Quantum Detailed Data Question

The Rocketry Forum

Help Support The Rocketry Forum:

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

Zbench

Well-Known Member
TRF Supporter
Joined
Jan 30, 2018
Messages
476
Reaction score
477
Location
Broadview Heights, OH
On Saturday at our local NOTRA Tripoli Launch, we sent a 4" LOC Goblin up on an I327DM. I used a quantum in the nose, only using the drogue function to deploy the parachute which was bundled with a chute release and set to release at 500 feet.

I downloaded the detail file from the quantum just now because I wanted to see how efficient the parachute I stitched up was on the decent. The graph of both raw altitude and filtered altitude are show below vs the elapsed time.

My question is this, you can see where the chute smooths out the rate of decent at the point shown below. I calculated the actual descent rate ( I think), by subtracting the altitude where the rate stabilized from the altitude on landing, and in a similar fashion, the elapsed time. That came to 352 feet in 41 seconds. By the powers of higher math, I calculate that to be 8.6 F/S.

In excel, mousing over the same data range for the velocity in both raw velocity and filtered velocity, excel calculates the average at 10.0 and 10.9 respectively. While those are pretty close to 8.6, which I believe to the no kidding decent rate, I am surprised by the discrepancy. Anyone have any theories?


1656984929726.png
 
Since the Quantum only has a Baro chip it only directly measures Altitude.

The Velocity it records is the derivative of the Altitude. And since there is some noise in the Altitude data V can vary a bit.
Even using the Filtered Altitude and then filtering the Velocity there is variations.

I would graph in Excel ONLY the main Chute decent. This should be close to a straight diagonal line.
Then Apply a Linear curve fit (select Curve, right click and 'add Trendline...'). Select displaying the equation. The Slope ( y = slope * x + intercept) in the equation will be the decent rate.

Do this also for the Filtered Velocity, only the Main chute decent portion.
This will average out any variation over the length of the main chute decent.
What numbers does this give you?
Do they match closer?
 
Walt,

Actually easier to do that using Minitab which I use for my day job. I only used the values in the chute part, the graph is just for visual representation... I used the actual values in the file for the calculations. I get the difference in what part of the curve you pick, etc. What I can't understand is if you use the same subset of data, and calculate descent rate by total feet / total time, that doesn't yield the same value for the same data set when you take the average of the velocity data. They are close (within 1 ft/sec), but I would think that they should be closer all things considered. I guess the filtering algorithm as you suggest introduces some variation.

Thanks for your thoughts.

Pete
 
Does your method with Minitab do a linear curve fit of all the data points?
Or just use the 'end points". A linear fit would 'average' the variations.

Ok, with only Baro data I would consider 8.6f/s to be close enough to 10f/s.

I have tried other calculations based on Quantum and Quark data and found similar mismatching and/or lots of noise.

Also, The Quantum derived values never seem real good. Example is the reported Max Acceleration is rarely close to something reasonable.

You could try starting with the Raw Altitude data then do the math to smooth and calculate Velocity.

I did see the graphs you posted in the other thread. The Alt and Vel curves are pretty smooth compared to ones I have from smaller rockets.
I have a 4" Goblin with a Mobius video camera onboard and the Goblin is very smooth both in accent and decent.
 
Walt,

Minitab uses sum of least squares through all the points to get the smallest set of residuals for the line slope, complete with ANOVA table.

I appreciate your comments, thanks!

Pete
 
Back
Top