120 G flight yesterday

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Adrian A

Well-Known Member
TRF Sponsor
TRF Supporter
Joined
Jan 21, 2009
Messages
3,185
Reaction score
2,912
Location
Lakewood, CO
Yesterday I had a fun flight at Tripoli Colorado in Hartsel that tied or maybe broke my personal record for highest-G boost at 120.1Gs.

TBH, there's not a lot to see in this video below, though I think one frame did catch the rocket. The sound is kind of fun.

View attachment 20230723_111032~2.mp4


1690235084386.png
Some fun facts from the Blue Raven data:
  • The boost was 0.50 seconds long, but in the video it sounds longer because it broke Mach 1.0 at 0.38 seconds, so the sounds got stretched out.
  • It broke Mach 1 just 200 feet above the pad.
  • The coast time was 97.8% of the total time to apogee.
  • It cleared the12 foot tower in 0.112 seconds, at 239 feet/second (163 mph)
The rocket is a 29mm minimum diameter rocket that I use for testing electronics. It weighs 228 grams without the motor.

IMG-1326.jpg

The nosecone is an Apogee 29mm plastic cone that I turned into a VK nosecone by molding some filled epoxy around it, in a silicone mold I made around 2011. The main chute compartment is a standard FG tube, and the aft end is built of some leftover tubes I had from way back when, joined by a coupler that also serves as the motor thrust ring. The av-bay shown in front is from 2 pre-production Blue Raven 29mm altimeters, and two 110 mAhr batteries. The batteries held up without a recharge for 2 flights yesterday. The av-bay, including some sealing bulkheads, harness attachment bulkheads, and charge holders that aren't shown, weighs 32 grams.

Unfortunately, the fins aren't nearly straight enough, which I saw as soon as I looked at the Blue Raven flight summary that was available on my phone as soon as I walked up to the rocket. Note the time to gyro overload and the number of ascent rolls:
IMG-1325.PNG

After downloading the data, indeed the gyro got pegged at 2294 deg/seconds right away:
1690236602613.png
Despite that, it had reasonable results for the tilt angle, with the GPS flight angle and the tilt passing through 90 degrees at close to the same time:
1690239934250.png
The short-duration oscillations are the rocket coning. Each little bump represents another rotation. (Those fins were really not on very straight). The future tilt is what the Blue Raven calculates during the ascent for the predicted tilt 3 seconds into the future. This is useful for triggering airstarts that have some ignition delay.

Here's the altitude data, zoomed out. The Baro sensor has the expected under-prediction of altitude due to using the standard atmosphere model when it's hot out.
1690237618104.png
Zooming in on the first part of the flight you can see baro altitude artifacts from going above and below Mach 1:
1690238152457.png
The rocket took less than 1 second to get the first 1000 feet.
The GPS took a little while to figure out that someone hit it with a near-instant impulse to Mach 1.28. I was happy with the GPS performance, though. This is the updated Featherweight GPS tracker with the latest UBlox M10 receiver, which tracks an impressive number of satellites from all 4 constellations:

1690237787370.png
Here are the accelerations during the boost:

1690238722055.png
While I'll claim this as a 120G boost, mostly it's around 100 Gs. You can see the rocket bouncing within the rails of the 12' tower in the X and Y axes in the first 0.12 seconds.

I'll post more data in a minute, with a fresh allocation of images
 

Attachments

  • 1690238642270.png
    1690238642270.png
    41 KB · Views: 0
That would have been fun to see... or not, since you probably saw a puff of smoke on the pad and that was about it. You're using an ADXL375 for z-axis accel, right?
 
What warp 9 motor was used?
CTI H410, sadly one of the few remaining.
That would have been fun to see... or not, since you probably saw a puff of smoke on the pad and that was about it. You're using an ADXL375 for z-axis accel, right?
No, the STM H3LIS331 for 3-axis for up to 400 Gs, along with the LSM6DS032 when it's less than 32 Gs.

Here's the apogee deployment. YO ucan see the 3 shock cord reefs and then when the shock cord loses its slack. This is right where I want to have this deployment charge. It was 0.28 grams.
1690244230586.png
The main chute deployment was more violent:

1690244321457.png

The first negative impulse is the charge firing, and then the big positive impulse is the nosecone getting to the end of the cord. There is only 0.06 seconds between them, so I'd say this charge was a little over-sized even though it was just 0.24 grams.
 
Just curious whether you're running the H3LIS331 at 100, 200, or 400 g's? I'm at 200 on my setup.
 
That's awesome!! I've got 2 of those motors that I'm looking forward to doing a similar rocket with.
Great looking flight and data!
 
120 g's is very impressive.
I did the motor for about a 100 g flight. Was in 2002 so electronics were '90s technology. Was an accelerometer altimeter imbedded in epoxy. Recovered intact.
L5300, 0.54 seconds

Everyone is looking at launch pad!

View attachment 593862
View attachment 593863
Mark , the start up pressurization noise the core of that motor made was epic.
 
Just curious whether you're running the H3LIS331 at 100, 200, or 400 g's? I'm at 200 on my setup.
The 400G range, though I record the data with 16 bit integers and .01 G resolution, so anything beyond 327.68 Gs gets a sign inversion. Early on when I was experimenting with it I found that the lower ranges for that chip didn't actually provide more resolution; they just bit-shift the result and add a zero. Not sure if that would have been the case for other combinations of settings. But deployment charges and landing impacts are often more than 200 Gs anyway.
 
The 400G range, though I record the data with 16 bit integers and .01 G resolution, so anything beyond 327.68 Gs gets a sign inversion. Early on when I was experimenting with it I found that the lower ranges for that chip didn't actually provide more resolution; they just bit-shift the result and add a zero. Not sure if that would have been the case for other combinations of settings. But deployment charges and landing impacts are often more than 200 Gs anyway.
Interesting, thanks. I was pretty locked into "no way I'd peg 200 g's on boost" without thinking much of other events that might be interesting.

Yes, it's been a while since I looked at it, but I remember it throwing me for a loop until I took a closer look at the documentation. I think they output a 12 bit number.
 
Very Nice, Adrian !

Q1: Why did the Inertial Altitude data abruptly stop at about 2.5 sec in the Inertial, Baro, GPS Altitude -vs- Time Plot ?
Adrian-Unit-B-Altitude-C30723-510656-5f6c4a287c7147c6655a5eeafb61e971.png
Q2: Have you ever seen a correlation between (x,y,z) acceleration and tilt angle -vs- time ( maybe capturing induced drag while the rocket is coning ) ?

Thanks !

-- kjh( :) living vicariously thru your incredible flight data :) )
 
120 g's is very impressive.
I did the motor for about a 100 g flight. Was in 2002 so electronics were '90s technology. Was an accelerometer altimeter imbedded in epoxy. Recovered intact.
L5300, 0.54 seconds

Everyone is looking at launch pad!

Me too, @MClark :)

Amazing motor and flight, not to mention that you recovered the rocket intact !

I took that new job and quit rocketry too soon ( circa 2000 ) :(

-- kjh
 
Very Nice, Adrian !

Q1: Why did the Inertial Altitude data abruptly stop at about 2.5 sec in the Inertial, Baro, GPS Altitude -vs- Time Plot ?
View attachment 593890
Q2: Have you ever seen a correlation between (x,y,z) acceleration and tilt angle -vs- time ( maybe capturing induced drag while the rocket is coning ) ?

Thanks !

-- kjh( :) living vicariously thru your incredible flight data :) )

I cut the plot off there because that's about the last point where the the inertial estimate was providing better information than the other measurements. The roll rate gyro maxed out at 2294 deg/second, which is 6.4 revolutions per second (384 rpm), and this rocket was spinning that fast 0.3 seconds after ignition, a little more than halfway through the burn. After that, the inertial navigation couldn't keep up and so it didn't couldn't keep track of where the lateral acceleration axes were pointing when they were measuring the large centripetal accelerations from the spinning. Some of that got erroneously estimated to be in the vertical direction, so the inertial navigation estimate for apogee was over 12,000 feet. I need to rebuild the aft section of this test rocket with straighter fins so that the the roll rate stays reasonable and the inertial navigation has a chance.

Q2 I'm not sure how you would separate induced drag from coning from regular drag for most flights. But if any dataset shows that, it would be the sustainer separation from my NSL 2-stage flight that we have discussed, where I had a poorly-aligned nosecone and the drag went up after separation from the booster that was keeping it straight.
 
Adrian, you do absolutely silly things. And its absolutely awesome!

Placing an order for a GPS in a moment to augment the Raven Blue. Gotta spend the rest of my credit card rewards on something. lol.
 
Thats awesome. I need some sort of electronics like that soon, but I am clueless right now to that. I've been watching videos and looking at Eggfinders and really want an Altimeter 3 from JL. But, might have to do something different....Very Cool!!
 
The 400G range, though I record the data with 16 bit integers and .01 G resolution, so anything beyond 327.68 Gs gets a sign inversion. Early on when I was experimenting with it I found that the lower ranges for that chip didn't actually provide more resolution; they just bit-shift the result and add a zero. Not sure if that would have been the case for other combinations of settings. But deployment charges and landing impacts are often more than 200 Gs anyway.
I have some vague plans for a rocket that will hit some truly stupid accelerations. What would happen if you exceeded 327.68 Gs on Blue Raven? Could it be run in baro only, and still be mach safe? Or would it be safer to use a barometric altimeter for deployment and just use the Blue Raven as a data logger?
 
Yesterday I had a fun flight at Tripoli Colorado in Hartsel that tied or maybe broke my personal record for highest-G boost at 120.1Gs.

TBH, there's not a lot to see in this video below, though I think one frame did catch the rocket. The sound is kind of fun.

View attachment 593601


View attachment 593775
Some fun facts from the Blue Raven data:
  • The boost was 0.50 seconds long, but in the video it sounds longer because it broke Mach 1.0 at 0.38 seconds, so the sounds got stretched out.
  • It broke Mach 1 just 200 feet above the pad.
  • The coast time was 97.8% of the total time to apogee.
  • It cleared the12 foot tower in 0.112 seconds, at 239 feet/second (163 mph)
The rocket is a 29mm minimum diameter rocket that I use for testing electronics. It weighs 228 grams without the motor.

View attachment 593777

The nosecone is an Apogee 29mm plastic cone that I turned into a VK nosecone by molding some filled epoxy around it, in a silicone mold I made around 2011. The main chute compartment is a standard FG tube, and the aft end is built of some leftover tubes I had from way back when, joined by a coupler that also serves as the motor thrust ring. The av-bay shown in front is from 2 pre-production Blue Raven 29mm altimeters, and two 110 mAhr batteries. The batteries held up without a recharge for 2 flights yesterday. The av-bay, including some sealing bulkheads, harness attachment bulkheads, and charge holders that aren't shown, weighs 32 grams.

Unfortunately, the fins aren't nearly straight enough, which I saw as soon as I looked at the Blue Raven flight summary that was available on my phone as soon as I walked up to the rocket. Note the time to gyro overload and the number of ascent rolls:
View attachment 593779

After downloading the data, indeed the gyro got pegged at 2294 deg/seconds right away:
View attachment 593783
Despite that, it had reasonable results for the tilt angle, with the GPS flight angle and the tilt passing through 90 degrees at close to the same time:
View attachment 593811
The short-duration oscillations are the rocket coning. Each little bump represents another rotation. (Those fins were really not on very straight). The future tilt is what the Blue Raven calculates during the ascent for the predicted tilt 3 seconds into the future. This is useful for triggering airstarts that have some ignition delay.

Here's the altitude data, zoomed out. The Baro sensor has the expected under-prediction of altitude due to using the standard atmosphere model when it's hot out.
View attachment 593785
Zooming in on the first part of the flight you can see baro altitude artifacts from going above and below Mach 1:
View attachment 593788
The rocket took less than 1 second to get the first 1000 feet.
The GPS took a little while to figure out that someone hit it with a near-instant impulse to Mach 1.28. I was happy with the GPS performance, though. This is the updated Featherweight GPS tracker with the latest UBlox M10 receiver, which tracks an impressive number of satellites from all 4 constellations:

View attachment 593786
Here are the accelerations during the boost:

View attachment 593802
While I'll claim this as a 120G boost, mostly it's around 100 Gs. You can see the rocket bouncing within the rails of the 12' tower in the X and Y axes in the first 0.12 seconds.

I'll post more data in a minute, with a fresh allocation of images
Awesome !!! What are your thoughts about the tilt calculation staying accurate (or at least useful) even though the roll gyro saturated for quite a few seconds ? For my homebrew systems the integration of the roll gyro factors into the overall tilt calculation. (Full disclosure - I don't use quaternions. I'm scared LOL)
 
Are those contiguous frames ?

It looks that way, but nope. Done by Cro-Magnon method of pause, then click the tiniest amount forward, using whatever player comes up in post #1 in full screen mode. Repeat about 50 times, and those are the only 4 frames I could get with the rocket moving.

Screenshot, crop, and add to new panoramic image with IrfanView.

There’s correlation between time and height. Since time is in whole seconds, you couldn’t do real calculations without advancing the video more precisely.
 
It looks that way, but nope. Done by Cro-Magnon method of pause, then click the tiniest amount forward, using whatever player comes up in post #1 in full screen mode. Repeat about 50 times, and those are the only 4 frames I could get with the rocket moving.

Screenshot, crop, and add to new panoramic image with IrfanView.

There’s correlation between time and height. Since time is in whole seconds, you couldn’t do real calculations without advancing the video more precisely.
You guessed at what I was wondering, Tim ( calculating Velocity from distance traveled / time in your screen shots )

-- kjh
 
I have some vague plans for a rocket that will hit some truly stupid accelerations. What would happen if you exceeded 327.68 Gs on Blue Raven? Could it be run in baro only, and still be mach safe? Or would it be safer to use a barometric altimeter for deployment and just use the Blue Raven as a data logger?
Internally, the Blue Raven handles the full +/- 400 G range of the accelerometer; the 327.68 limitation is just on the recording. Beyond 400 Gs, the acclerometer probably just rails and the default apogee detection would work fine, since it's voting 2 out of 3 sensors, including the baro sensor and the gyro tilt.
With careful clicking, I managed to freeze 4 frames that captured the rocket. Open full-size to see it:
Nice!!
Awesome !!! What are your thoughts about the tilt calculation staying accurate (or at least useful) even though the roll gyro saturated for quite a few seconds ? For my homebrew systems the integration of the roll gyro factors into the overall tilt calculation. (Full disclosure - I don't use quaternions. I'm scared LOL)

I would recommend just using quaternions. Algorithms are found online and they're not that hard to implement.

I would guess that the reason why the tilt estimation was reasonable even though it lost track of the roll orientation is because the motion in the pitch and yaw axes was small and oscillating. Even if they weren't in the Earth-centered right direction, they probably mostly cancelled each other out on this flight. If the rocket had a lot of weathercocking or some other larger pitch event during the flight it probably would have been a different story.
Neutronium95,
I exceeded the acceleration several times with the RavenV4 (+100G). Adrian recommended I still use acceleration based apogee. It works every time, the data and graph just max out and shows you a flat line.
If the acclerometer rails at 400 Gs (the STM sensors tend to actually rail at spec value + ~10%) then the Blue Raven would underestimate the vertical velocity and estimate apogee somewhat early.
 
Internally, the Blue Raven handles the full +/- 400 G range of the accelerometer; the 327.68 limitation is just on the recording. Beyond 400 Gs, the acclerometer probably just rails and the default apogee detection would work fine, since it's voting 2 out of 3 sensors, including the baro sensor and the gyro tilt.
Thanks for clarifying that. I was just concerned that there might be some weird overflow error when exceeding 327.68 G. I haven't nailed down many of the precise details of my design, but I think that I can say for certain that it'll only hit 400G if the motor fails.
 
The 400G range, though I record the data with 16 bit integers and .01 G resolution, so anything beyond 327.68 Gs gets a sign inversion. Early on when I was experimenting with it I found that the lower ranges for that chip didn't actually provide more resolution; they just bit-shift the result and add a zero. Not sure if that would have been the case for other combinations of settings. But deployment charges and landing impacts are often more than 200 Gs anyway.
Adrian

You might find this data interesting from a H3LIS331. This is a 900 G lithobraking recovery captured at 500Hz with the sensor set to the 200 G range. The rocket and electronics survived, thanks to the mentoring from the late Dr. Gerald Bull. 900 G Raw Data Impact.png
 
Back
Top