves like that I cant say with certainty but some of the difference will be caused by our using different values, especially in the structural mass regime. As the structural mass and payload in your upper section increases the graph will tend to shift towards a larger first stage. As your structural mass decreases it will shift towards a larger second stage. The varying ISPs will also affect your system quite a bit. Unfortunatley here there are a lot of variables that all affect how the graph looks that it is sort of hard to give a proper explanation of how exactly it gets that shape. -
That would be highly unadvisable. Use of a normal catalyst either platinum or a permangenate once in the chamber would be much more advised. I will address the rest of the thread since I was on last night when I have a chance to make a long post.
- 66 replies
-
- monopropellant
- engine design
-
(and 1 more)
Tagged with:
-
Rocket math for custom fuel tanks in RSS
A Fuzzy Velociraptor replied to Halo_003's topic in Science & Spaceflight
I can't say I know what RSS is in this context or at all so some of what I'll say may not make sense or apply for your application. However, there are a couple issues with your system estimations I can see. I am not sure about the validity of saying that 95% percent of the cylinder available for propellant, though since I don't think you are trying to actually design something it is probably fine. Also I don't know where or how you have calculated your structural mass fractions but since your system uses a LH2/LOX you should anticipate your structural mass to be no less than 10% the mass of your system and likely closer to 12-15%. Assuming you plugged the stuff into the equation correctly (be sure to take into account the mass of the second stage in your first stage delta V estimation) your delta V estimations should be accurate. So yes, your payload would be less than 10kt. Increasing the payload capacity of your vehicle would actually be fairly easy to do. Vary the ratio of the mass of your lower stage and upper stage until you hit a peak. This can be done pretty easily through creating a plot in excel or running a program through matlab/fortran etc. However, increasing your capacity to 20kt from there may be quite a bit harder than just optimizing your system. You will likely need to increase the mass of your vehicle or the number of stages of your vehicle including adding parallel staging. ___________________________________ I went ahead a wrote a quick program for you that will work with matlab at least. By increasing your ratio so that your first stage is about 83% of the total mass of the system not including your payload and your second stage to about 17% the total mass you can get your payload up to 10kt with a total mass of 220kt. I added in the assumptions of your structural mass being 10% for the first stage and 15% for the second stage. I also added in your ISP values though I accidentally used 386 instead of 381. Program: x=[1:100]; y=zeros(1,100); c=1; % Mass Fractions nuL=0.1; nuU=0.15; % Effective Velocity EVL=386*9.81; EVU=455*9.81; % Total Rocket Mass Mt = 200000; % Payload Mass Mp = 10000; for c=1:100 % Lower Stage Mass Ml=Mt*x(c)/100; % Upper Stage Mass Mu=Mt*(100-x(c))/100; % Delta U estimation delU = EVL*log((Ml+Mu+Mp)/(nuL*Ml+Mu+Mp))+EVU*log((Mu+Mp)/(Mu*nuU+Mp)); y(c) = delU; end plot(x,y) title('Delta U Estimation'); xlabel('Percentage First Stage of Total Vehicle Mass'); ylabel('Delta U') grid on -
Smallest possible orbital rocket
A Fuzzy Velociraptor replied to sevenperforce's topic in Science & Spaceflight
A lot of Kerbal physics while fine for a game don't translate well into real-life. Also while you may be able to decrease gravity losses through a faster burning engine you will in-turn increase your drag as well as structural requirements. On the topic of very small rockets, there is a small company called Aster Choronautics, that claims their launch vehicle would be able to carry a 5kg payload to orbit on a vehicle with a mass of 1000kg in a single stage version or 50kg with 1250kg in two stages.