Alt3 vs Thrustcurve

[dhkaiser]

Recently flew my SD Madcow DX3 2.6" TWFG on a Aerotech I140 to 3473ft. Thrustcurve returned a max acceleration of 10.4 G's. Alt3 returned a max of 14.36 G's. This is my first time comparing the two. Should I expect this kind of variation between the two. Or is this an accepted +/- in the published performance of the motors?

 

[Buckeye]

Thrustcurve is 1D. Alt3 measures 3D. Is the 14.3G the total acceleration magnitude or just the component along the rocket axis? How vertical was the flight? Mass entered correctly in Thrustcurve?

Since max and min accelerations usually occur during the motor burn, then the usual caveats about motors not performing to their data sheets apply.

 

[JohnCoker]

The simulator in ThrustCurve.org is very simple; it's main purpose is to give you a feeling for what a motor will do for your rocket and whether it will provide enough thrust for a safe lift-off. That said, you will see a difference between simulation and actual flight even with more complete simulators because they don't have all the relevant data. For example, various factors related to the rocket's shape and surface are wrapped into the C_D (coefficient of drag). One thing you can do is adjust the C_D in simulations to make the simulated altitude match the measured one, then future simulations will be more accurate.

For more info on how simulators work see my Flight Physics article.

 

[Buckeye]

One thing you can do is adjust the C_D in simulations to make the simulated altitude match the measured one, then future simulations will be more accurate..

Not always. This may help the one flight in question, but may not extrapolate into the future. You are using just one tuning parameter (Cd) to compensate for everything that is wrong in the flight/simulation, like motor thrust, weather, and mass, in addition to aero drag.

Discrepancies between motor thrust data and real motor burn is by far the major factor influencing sim and flight correlation. Thrustcurve motor guide, being crowd-sourced, uses the average of all entries for a given motor. This gives a good general feel for performance and safety, but makes it hard to dial in an exact correlation based on just one observation.

Anyway, for the question posted here, since we are looking at peak accel, which occurs within fractions of a second of ignition, my guess is that simulated vs. actual thrust curve is the culprit. There really is not enough time for errors in Cd to accumulate.

 

[John Beans]

Two things:
1. "Max" values are random and noisy, because motor to motor you can see hard-to-predict spikes that just naturally occur for tiny moments during burn. So the very "peak" acceleration measures are rarer events than you would care about really. The total area under the curve or averages during burn are more meaningful, and will more accurately translate to speed and altitude.

2. Acceleration due to thrust is the AltimeterThree total acceleration reading - 1G.
The accelerometer will read 1G when it's sitting on the pad before launch. The reading minus 1G is due to engine thrust.

 

[Buckeye]

Two things:
1. "Max" values are random and noisy, because motor to motor you can see hard-to-predict spikes that just naturally occur for tiny moments during burn. So the very "peak" acceleration measures are rarer events than you would care about really. The total area under the curve or averages during burn are more meaningful, and will more accurately translate to speed and altitude.

2. Acceleration due to thrust is the AltimeterThree total acceleration reading - 1G.
The accelerometer will read 1G when it's sitting on the pad before launch. The reading minus 1G is due to engine thrust.

Yes, measurements can be misleading, too, not just the simulations!

Like the old engineering adage says, "Everybody trusts the test result, except the guy that ran it."

 

[cerving]

Motor tests are done horizontally on a static stand to eliminate any other factors. When you try to get the same results in flight, you have drag, wind, roll, CP/CG movement, etc. that might affect the data from your accelerometer. You also need to sample quickly, but filter the samples to get rid of any spikey readings. I'd trust the test stand data over the flight data for motor characterization, which shows what your ROCKET is doing, not necessarily what your MOTOR is doing.

 

[Buckeye]

Motor tests are done horizontally on a static stand to eliminate any other factors. When you try to get the same results in flight, you have drag, wind, roll, CP/CG movement, etc. that might affect the data from your accelerometer. You also need to sample quickly, but filter the samples to get rid of any spikey readings. I'd trust the test stand data over the flight data for motor characterization, which shows what your ROCKET is doing, not necessarily what your MOTOR is doing.

Right, of course. Same thing for drag. A wind tunnel test or CFD simulation is the best way to get Cd, and not from flight back-calculating.

I think the OP was simply looking at peak rocket accel only, one from simulation and one from acclerometer. Mr. Beans suggested to not entirely trust the peak value from the Alt3.

 

[cerving]

Right, of course. Same thing for drag. A wind tunnel test or CFD simulation is the best way to get Cd, and not from flight back-calculating.

I think the OP was simply looking at peak rocket accel only, one from simulation and one from acclerometer. Mr. Beans suggested to not entirely trust the peak value from the Alt3.

As long as it's not for motor characterization, the most useful number you're going to get is average acceleration from launch to burnout (which you can get by setting a timer, or get from maximum velocity). That eliminates most of the noise that you might get. That's what we do with the Eggtimer, Quantum, & TRS; most of the time it's a pretty reasonable number once you download the flight profile and look at the graph.