Jones Boys test fire

Added by JC Jones almost 4 years ago

We did the first test fire with a known engine using the Shepard v1.1 software and found an interesting anomaly.The time stamps, when checked against the video seem to be off. Total time seems to be twice as long as the event. Any ideas of why this is happening?

Video here: https://www.youtube.com/watch?v=Zf-JY_Vloxc
Data file attached.

Thanks,

John & Christopher


Replies (18)

RE: Jones Boys test fire - Added by J. Simmons almost 4 years ago

Hey guys,

I am probably going to have to defer to Jeremy and Chris on this one. But, let me at least take a look at things. Just to check, is the attached data file the raw file from the DAQ software, or is it edited in any way? I ask because there is a jump in time between lines 50 and 51. I also saw a couple of negative time stamps, which seems odd to me.

Out of curiosity, do you have a copy of the expected thrust profile for this motor? I am having a little bit of difficulty sorting out where the motor actually fired in the data. I am guessing that could be related to the vertical orientation of your test stand (there appears to be quite a bit of data with about 1-3 N of force on the sensor until about line 225 or so, where the force really picks up and peaks at around 45 N near line 443).

In reviewing the video, I see now that the ignition sequence is quite noisy, which may also account much of the jitter in the early part of the data. Assuming the real burn starts at about line 225 (which has a timestamp of 3.746 s) and that the burn continues until the sensor values taper off to around 1 N at line 3278 or so (timestamp of 6.6 s), then the actual burn is about 2.9 s in length. As I watch the video, it looks like the actual burn starts at about time index 0:32 and continues until about time index 0:35 s, giving an approximate burn time of 3 s. That seems pretty close to the 2.9 s time in the data. And, if I use a timer with the video, I get a burn time of about 2.8 s.

So, it looks to me that the burn time in the data is correct (other than the odd timestamps). Can you walk me through your analysis of the data so I can see the timing difference you are seeing?

Thanks.

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

Sorry for the delayed reply. Pesky day job...

Your data shows a -203.159 for the timestamp every once in awhile. There's also that jump between lines 50 and 51 (that J mentioned) that concerns me. The jump may be explained below, but the negative timestamp concerns me a little bit.

John & Christopher - In addition to what J has asked for, can you define the "event"? Is that the entire test from recording enable to disable?

The way the Processing sketch is written, as long as the "Enable Recording" button is clicked and the thrust is greater than 0, data will be recorded. With v1.0 of the test stand structure we have run into situations where the slide will "stick" and cause the load cell to stay above 0. That means that data will be recorded until you hit the button again to disable recording. Did you happen to notice if the thrust value was sitting above 0 when you went back to the computer? Also, if at any point in the test firing your load reads 0 for any reason, the recording will stop. If the value goes back above 0 again, the recording will start again. That could explain the jump in the timestamp between 50 and 51 if you had a mechanical or firing anomaly, or if your load cell sensitivity is low enough that low thrusts read as 0.

I don't have a set of Shepard hardware with me right now, but I should be able to rig a test circuit up tomorrow morning. Then I can do some tests to see if I can reproduce these issues.

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

I have been able to reproduce the negative time issue here. I have a theory on why it's happening, and I'll work on troubleshooting today as I get time.

I also just noticed that the Jones Boys' data and mine both show the temperature maxed out at 382.23 degrees C. This is a bug too, since when the sensor is disabled it should go to -273.13 degrees C.

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

At least on my test setup, the thrust value will hop above 0.01 (the trigger thrust) every once in awhile. When the Processing UI sees a value above 0.01 and recording is enabled, it will start the counter. If the thrust hops above 0.01 before the test firing starts, it could cause jumps in the timestamp and the software timer getting out of sync with the test fire's time. This trigger value could be increased to lessen the risk of this happening, but it would be a balancing act between eliminating false triggers and losing a little bit of data from the front of your thrust curve.

This brings a couple of things to mind.

  1. In theory, it should be possible to use a USB-to-serial adapter on a laptop and run that straight into the TX/RX pins on the Arduino. If the current Shepard DAQ could use serial that way, it would allow the user to move the laptop further away from the test stand than you can with USB. I don't think this idea will make it into Shepard proper because it complicates the operation, but it might be a useful adaptation for Jones Boys' Rocketry to try out.
  2. With a change to the Processing UI, it should be possible to eliminate the thrust trigger and just save data from recording enable to recording disable. Again, probably not a feature that will be mainline for Shepard.

Both of these features combined would allow the John and Christopher to have the laptop with them outside of the safety zone (assuming it's less than a 40 or 50 foot radius). They could then enable recording to start collecting data and disable it again to stop recording. That way there's no question about whether there's a false trigger or not. The user controls the data collection. That would also eliminate the need for a delayed stop like CCSSC has requested.

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

I've committed a Processing sketch change to the repository that will discard a sample if the timestamp is invalid.

John and Christopher - Can you check the current revision (58) out of the repository and try the Processing sketch again? I would suggest checking it out into a different directory and running it from there. If you try to do an 'svn up' on your current copy of the software with the changes you've made, it's probably not going to work right.

I would look at the following line at the top of the Processing sketch to fix the timestamp jump/sync issue.

float triggerThrust = 0.01; //The thrust level that will tell the software to start recording data.

In my tests here, I saw that the raw value for the load cell will float between 0 and 2 with the load cell sitting on the desk. With my calibration that's not enough to bring the value above 0.01, but on your setup it might. I would play around with the trigger value to eliminate false triggers. The procedure I would follow is to set your load cell up in its firing configuration, with a motor in the tube (no igniter of course). Enable recording and let it sit for awhile there at steady state. If you end up with data in the csv file (and probably on the graph too), you need to increase the trigger thrust. Once that's done, I would add just a little safety margin to that trigger value since the actual test firing will be outdoors.

Please let me know if that helps.

RE: Jones Boys test fire - Added by JC Jones almost 4 years ago

Hi Guys,

Sorry for the delayed response… we've been distracted by building motors. : ) To answer J.'s questions, that is the un-edited data file. The expected thrust curve for that motor is here: http://www.thrustcurve.org/simfilesearch.jsp?id=694 You're right that the recording probably stopped momentarily after the chuffs that happened as the motor ignited, but the time stamp shouldn't have jumped that much. However, as you said, once adjusted for that, it seems pretty close. I was using line 131 as the start for the full fire. With the adjusted time it gives me a duration of about 3.2 sec while measuring the time in the video of what I think is the recored event (by event I mean the actual firing event), it comes out to about 2.9 secs. Pretty close.

Don't know where that negative time is coming from.

John & Christopher

RE: Jones Boys test fire - Added by JC Jones almost 4 years ago

Thanks Jeremy,

We'll give that a try!

John

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

Hi John and Christopher.

Any updates?

RE: Jones Boys test fire - Added by JC Jones almost 4 years ago

Hi Jeremy,

No… We were hoping to do a test fire today but ran into some issues with the motor prep.

John & Christopher

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

Hopefully the issues were nothing major. We're excited to hear about how your test firing goes when you do get to do it.

Out of curiosity, how did you hear about Mach 30?

RE: Jones Boys test fire - Added by JC Jones almost 4 years ago

One of the component of the motor casing was manufactured incorrectly and we had to wait for a replacement. Not major but a significant delay. We were finally able to get the test firing done over the past few days. Still crunching the numbers but I think the data is good.

We found you guys through the youtube video while searching for test stand info.

John & Christopher

Untitled_4.jpg (791.4 kB)

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

Sorry to hear that you had supplier issues, but I'm glad you're up and running now.

That's great that the data looks good. If you find any other issues let us know and we'll see what we can do.

What kind of analysis are you doing?

RE: Jones Boys test fire - Added by JC Jones almost 4 years ago

You can see the analysis in the attached. He had to turn it in for grading today. Remember, this is a high school freshman project. ; ) We cited opendesignengine.net in the picture captions but will have a more proper citation in the project notebook.

Thanks again for all you help!!!

John & Christopher

Full-Layout-sm.jpg (1017.5 kB)

RE: Jones Boys test fire - Added by Jeremy Wright almost 4 years ago

I'm very excited to see that and I think the layout looks very well done. I hope that he gets a great grade.

Most of our hardware team is wrapped up in another project right now, but when we get to focus more effort on Shepard again I'll make sure I get the team together to review his layout.

Thanks!

RE: Jones Boys test fire - Added by Jeremy Wright over 3 years ago

John & Christopher,

I wanted to touch base with you to see how Christopher did on his project. I also wanted to gauge your interest in maybe doing a live demo of your test stand during our Yuri's Night event this year.

http://yurisnight.net

So, any interest?

RE: Jones Boys test fire - Added by JC Jones over 3 years ago

Hi Jeremy,

Christopher's project turned out great. He got an A for his class but didn't get any awards at the science fair. He did however get invited to participate in the regional competition which is this weekend so we'll see how that goes.

As for the the Yuri's Night event, he thought it would be cool to do but he's already taken the the electronics apart and started another project. (Something very cool: involves UAVs with FPV and rocket payloads!)

Sorry, maybe we can just join to watch,

John and Christopher

RE: Jones Boys test fire - Added by Jeremy Wright over 3 years ago

I'm glad to hear that Christopher got an A, and it's exciting that he's been invited to participate in the regional competition.

It would be great if you guys wanted to watch our Yuri's Night event, but if we have enough space in the Hangout would you be interested in attending directly as guests?

Also, can we get your permission to play your test firing video (linked in the top of this discussion) during the event?

RE: Jones Boys test fire - Added by JC Jones over 3 years ago

Sorry I haven't gotten back to you. It's been a busy couple of weeks. I've been working with a TARC Team https://plus.google.com/photos/110261994983202529442/albums/5991819839710277665?authkey=CKuqwKDBp-j0rQE. Christopher competed in the regional ISEF Science Fair and took a First Place in physics for his group (9th grade)! He got a "sponsor's prize" of second place from The American Institute of Aeronautics and Astronautics which included a $100 cash award and a 3 day workshop at Goddard. He also got an honorable mention from the USAF Office of Scientific Research. We're waiting to hear if the county is going to select him as one of the kids they take to State. Thanks again to you guys for all your help in making this possible.

We haven't had a chance to even think about Yuri's Night. He did put the test stand back together to display at the fair but we haven't tested it. I posted a video of the actual tests here: https://vimeo.com/89640260 You are certainly welcome to play that one.

John

(1-18/18)