Jolly Login Chute Release -- physics question.

The Rocketry Forum

Help Support The Rocketry Forum:

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

Bat-mite

Rocketeer in MD
Joined
Dec 5, 2013
Messages
11,959
Reaction score
2,797
Location
Maryland
Figure this was better as a separate thread than sticking it in John B's product announcement thread.

I had an experience recently where I was helping a fellow rocketeer prep a very large L3 rocket. He asked me to turn on his altimeters, which I did, but then he thought it would be fun to show me the inside of the AV bay, which was accessed with a door in the airframe.

About a second after he removed the door, the deployment charge fired and ejected the 50-pound nose cone.

We realized after the fact that when he opened the door, the air rushing into the bay caused a change in barometric pressure that fooled the altimeter into firing.

That's the background to this question. In a normal AV bay, the altimeter remains inside the bay, and its only access to the atmosphere is through the static ports. But with the Jolly Logic Chute Release product, the altimeter will start out inside the booster, sampling with a static port, but when the wrapped chute is deployed at apogee via motor ejection, the altimeter comes out with the chute.

Wouldn't this cause the same thing to happen as in the example above? The altimeter suddenly goes from being inside the bay to fully out in the air, subject to the wind.

How does this pressure change not cause the altimeter to deploy early?

Thanks in advance to John Beans, or anyone else who can answer.
 
...
We realized after the fact that when he opened the door, the air rushing into the bay caused a change in barometric pressure that fooled the altimeter into firing.
...

I think you have that wrong. The pressure outside and inside the bay should have been the same. But when he opened the door the air pressure inside the bay decreased for a split second and then normalized again. IMO that is what caused the false reading and triggered the main chute.
 
Jolly Logic has lots of experience filtering out ejection transients since the release of the original AltimeterOne. All three models of the JL Altimeters do this. Also, the Chute release looks for a particular altitude on the way back down....

Hopefully John will chime in with specifics. I could post AltimeterThree data showing ejection event spikes pretty well smoothed if you're interested in that.

In the event that you described, I'd wonder if there might have been a hand inadvertently covering the static ports while the access door was being opened.....
 
In the event that you described, I'd wonder if there might have been a hand inadvertently covering the static ports while the access door was being opened.....

That certainly seems like a possibility. Or, perhaps opening the door cause a backdraft resulting in "the air pressure inside the bay decreased for a split second and then normalized again" of the previous post.

In any case, it seems pretty sure that just exposing the altimeter to the outside air was not the culprit.
 
This is one reason why I like my Raven 3 altimeters that have both accelerometers and barometric altimeters. They won't fire any pyro charges until after they detect launch (at least 3 g's).
 
What altimeter(s) fired?

Did by chance you touched the cap when turning them on?
This can cause them to fire the charge(s)..

JD
 
You're hitting on what makes it hard to write algorithms for sensors, and that applies to altimeters as well as things like a Nike Sports Band, for instance.

It's one thing to write an algorithm for the "ideal" case, it's quite another to be able to handle the rather random and sometimes rare conditions that crop up (either in proper OR improper use of the product). It's what we spend most of time doing once the basic mechanics of the product are worked out.

BEC touches on the gist of this: to make a product robust, you have to test and test and test and learn to handle a wide variety of environmental and human events. It's not to difficult to wire a pressure sensor to a microcontroller and detect altitude. But it's pretty hard to make a reliable altimeter that "just works" all the time, even when odd random stuff occurs.

The filters on Chute Release get updated just about each time we test. One of our testers noticed a false release (like you described) during loading and checking (pressure and suction as the nose cone was taken off and on). To a pressure sensor, it takes a ridiculously small pressure change to simulate 100+ feet of altitude gain. So if you want a little academic exercise, try to think through the logic that would be able to separate a sudden drop in pressure, then an increase versus a flight (which looks like a pressure drop, followed by an increase). If you design it right, you always ignore loading, but always react properly to a flight.

This is analogous to how AltimeterTwo (which uses an accelerometer to sense the start of flight) is able to ignore handling and dropping, but can still capture flights with so few errors.

Writing algorithms is a delicate balancing act, not a "more is better" exercise. Sometimes I'll write an algorithm that is "too clever," and actually reduces reliability. For instance, I'm getting ready to "dumb down" the Adaptive Ground Set filter for Chute Release. The purpose of the filter was to adapt the ground level to weather changes as the rocket sat on the pad for long periods. In some cases, it was taking as "ground level" a higher pressure after being loaded inside a poorly vented rocket. Having a "too high" ground pressure is very bad for a chute release, because a "too high" ground pressure implies a "too low" ground level, and the rocket will crash into the ground before it thinks it is getting close to where it thinks the ground should be. In this case, it's probably better to just lock in the ground level prior to loading and have a rare 25 foot error on some flights in rapidly changing weather than it is to introduce the possibility that the chute won't open. I'm not saying we won't eventually get back to an adaptive filter that works more robustly, but until it's more reliable: it's not worth it.
 
What altimeter(s) fired?

Did by chance you touched the cap when turning them on?
This can cause them to fire the charge(s)..

JD

He had two RRC3s. I had actually only turned on one of them when he opened the door, and nothing was touched. But like others have said--the sudden suction from opening the door, followed by a rush of air looked like ascent and descent to the altimeter. So the one that was turned on fired a split-second after the door was opened.

A 50-lb NC shot about twenty feet into the air. Did we run? Did we yell, "Look out?" Did we put our hands over our heads? No. We stood there with our mouths hanging open, looking up and watching it. Thank God the shock cord got hung up on the airframe as it came down, or someone would have gotten hurt for sure. :facepalm:
 
He had two RRC3s. I had actually only turned on one of them when he opened the door, and nothing was touched. But like others have said--the sudden suction from opening the door, followed by a rush of air looked like ascent and descent to the altimeter. So the one that was turned on fired a split-second after the door was opened.

A 50-lb NC shot about twenty feet into the air. Did we run? Did we yell, "Look out?" Did we put our hands over our heads? No. We stood there with our mouths hanging open, looking up and watching it. Thank God the shock cord got hung up on the airframe as it came down, or someone would have gotten hurt for sure. :facepalm:

Hey... I seem to remember a similar occurrence at LDRS.... :p
 
Here's a flight from about an hour ago that demonstrates ejection shock filtering.

It was a flight on an F motor to 868 feet.

Here's the tale of the tape:

Flight15.png


Here you can clearly see the ejection spike (AltimeterThree was in the fuselage with Chute Release and the parachute. Notice that both products ignore the pressure spike (otherwise AltimeterThree would have gotten the apogee wrong).

The ejection was a little early, but both products (and the parachute, which was folded up nicely and thus protected from damage) did well.
 
I like what I see. Between single and double deploy. I call it 1.5x deploy :)
 
Back
Top