Getting high...really high....how do I know I'm there????

[FredA]

Jim,

Thanks for the input.
I'll go back and talk to Keith again about using tilt - not a bad idea.

We hope NOT to spin the rocket as the primary "mission" goal is to be a camera platform.
However, I wouldn't be surprised if the FAA demands spin....we'll see on that.

If we are forced to spin, deployment as close as possible to apogee will become even more desirable.
Deployment will arrest much of the spin which will be required for decent photography.

The rocket is a rebuild of our half-scale Nike Apache shown below taking off at Balls-20 on a "O."
The booster had a deployment issue and broke a fin on landing - thus we want to rebuild it and plan to make it "all motor" to get to the "R."
The sustainer recovered fine and we don't plan to mess with it much - just an avionics refresh and new/more cameras.
There isn't room in the sustainer for any anti-roll device.

 

[jderimig]

The problem with tilt or MAD is that if you use just a little spin stabilization there is not enough atmosphere to assure that the rocket will tilt over at apogee. Gravity doesn't provide the moment to pitch the rocket, drag does. You can back slide quite a bit until some atmosphere is encountered. By that time your ~100K barometric sensors will start reading again, either way your apogee detection will be very late.

I am working also on a very high BALLS rocket design. Working on the same problem. Don't need GPS,

PS. If you spin that makes tilt detection with consumer gyro devices possibly ruled out unless you can be sure you can keep the spin rate below that max limit of the gyro. If you do that you also will be operating the gyro at its lowest resolution setting probably.

 

[cerving]

I would think this might be a 3-tier prediction. 1) Use baro-only to about 30,000' while adjusting the accelerometer based on the baro data and/or IMU for off-axis. 2) Fuse the baro-accelerometer, reducing the baro-accelerometer ratio dynamically until you're at the limit of the baro (which according to TE-MeasSpec is about 80,000' with a resolution of about 100'). 3) At 80,000', just set a timer for the final 20,000' or so based on your predicted apogee from that point. That would take care of just about all of the negatives.

 

[JimJarvis50]

The problem with tilt or MAD is that if you use just a little spin stabilization there is not enough atmosphere to assure that the rocket will tilt over at apogee. Gravity doesn't provide the moment to pitch the rocket, drag does. You can back slide quite a bit until some atmosphere is encountered. By that time your ~100K barometric sensors will start reading again, either way your apogee detection will be very late.

I am working also on a very high BALLS rocket design. Working on the same problem. Don't need GPS,

PS. If you spin that makes tilt detection with consumer gyro devices possibly ruled out unless you can be sure you can keep the spin rate below that max limit of the gyro. If you do that you also will be operating the gyro at its lowest resolution setting probably.

Keith could correct me, but I believe their products are set at 2000 degrees/sec by default (the highest available setting for the chip). I suppose backsliding could occur (I've never been that lucky), so I would still say to use a timer to ensure that something happens no matter what. In my latest flight, I calculated the time to apogee and then set the timers about 20 seconds early to account for non-ideal flight and because early deployment wouldn't harm anything if it was indeed early.

Jim

 

[jderimig]

I would think this might be a 3-tier prediction. 1) Use baro-only to about 30,000' while adjusting the accelerometer based on the baro data and/or IMU for off-axis. 2) Fuse the baro-accelerometer, reducing the baro-accelerometer ratio dynamically until you're at the limit of the baro (which according to TE-MeasSpec is about 80,000' with a resolution of about 100'). 3) At 80,000', just set a timer for the final 20,000' or so based on your predicted apogee from that point. That would take care of just about all of the negatives.

I think existing Kalman filter code will do exactly that (this will be my approach). You can change the Kalman gain dynamically as altitude increases to trust the integration step more and baro correction less. I think the timer is unnecessary and has the same problem as off axis. You cannot know the proper delay unless you know the off axis amount.

 

[JimJarvis50]

One thing I learned preparing for my last flight was that the impact of "early" deployment on a high altitude flight was pretty minimal. About 5K feet as I recall, and that assumes that the rocket stops going up after deployment, which it won't. But, that early deployment gives you some margin in case the flight is off axis or doesn't reach the simulation for some other reason.

I wouldn't really expect off axis flight to be a significant issue with respect to apogee time. I would expect that Fred will need tilt inhibition to get his flight approved, so large off-axis values shouldn't occur with an appropriate tilt criteria.

I would strongly recommend using a timer no matter what else might be included for apogee detection. I think we could probably invent a scenario (such as back sliding) that might fool any other apogee detection method that might be used. You want something that will positively happen if other methods fail. Cheap insurance.

Jim

 

[FredA]

I do own [one of the few] tilt-o-meters.
That and a MARSA-54 is the current sustainer avionics....but that was for our previous flight which was staying under 100k.
So we will have tilt-inhibit onboard.
Just the apogee sense that I'm pondering.

Setting a timer to 20-seconds early seems [excuse me Jim] a little foolish.
Wouldn't "on-time" be better with the assumption that 20-seconds late is inconsequential if apogee occurs as planned.
Thin air is thin air whether you are going up or down and velocity on the "far" side of the parabola will be less or equal to the near side.

 

[JimJarvis50]

I do own [one of the few] tilt-o-meters.
That and a MARSA-54 is the current sustainer avionics....but that was for our previous flight which was staying under 100k.
So we will have tilt-inhibit onboard.
Just the apogee sense that I'm pondering.

Setting a timer to 20-seconds early seems [excuse me Jim] a little foolish.
Wouldn't "on-time" be better with the assumption that 20-seconds late is inconsequential if apogee occurs as planned.
Thin air is thin air whether you are going up or down and velocity on the "far" side of the parabola will be less or equal to the near side.

I've had experience with both the Telemetrum products and the tiltometer. I would not recommend that you use the tiltmeter for your project. The technology on the Tiltometer is behind that of the Telemetrum stuff. Also, the apparent tilt on the Tiltometer goes up with positive acceleration and then down with the subsequent decelleration. I have had flights where that acceleration is enough to reach the abort criteria, and I had one flight abort because of that. Assuming you pick a tilt criteria that passes the smile test, say 10 degrees, and since the Tiltometer will permanently inhibit the flight if that criteria is reached, you might shut down the flight when tilt isn't actually there. I think you would have a better chance of success with the Telemetrum product.

The 20 seconds early is not foolish. The 20 seconds is the time less than the calculated apogee from the simulation. My flights typically go to 90% of the simulation, so I'm simply allowing for my experience to date. If you use RasAero to do your simulation, you will get a result that you probably won't exceed. That is, you will likely be lower than the simulation but not likely higher. Setting the time a little on the short side just lets you cover the range where the flight likely will be.

I typically use barometric apogee if apogee occurs below 90K feet. That covers something that is really below the simulation, or the motor not lighting or whatever. The problem is the gap between 90K and whatever you think apogee might actually be. If you were to use something like tilt to cover that gap, then you could be a little more generous on the timer (but include the timer in any case).

I have a couple relevant flights that might give an indication how the Telemetrum might work using tilt for apogee. Let me dig up some data.

Jim

 

[JimJarvis50]

I have a couple relevant flights that might give an indication how the Telemetrum might work using tilt for apogee. Let me dig up some data.

Jim

Actually, I only have two relevant flights. One was a two-stage but only to 20K. Tilt apogee would have been fine. On my three-stage flight to 130K, tilt was under-predicted. I suspect this was due to rocket spin, although all readings were on-scale. Tilt was only 35 degrees when I know the rocket was horizontal. Maybe not such a good idea on the long flights.

Jim

 

[VernK]

I've had experience with both the Telemetrum products and the tiltometer. ...[snip]... Assuming you pick a tilt criteria that passes the smile test, say 10 degrees, and since the Tiltometer will permanently inhibit the flight if that criteria is reached, you might shut down the flight when tilt isn't actually there. I think you would have a better chance of success with the Telemetrum product.

I was just looking at the datasheet for the IMU used in the Altus Metrum TeleMega and EasyMega (which I think are the products that provide the tilt limits.) The IMU is an Invensense MPU-6000 according to their website. The datasheet specifies the gyro sensitivity to linear acceleration as 0.1 Deg/Sec/G "typical" (not "max"). Therefore, if the boost is say 25 G's for 8 seconds, then the gyro will typically drift 25*8*0.1 = 20 Degrees. On top of that would be drift from any other sources. So it seems to me that a 10 degree tilt limit is not really doable. Unless they are doing something really clever in the firmware to reduce this sensitivity.

 

[JimJarvis50]

I was just looking at the datasheet for the IMU used in the Altus Metrum TeleMega and EasyMega (which I think are the products that provide the tilt limits.) The IMU is an Invensense MPU-6000 according to their website. The datasheet specifies the gyro sensitivity to linear acceleration as 0.1 Deg/Sec/G "typical" (not "max"). Therefore, if the boost is say 25 G's for 8 seconds, then the gyro will typically drift 25*8*0.1 = 20 Degrees. On top of that would be drift from any other sources. So it seems to me that a 10 degree tilt limit is not really doable. Unless they are doing something really clever in the firmware to reduce this sensitivity.

I think if the boost is 25G's for 8 seconds, tilt will be the least of your problems!

Data from my three-stage flight a year ago is attached. The EasyMega did well in comparison to estimated tilt from the video, but started to diverge about 50 seconds into the flight. By apogee, it wasn't close. I guess this is just drift due to the accelerometers being turned off for the purpose of correcting gyro drift.

10 degrees may be a challenge. However, it's necessary for staying within the waivered radius that we now have. I used 20 degrees on the three stage flight that is attached. That was too high. This year, my setting was 10 degrees. If you run some simulations on a 200K flight, you'll find that you won't really want the motor to light if you're above 10 degrees.

Jim

 

[FredA]

Boost will be a progressive motor.
Take off will be at ~6G's and the motor (given the reduced mass) will be pushing something in the vicinity of 25G's at burnout.

Agree that lighting the sustainer with more than 10 degrees of tilt will put it in Winnemucca...

All this just underlines the predicament of needing really good avionics that don't exist at a amateur level.
Hard to swallow an avionics budget that exceeds the rest of the project..............becomes a show stopper for a one-of flight.