Rough Accelerometer Data from a Blue Raven in a 38mm Av-Bay -- Is the noise for real ?

[kjhambrick]

Maybe this should be a private PM for @Adrian A but then again, I wonder if anyone else has seen similar noise in their raw Blue Raven acceleration data ...

I've got six flights now with a Blue Raven -- the first four were with a Power Perch on a plank and the two most recent flights were in a Featherweight 38mm Av-Bay.

I am flying the Featherweight 38mm Av-Bay package 'Active Side Down' so that if I ever need the 4th e-match output channel, it will be easily accessible in the booster.

This means my axial accelerometer and gyro data is inverted which the Blue Raven handles automatically for event control.

My first flight with the FW Av-Bay was an AT H180W to about 3500 ft in "T'Pring's P'Toy" -- a 675 gram ( dry mass ), 1.9-inch rocket.

These are Raw Pressure Altitude, Density Altitude and Inertial Altitude -vs- time and Acceleration and Air Density -vs- Velocity for the H180W flight:


Note the very rough acceleration data in the 2nd plot.

After that first flight, I wondered if the free-end of the Blue Raven was 'ringing' like a tuning fork so I added a nylon hold-down for the top, free end before my 2nd flight in the same rocket:


My 2nd flight in T'Pring's P'Toy was on an AT I161W to about 5000 ft:

The top-end hold-down didn't seem to fix the noisy acceleration data ...

I don't think the rough acceleration is due to any real physical phenomena because I would think something would also show up in the gyro data and thereare no obvious blips in the gyro data.

These are axial inertial and 3-DoF gyro data during the thrust phase for the two flights:

I've also got four flights with the same Blue Raven in a Power Perch where I fly the Blue Raven right-side up and none of those flights exhibit the accelerometer noise !

So I finally looked at the end of the flight summary.csv files. This is the most recent I161W flight:

Direction of 1st motion | -0.9998| -0.0085| -0.0175
Gravity direction on pad | -0.9992| 0.0029| -0.0395
Low-G accel pad offset | 0.0092| -0.0000| 0.0004
Hi-G accel pad offset | -1.53| -20.45| 0.94
Pad gyro rates (deg/sec) | 0.16| -0.19| 0.47
Pad Hi-G Accel | -2.53| -20.45| 0.90
Pad low-G Accel | -0.99| 0.00| -0.04

Q1: Could the noise be due to the Blue Raven Firmware switching between the Low Range and High Range Accelerometer ?

Q2: Do I need to recalibrate my Blue Raven ?

Thanks !

-- kjh

EDIT: P.S. I have swapped out Blue Raven S/N 0236 from above with S/N 0828 which does not show the Hi-G offsets:
Direction of 1st motion |-0.9947 |-0.0157 |0.1014
Gravity direction on pad |-0.9999 |-0.0152 |0.0006
Low-G accel pad offset |0.0009 |0.0000 |-0.0000
Hi-G accel pad offset |-0.00 |0.83 |2.41
Pad gyro rates (deg/sec) |-0.89 |0.73 |0.64
Pad Hi-G Accel |-1.00 |0.81 |2.41
Pad low-G Accel |-1.00 |-0.02 |0.00

I'll try to fly an H180W in the same rocket at the next AARG launch or maybe at the TNT Seymour Launch next weekend if I can make the launch and I am feeling lucky

 

[Adrian A]

Those motor boost periods look plausible to me. These are just noisy motors, and the 38mm av-bay setup makes a stiff connection from the airframe to the altimeter. The high-G accelerometer only gets used when its measured value goes over 32 Gs.

The pad offsets are measured on the pad based on the low-G accelerometer and subtracted off before the measurements are used, so I'm pretty sure that all the above accelerations are measured by the low-range accelerometer. It wouldn't hurt to re-calibrate your unit, which will make those pad offsets go down, and should slightly improve your inertial navigation estimates.

 

[kjhambrick]

Those motor boost periods look plausible to me. These are just noisy motors, and the 38mm av-bay setup makes a stiff connection from the airframe to the altimeter. The high-G accelerometer only gets used when its measured value goes over 32 Gs.

The pad offsets are measured on the pad based on the low-G accelerometer and subtracted off before the measurements are used, so I'm pretty sure that all the above accelerations are measured by the low-range accelerometer. It wouldn't hurt to re-calibrate your unit, which will make those pad offsets go down, and should slightly improve your inertial navigation estimates.

Thanks Adrian !

I'll install S/N 0236 on the Power Perch and recalibrate just for fun

I'll also follow up with the H180W flight data when I have something to post.

-- kjh

p.s. I just now installed version 1.0.9 (309) on my phone which reminded me why I was confused on 32G -vs- 16G for the low-G accelerometer.

The doc's mention that the low-G accelerometer is +/- 32G but the Blue Raven Android App shows "Accel 16g" on-screen so I wasn't sure where the switch from the Low-G accelerometer to the High-G accelerometer was happening in the on-board Blue Raven ...

 

[Spacedog49Krell]

I find some batches of White Lightning propellant to be very noisy burning. The noise resolution is compounded when you're capturing data at 500Hz.

 

[kjhambrick]

I find some batches of White Lightning propellant to be very noisy burning. The noise resolution is compounded when you're capturing data at 500Hz.

Thanks @Spacedog49Krell

All my past experience is with AltAcc data ( 8-bit data at 16 Hz ) so I have never witnessed high-frequency events like these before.

I am still trying to wrap my head around how the net force on the system could cause a 645 gram rocket to undergo the rapid high delta-v indicated by the axial accelerometer readings.

But based on @Adrian A's and your posts, I believe the readings ...

My next flight will be the same rocket and another H180W with a different Blue Raven to compare accelerometer data.

-- kjh

p.s. I've got three more recent flights to post on the Off the rail velocity POLL thread ( I357T, H180W, I161W ) -- punch line is that I have been able to identify the rocket leaving the rail in all three flights and the axial distances and velocities are all in the ballpark ...

 

[kjhambrick]

Hmmm ... @Adrian A said ... noisy motors and a stiff connection to the airframe ... and @Spacedog49Krell said ... some batches of White Lightning propellant are very noisy burning ...

I am using four, 5.25-inch, #2-56 316-stainless steel threaded rods to install the Featherweight 38mm Av-Bay in a 4-inch Aerotech 1.9 inch coupler tube:

Maybe my threaded rods are buzzing with the motor noise ?

I'll try adding a stretched rubber band or a tight wrap of masking tape around the four rods before I button up the Av-Bay for my next flight ...

-- kjh

 

[Adrian A]

p.s. I just now installed version 1.0.9 (309) on my phone which reminded me why I was confused on 32G -vs- 16G for the low-G accelerometer.

The doc's mention that the low-G accelerometer is +/- 32G but the Blue Raven Android App shows "Accel 16g" on-screen so I wasn't sure where the switch from the Low-G accelerometer to the High-G accelerometer was happening in the on-board Blue Raven ...

Wow, I can’t believe I have been overlooking that!

 

[kjhambrick]

Wow, I can’t believe I have been overlooking that!

NP Adrian.

Being a software developer, all-too-often I 'see' what I expect to see instead of what's actually there

-- kjh

 

[Spacedog49Krell]

Thanks @Spacedog49Krell

All my past experience is with AltAcc data ( 8-bit data at 16 Hz ) so I have never witnessed high-frequency events like these before.

I am still trying to wrap my head around how the net force on the system could cause a 645 gram rocket to undergo the rapid high delta-v indicated by the axial accelerometer readings.

But based on @Adrian A's and your posts, I believe the readings ...

My next flight will be the same rocket and another H180W with a different Blue Raven to compare accelerometer data.

-- kjh

p.s. I've got three more recent flights to post on the Off the rail velocity POLL thread ( I357T, H180W, I161W ) -- punch line is that I have been able to identify the rocket leaving the rail in all three flights and the axial distances and velocities are all in the ballpark ...

Building my own flight computers and making progressive speed improvements from 24Hz, 37Hz, 50Hz, 100Hz, and 400/500Hz. I wasn't as shocked by the data resolution changes, as you, going from 16Hz to 500Hz. Also, one of your "Off the rail velocity" graphs was on a I225FJ Black Max. AT's Classic and Black Max propellants are smooth burning formulations compared to White Lightning or Dark Matter formulations.

One real event I see in your data is the gyro noise on the I161W flight. I've seen other friend's flight data showing aerodynamic gyro noise. The 800-1100 ft/sec region is very noisy.