We are leaving for the Desert/FAR about 6pm tonight.
We hope to hover the silver ball and get some performance measurements.
We will tweet results, we may post from the site in the evenings if our new MiFi works from there.
.
Wednesday, September 30, 2009
Monday, September 28, 2009
Accuracy...
It is very possible that there will be prizes determined by accuracy this year. I've spent the weekend working with the high precision RTK GPS and the helicopter.
If we can do the 180 seconds (big if) then I'm pretty comfortable beating the rumored 0.8m accuracy of Armadillo, beating the rumored 10cm accuracy of Masten is going to be a lot harder.
I still have a fair bit of work to do, but it looks promising. I'm not going to say much more about this as I don't want to give to much away. The chalange is in the vehicle dynamics, not the GPS.
Any sufficiently advanced technology is indistinguishable from Magic. Multifrequency RTK GPS stuff is sufficiently advanced. When you realize your making sub 1 cm scale measurements over a 20000km distance in real time at many Hz it is stunning. That is 0.00000005 percent. or 0.0005 ppm.
If we can do the 180 seconds (big if) then I'm pretty comfortable beating the rumored 0.8m accuracy of Armadillo, beating the rumored 10cm accuracy of Masten is going to be a lot harder.
I still have a fair bit of work to do, but it looks promising. I'm not going to say much more about this as I don't want to give to much away. The chalange is in the vehicle dynamics, not the GPS.
Any sufficiently advanced technology is indistinguishable from Magic. Multifrequency RTK GPS stuff is sufficiently advanced. When you realize your making sub 1 cm scale measurements over a 20000km distance in real time at many Hz it is stunning. That is 0.00000005 percent. or 0.0005 ppm.
Thursday, September 24, 2009
Testing Results, mostly good...
For the next 36 days Updates will be short and probably not include much video or pictures, they just take to long to edit and prepare and I'm out of time.
We static tested the silver ball today, on the first static test we were expecting bright orange hydrocarbon flames(We were running gasoline as the fuel) . The exhaust was clear loud and steady. We though the fuel valve did not open and made some minor changes to the static test software refilled with peroxide and ran it again....
During the second run again clear loud steady rocket exhaust, but again no expected orange,
although about 10 seconds from the end of the run there was a noticeable change in the exhaust note.
On closer inspection the fuel tank was empty and there were signs of light soot on the thermal coating in the motor. Based on the run time and amount of fuel we loaded the pitch change at 10 seconds from shutdown would have been when we ran out of fuel. This is the first gasoline fueled motor we've ever seen fired so its possible the only thing wrong was our expectations.
We switched from kero to gas for more volatile fuel and slightly better (1 or 2 sec) ISP.
Given that we then prepared to try a tethered hover with the gimbals, and we ot bitten by small details. On the blue ball the only vent is the emergency vent we fill through the vent and have never had problems. On the silver ball during fueling for the two static tests the purge vent had been open from the main tank utalage to the motor. Before fueling for the potential tethered hover we closed that vent so we could slightly pressurize the vehicle, hang it and see if the roll control thrusters were working correctly. (They were) this caused us to leave the purge vent closed. Since the blue ball has no such vent we did not think it was an issue.
We put the fill tube down the vent with lots of clearance and started pumping.... a minute or so into the fill the hose comes flying back out and sprays everything with peroxide. We immediatly douse the vehicle, and perople with the standby water. This water deluge prevented any real damage, but also caused some of the electronics to be glitchy. So we called it a day and brought the vehicle home to dryout and test the wet parts.
The reason this happened is subtle, the main vent attaches to the tank with a 1.5" Sanitary fitting. A 1.5 inch sanitary fitting is very similar to a 1" sanitary fitting, ie they are identical other than the hole in the middle of the gasket. We had put a 1" gasket on the 1.5" vent. So when the 1" fill tube is shoved down into the vent it seals with this gasket.
Now that I'm home I will review the data in detail and look at the video.
We static tested the silver ball today, on the first static test we were expecting bright orange hydrocarbon flames(We were running gasoline as the fuel) . The exhaust was clear loud and steady. We though the fuel valve did not open and made some minor changes to the static test software refilled with peroxide and ran it again....
During the second run again clear loud steady rocket exhaust, but again no expected orange,
although about 10 seconds from the end of the run there was a noticeable change in the exhaust note.
On closer inspection the fuel tank was empty and there were signs of light soot on the thermal coating in the motor. Based on the run time and amount of fuel we loaded the pitch change at 10 seconds from shutdown would have been when we ran out of fuel. This is the first gasoline fueled motor we've ever seen fired so its possible the only thing wrong was our expectations.
We switched from kero to gas for more volatile fuel and slightly better (1 or 2 sec) ISP.
Given that we then prepared to try a tethered hover with the gimbals, and we ot bitten by small details. On the blue ball the only vent is the emergency vent we fill through the vent and have never had problems. On the silver ball during fueling for the two static tests the purge vent had been open from the main tank utalage to the motor. Before fueling for the potential tethered hover we closed that vent so we could slightly pressurize the vehicle, hang it and see if the roll control thrusters were working correctly. (They were) this caused us to leave the purge vent closed. Since the blue ball has no such vent we did not think it was an issue.
We put the fill tube down the vent with lots of clearance and started pumping.... a minute or so into the fill the hose comes flying back out and sprays everything with peroxide. We immediatly douse the vehicle, and perople with the standby water. This water deluge prevented any real damage, but also caused some of the electronics to be glitchy. So we called it a day and brought the vehicle home to dryout and test the wet parts.
The reason this happened is subtle, the main vent attaches to the tank with a 1.5" Sanitary fitting. A 1.5 inch sanitary fitting is very similar to a 1" sanitary fitting, ie they are identical other than the hole in the middle of the gasket. We had put a 1" gasket on the 1.5" vent. So when the 1" fill tube is shoved down into the vent it seals with this gasket.
Now that I'm home I will review the data in detail and look at the video.
Wednesday, September 23, 2009
Leaving for Test Site within the hour...
We are going out to static test and maybe over the silver ball.
News when we return.
News when we return.
Tuesday, September 22, 2009
Unreasonable Souvenir.
Garage Space is always at a premium. We usually take broken parts to the scrap metal dealer once a month. Since the rocket and camera died in such a fun way we thought we'd sell the pair on e-bay as a souvenir.
Also here is the promised picture of the silver ball in all its glory....
Also here is the promised picture of the silver ball in all its glory....
Silver Status...
I finished the integration testing on the Silver ball last night.
the GPS cable turned out to be an intermittent antenna problem.
Swapped out the antenna and all is well.
We ended up about 30lbs over the target weight, at least 10 of these pounds are in the new stainless motor. In the present condition using carefully calculated ISPs adjusted for under over expansion we get a hover time of 182 seconds. we never ever get theoretical ISP, so we will need to do something. Most likely a careful weight diet and the addition of some external pressurization so we can fill past 50% point. If we fill the tanks till we have a takeoff Thrust to weight of 1.2 I get a hover time (using our historically achieved ISP fudge factor) on the order of 200 seconds. Some Parts like the Motor are really massive,....
Some parts like the roll control thrusters are as light as I can imagine...
We hope to test this coming weekend.
If I get home before its dark today, I'll drag it out in front of the garage and try to take an overall picture of the beast.
With the blue ball the "vector" point for the vanes is on the bottom of the motor so that there is always an offset from this point to the center of gravity ie CG. With the silver ball this point is the gimbals hinge and with the heavy motor we were worried abut the this getting too close to the CG, so we mounted the payload up high, since we mounted it up high. The (easily removable) leg weights we use on the blue ball won't work. The silver payload are steel blocks bolted to the top of the vehicle. This combined with the increased weight makes the silver ball a lot harder to move around. My son and I can still carry it, but we cant lift it 3 feet in the air to put it in the truck without assistance.
UPDATE
I've done some thinking and calculating. If I leave the system as blow down and do all the weight reduction things I can think of its still pretty marginal. If I put a pressurization system on the vehicle I can add 30 more pounds and get 75% of theoretical ISP (This is 0.75 * the number beyond already accounting for over/under expansion etc...) and still do the task.
If this weekends test goes well it looks like engineering a pressurization system gives me the most bang for the time invested. It has the side effect of almost doubling my propellant usage and that makes my propellant supply look tighter than I'd want for through testing, but it still looks like the best path.
A side note on calculating ISP for over or under expansion...
For under expansion I just use the ISP I calculate when the chamber pressure is exactly right for proper expansion, thus I get no ISP credit for the actual higher chamber pressure.
For over expansion I take the theoretical ISP for perfect expansion at this pressure, then calculate the area difference between the actual expansion cone and the perfect one. I then calculate this difference in area as a thrust equal to the (ambient pressure * this area).. in the wrong direction and use tht as a reduction in ISP.
I have some C code that will run propep as a sub process and can make these calculations automatically. So as an excercise I varied the actual expansion ratio up and down searching for the flight performance maximum. It turns out the simplifying assumptions I made in January 09 blog post of using the average flight chamber pressure was withing 1% of the more rigorious method, well withing the uncertanty in all these numbers.
the GPS cable turned out to be an intermittent antenna problem.
Swapped out the antenna and all is well.
We ended up about 30lbs over the target weight, at least 10 of these pounds are in the new stainless motor. In the present condition using carefully calculated ISPs adjusted for under over expansion we get a hover time of 182 seconds. we never ever get theoretical ISP, so we will need to do something. Most likely a careful weight diet and the addition of some external pressurization so we can fill past 50% point. If we fill the tanks till we have a takeoff Thrust to weight of 1.2 I get a hover time (using our historically achieved ISP fudge factor) on the order of 200 seconds. Some Parts like the Motor are really massive,....
Some parts like the roll control thrusters are as light as I can imagine...
We hope to test this coming weekend.
If I get home before its dark today, I'll drag it out in front of the garage and try to take an overall picture of the beast.
With the blue ball the "vector" point for the vanes is on the bottom of the motor so that there is always an offset from this point to the center of gravity ie CG. With the silver ball this point is the gimbals hinge and with the heavy motor we were worried abut the this getting too close to the CG, so we mounted the payload up high, since we mounted it up high. The (easily removable) leg weights we use on the blue ball won't work. The silver payload are steel blocks bolted to the top of the vehicle. This combined with the increased weight makes the silver ball a lot harder to move around. My son and I can still carry it, but we cant lift it 3 feet in the air to put it in the truck without assistance.
UPDATE
I've done some thinking and calculating. If I leave the system as blow down and do all the weight reduction things I can think of its still pretty marginal. If I put a pressurization system on the vehicle I can add 30 more pounds and get 75% of theoretical ISP (This is 0.75 * the number beyond already accounting for over/under expansion etc...) and still do the task.
If this weekends test goes well it looks like engineering a pressurization system gives me the most bang for the time invested. It has the side effect of almost doubling my propellant usage and that makes my propellant supply look tighter than I'd want for through testing, but it still looks like the best path.
A side note on calculating ISP for over or under expansion...
For under expansion I just use the ISP I calculate when the chamber pressure is exactly right for proper expansion, thus I get no ISP credit for the actual higher chamber pressure.
For over expansion I take the theoretical ISP for perfect expansion at this pressure, then calculate the area difference between the actual expansion cone and the perfect one. I then calculate this difference in area as a thrust equal to the (ambient pressure * this area).. in the wrong direction and use tht as a reduction in ISP.
I have some C code that will run propep as a sub process and can make these calculations automatically. So as an excercise I varied the actual expansion ratio up and down searching for the flight performance maximum. It turns out the simplifying assumptions I made in January 09 blog post of using the average flight chamber pressure was withing 1% of the more rigorious method, well withing the uncertanty in all these numbers.
Sunday, September 20, 2009
Beaten by a cable...
I expected to take the silver Ball to FAR on Friday.
I've worked for 18 hrs a day for three days and I'm still 3 hours away from being ready.
I think I have everything working except the GPS.
On The Blue Ball the GPS antenna sticks up and we've broken it off three times. So on the Silver ball I machined a real light aluminum dome that sits on top of the computer cover and the GPS antenna sits on that. It all lines up so it goes directly into the RF tight box the GPS lives in.
Nothing sticking out to break off.... alas every time you open or close the computer box you must unplug and replug the GPS cable. After opening the box 6 million times over the last few days the cable died. If I replace it with a different cable everything works, alas the mechanical fit is such that I can't assemble things with the different cable. We've passed all our integration tests except the last one a simulated takeoff, we can't do that without a GPS lock and no antenna == no lock. I could go and do a static test, but I think I'll wait a day or two and try again. I've been pushing too hard and I'm too tired to think clearly not the right state to be in testing a new system.
One of these days I will learn how long it takes to do vehicle integration.
When you have a pile of individually tested and roughly rocket looking parts you think you have 4 hours of work, alas its days or weeks of work, not hours.
I've worked for 18 hrs a day for three days and I'm still 3 hours away from being ready.
I think I have everything working except the GPS.
On The Blue Ball the GPS antenna sticks up and we've broken it off three times. So on the Silver ball I machined a real light aluminum dome that sits on top of the computer cover and the GPS antenna sits on that. It all lines up so it goes directly into the RF tight box the GPS lives in.
Nothing sticking out to break off.... alas every time you open or close the computer box you must unplug and replug the GPS cable. After opening the box 6 million times over the last few days the cable died. If I replace it with a different cable everything works, alas the mechanical fit is such that I can't assemble things with the different cable. We've passed all our integration tests except the last one a simulated takeoff, we can't do that without a GPS lock and no antenna == no lock. I could go and do a static test, but I think I'll wait a day or two and try again. I've been pushing too hard and I'm too tired to think clearly not the right state to be in testing a new system.
One of these days I will learn how long it takes to do vehicle integration.
When you have a pile of individually tested and roughly rocket looking parts you think you have 4 hours of work, alas its days or weeks of work, not hours.
Tuesday, September 15, 2009
Congradulations to AA and good luck to Masten.
AA flew the 180 last week, and Masten is going to attempt the 90 second flight tomorrow.
I think that the 180 second event will be a stretch for both Masten and Unreasonable.
The 90 second event may come down to who has the best GPS equipment.
Toward that end I've recently purchased a pair of Trimble BD960 L1-L2 Glossnass RTK Rover and Base GPS units with all the accuracy enhancing options. I also got what I consider to be a very favorable ruling on from the judges on what can be counted as payload weight.
I've been flying with a very light L1 GPS from Crescent, its a nice GPS, and flies well,but as a L1 GPS its just not accurate enoough if the prize will be won or lost on GPS accuracy.
So I asked the judges the following question:
If a camera is carried and used to enhance landing accuracy, can it be counted as payload?
(The history here is yes it can be payload)
Since a GPS position reporting is also part of the required payload can I carry a 2nd really accurate GPS report its position for the payload requirements and manually use the information to correct my posiition, just like a camera. I expected them to rule that cameras used for landing accuracy could not be counted as payload, instead The answer I got back kind of suprised me, yes the 2nd GPS can be payload as long as it is not required for flight, and it does not matter how the data is communicated to the vehicle. So it can even be automatic. you might think that this is a lot of fuss over a few ounces of electronics, but to get really really accurate GPS you need a good survey grade choke ring antenna, a good one can weigh as much as 12 pounds or more.
Now I can count that as payload!!!!! This should give me accuracy in the 2cm range, if only I could control the vehicle that tightly. I've started working on that with the helicopter a so far I'm consistently better than what armadillo reported for accuracy on their flights. how much better I won't reveal until the end of October.(The Helicopter can't carry the weight of my really good antenna, so I'm using a lesser antenna hence the vehicle should be at least as good.)
A sobering though is that it would be possible to accomplish both the 90 and 180 second tasks and still walk away from the competition empty handed. If Masten,Unreasonable both accomplish their goals, its very likely that one of the three teams will get nothing.
For example:
Masten, Unreasonable and Armadillo all do the 180.
Masten and unreasonable are more accurate, Armadillo gets nothing. That is clearly the least likely scenario, but anything is possible.
Two years ago there was one team attempting and no winners.
This year its a real possibility that all the prizes will be won with more successes than winners.
Also today was the last possible day to register for the event, so we should also hear from Xprize in the next few days to learn if there are any surprise teams registering at the last minute.
I think that the 180 second event will be a stretch for both Masten and Unreasonable.
The 90 second event may come down to who has the best GPS equipment.
Toward that end I've recently purchased a pair of Trimble BD960 L1-L2 Glossnass RTK Rover and Base GPS units with all the accuracy enhancing options. I also got what I consider to be a very favorable ruling on from the judges on what can be counted as payload weight.
I've been flying with a very light L1 GPS from Crescent, its a nice GPS, and flies well,but as a L1 GPS its just not accurate enoough if the prize will be won or lost on GPS accuracy.
So I asked the judges the following question:
If a camera is carried and used to enhance landing accuracy, can it be counted as payload?
(The history here is yes it can be payload)
Since a GPS position reporting is also part of the required payload can I carry a 2nd really accurate GPS report its position for the payload requirements and manually use the information to correct my posiition, just like a camera. I expected them to rule that cameras used for landing accuracy could not be counted as payload, instead The answer I got back kind of suprised me, yes the 2nd GPS can be payload as long as it is not required for flight, and it does not matter how the data is communicated to the vehicle. So it can even be automatic. you might think that this is a lot of fuss over a few ounces of electronics, but to get really really accurate GPS you need a good survey grade choke ring antenna, a good one can weigh as much as 12 pounds or more.
Now I can count that as payload!!!!! This should give me accuracy in the 2cm range, if only I could control the vehicle that tightly. I've started working on that with the helicopter a so far I'm consistently better than what armadillo reported for accuracy on their flights. how much better I won't reveal until the end of October.(The Helicopter can't carry the weight of my really good antenna, so I'm using a lesser antenna hence the vehicle should be at least as good.)
A sobering though is that it would be possible to accomplish both the 90 and 180 second tasks and still walk away from the competition empty handed. If Masten,Unreasonable both accomplish their goals, its very likely that one of the three teams will get nothing.
For example:
Masten, Unreasonable and Armadillo all do the 180.
Masten and unreasonable are more accurate, Armadillo gets nothing. That is clearly the least likely scenario, but anything is possible.
Two years ago there was one team attempting and no winners.
This year its a real possibility that all the prizes will be won with more successes than winners.
Also today was the last possible day to register for the event, so we should also hear from Xprize in the next few days to learn if there are any surprise teams registering at the last minute.
Sunday, September 13, 2009
106 Second Hover
We did a 106 second stable hover with the blue ball.
We also started working on the PAD forms for the LLC pads.
More info later...
Also started tweeting as unrocket
Will probably keep that up until the contest....
47 days and 6 testing weekends left.
We also started working on the PAD forms for the LLC pads.
More info later...
Also started tweeting as unrocket
Will probably keep that up until the contest....
47 days and 6 testing weekends left.
Wednesday, September 09, 2009
Working on Rockets...
I've taken the week off from the day job to be Unreasonable. I finished the last of the machining on the new stainless motor, I picked up the chamber liner from flame spray and now all the parts are at the welder. I received all the turned stainless parts today, so I have spares.
To finish my earlier discussion about sending parts out I talked to the vendor and he actually did not have any scrap. All his problems or issues were found in dry runing the programs on his CAM software, so no scrap generated.
The Blue Ball software is ready to test again, we may do so this weekend.
I've still got to get the video working correctly.
51 days and 7 testing weekends to go.....
To finish my earlier discussion about sending parts out I talked to the vendor and he actually did not have any scrap. All his problems or issues were found in dry runing the programs on his CAM software, so no scrap generated.
The Blue Ball software is ready to test again, we may do so this weekend.
I've still got to get the video working correctly.
51 days and 7 testing weekends to go.....
Sunday, September 06, 2009
Mixed Results....
Good news:
The goal for this weekend was to resolve the "heavy wiggles" and fly a clean 90 second tethered flight. We resolved the wiggles and we flew a flight somewhere between 89 and 91 seconds.
We had significant takeoff throttle margin so I can add some more propellant and I'm sould be able to get to 95 or 100 seconds.
Bad News:
Our wireless video system did not work, it looks like the cable coming out of the video camera was intermittent. We did not have wiggle problems, but we had roll, or yaw problems holding a consistent heading. We actually "spun" to the point of the tethers winding up and forcing an abort on a flight. We tried to fix this 4 times and nothing we did had any effect.
After we got home and I reviewed te code, I learned it had no effect be cause the variable set up to be "Rudder" gain was not connected to the loop controlling the "Rudder" (Remember this software was developed on a helicopter and hence has helicopter type coordinates rather than more typical rocket names for the axis.) so I could turn it up to infinity and it would still do nothing.
So we need to try the test again next weekend.
Scary News:
We have 7 more testing weekends and 53 total days before the event .
The goal for this weekend was to resolve the "heavy wiggles" and fly a clean 90 second tethered flight. We resolved the wiggles and we flew a flight somewhere between 89 and 91 seconds.
We had significant takeoff throttle margin so I can add some more propellant and I'm sould be able to get to 95 or 100 seconds.
Bad News:
Our wireless video system did not work, it looks like the cable coming out of the video camera was intermittent. We did not have wiggle problems, but we had roll, or yaw problems holding a consistent heading. We actually "spun" to the point of the tethers winding up and forcing an abort on a flight. We tried to fix this 4 times and nothing we did had any effect.
After we got home and I reviewed te code, I learned it had no effect be cause the variable set up to be "Rudder" gain was not connected to the loop controlling the "Rudder" (Remember this software was developed on a helicopter and hence has helicopter type coordinates rather than more typical rocket names for the axis.) so I could turn it up to infinity and it would still do nothing.
So we need to try the test again next weekend.
Scary News:
We have 7 more testing weekends and 53 total days before the event .
Friday, September 04, 2009
FAR FAR Away again...
I just finished packing for a trip to FAR on Saturday, we will be attempting a 90 second tethered hover with the blue ball. This includes the live video so all the rules requirements for the vehicle are complete. Other than resolve the heavy weight stability issues we have no more fabrication
to do on the blue ball.
On the Silver ball I finished slotting the stainless chamber and machining the injector plate. Both parts were dropped off to be flame sprayed with a thermal barrier coating today.
(Last time I went by the flame spray shop the gave me a quick shop tour and it looked like there were a bunch of merlin sized chambers in there to be sprayed.
I also dropped the next set of catalyst screens off to be plated.
Next week if all goes well the silver ball will be hardware complete.
to do on the blue ball.
On the Silver ball I finished slotting the stainless chamber and machining the injector plate. Both parts were dropped off to be flame sprayed with a thermal barrier coating today.
(Last time I went by the flame spray shop the gave me a quick shop tour and it looked like there were a bunch of merlin sized chambers in there to be sprayed.
I also dropped the next set of catalyst screens off to be plated.
Next week if all goes well the silver ball will be hardware complete.
Tuesday, September 01, 2009
My weekend with the Stainless Bully...
Team Phoneicia's blog had a quote Aluminum is your friend stainless is the sadistic bully from third grade. The part above has not been cleaned up yet, but its basically done it started out as a 1" thick 8" diameter 316 stainless. The back side is taped for the nozzles, all I have to do is clean up and drill the holes coming in from the stubs sticking out.
Subscribe to:
Posts (Atom)