Discussion in 'Rocketry Electronics and Software' started by cvanc, Oct 8, 2012.
Tom, in the file you posted which values are the altitude?
The default if you open the file shows only acceleration and barometric altitude. You need to select other parameters separately. Peak altitude was about 2100 feet. (Hold CTRL and select velocity (accel ft/sec) for the velocity curve without clearing the others.)
Thank you. I had hoped the file included altitude values so that I can import the data in to my software.
Look in the lower left corner. 3 tabs - graph, summary, raw. Might be what you want?
I know the information is there, but you should contact Adrian for the exact file format.
Here is the Raven3 FIPa file from my flight today. The rocket is a modified Wildman Jart, with an altimeter in the (fully seperable) nose-cone shoulder (or coupler tube). The strange thing about this flight is that it appears that both the apogee and main channel fired at apogee. The nosecone is hollow and contains the ejection charge, chute, etc. There is burnt black powder inside the nosecone, and the shock cord affixed to the metal nose tip is blackened, so I know the charge fired. If it had been some other sort of separation, the nosecone deployment charge would have fired OUTSIDE the nose cone, and none of this evidence would have been left behind.
The trace shows some curious things:
1) The Barometric pressure increases sharply when the apogee (aft) charge is fired. This suggests to me that I have "leaks" in the double-bulkhead at the aft end of the altimeter bay. It does not look like this negative-going altitude spike went low enough to trigger the firing of the main charge. So what did? The main charge voltage dipped, but I thought this was a side-effect of #2 below. Given that the charge actually *fired* (as evidenced by the soot and discoloration in the nosecone), then this must actually be due to the channel firing.
2) Battery Voltage sagged pretty badly when the apogee charge was fired. The leads may be too long on this battery. I need to run some ground tests to see what caused this.
So, anyone know for certain what caused the main charge to fire at apogee?
View attachment 2012-10-14-ESL-173.FIPa
P.S., the more I look at this the more mystified I become. The main channel *clearly* fires much later in the trace. The voltage on that channel only appears to dip at apogee - and I am all but convinced this is due to the battery voltage sag (they look almost identical). The mysterious part is that the chute is bright orange, and was clearly deployed at apogee. I thought it was a mechanical separation due to shock (sheared the nosecone nylon shear pins when the aft shock cord reached full extension). The only problem with this is that the main ejection charge had relatively short leads, and the charge canister was not affixed inside the nosecone in any way. It is impossible for it to have fired later and still have been inside the nosecone. Yet there is all that burnt evidence in the nosecone (and smell of spent black powder).
Could you post some pictures of your setup? I think that might help. I agree the battery sag probably made the "main" data dip - after all how can any monitored parameter read higher than the instantaneous power supply voltage? But you say there is evidence the main actually went off at that moment. I don't get that. It looks like it really went off at 256 seconds and 700 feet.
If you scale the battery, apogee and main voltages so that they both go 0-5V, you can see that the main continuity voltage just dipped along with the battery voltage. Also, the current was reading over 13 Amps during the whole 1-second firing. So you had a short during the apogee firing, but the main output was not switched on until later when AGL < AGL1. What kind of battery and igniter did you use at apogee? The accel data looks like the main inflated at apogee.
You can right-click in the parameter window to copy or save the data for import somewhere else.
Holy cow Adrian, you're right. There is a huge spike in the current draw at apogee. I didn't look at that before. Can't be very healthy...
Thanks. I installed your software and exported the altitude values. I was then able to import the cleaned up data into my software. Worked like a charm.
View attachment M1297.FIPa
I want to make this one a mystery but ill leave some hints.
Question is what happened and why, anyone have guesses? This flight was in the spring and I know what I think happened and will share soon.
The "PS" in my post was an edit - after I had done exactly what you suggest here. I could see that the battery voltage sagged heavily, and that the main channel voltage was just following this envelope.
It didn't occur to me, however, to check the current draw. Wow, 14A! No wonder the voltage sagged. I am using a 600mah battery with a connector that mates to the power perch. I use the mounting screws of the perch to power my BRB beacon. The ejection canisters I am using are the mini ones from Pratt hobbies, which I believe use a crushed grain-of-wheat lamp as the igniter. The 14A draw did not last for the entire 1 second, though. If you expand the timescale it lasted only about 0.4 seconds. Still, half a second of 14A is ridonculous. I think you had recommended elsewhere an e-match that does not short when fired? Could you reiterate that here? Also, I think you recommended a small series resistor in case one anticipates such a short. I may need to incorporate this if I don't re-work these canisters with some better e-match type of initiator.
Yes, we observed that the main chute deployed at apogee along with the streamer. The thing that was vexing me before (how did the forward charge fire at apogee) has been resolved.
I HAD THE CHARGES REVERSED! :cyclops: I can't believe I did that. This is the fourth avbay I have built using a Raven, and I've never made such a basic mistake on any previous one.
The evidence shows that the apogee charge fired the nosecone off its shoulder by design - although unintentional The apogee charge was *supposed* to fire the fincan off the aft end of the nosecone shoulder. The main charge did not fire inside the bay. There is no soot or debris in the fincan bay, so now aside from correcting the channel reversal, the only mystery I need to sort is that somehow there was a mechanical separation of the aft bay. I also need to fix the pressure leak, that should be no big deal.
Thanks all (especially Adrian) for your insight.
Now *there's* an interesting file. I have absolutely no idea what happened. But I'd guess more than one thing went wonky on you?
The file says you impacted the ground at two thousand eight hundred miles an hour. That's gotta be wrong - you'd never get the Raven back to extract data from. (Right?)
Since there wasn't anything visible in the lateral axis before the disintegration, that indicates that it wasn't the usual shred cause of coupler failure, instability, or fins. That plus the consistent positive-G impulse during the burn suggests to me that it was a CATO.
You should make sure that the main FET wasn't damaged from that much current. It's in the range where it's possible that the FET overheated. If it was damaged, it will leak current when it's supposed to be off. You can put a multimeter in current mode to check how much leakage there is when it's off. I recommend not using a larger Lipo battery than the 130mAhr ones provided, for this reason. If you do still want to use a larger battery, you should use the 3rd or 4th outputs which have higher-capacity FETs. The Raven3's 3rd and 4th outputs are o.k. over 20 Amps, and possibly more.
Here is my L2 certification flight from LDRS
Looks good Les, but could you post the actual data file instead of a screen grab? Much more to look over that way...
Adrian, are you saying some shreds show an undamped, increasing oscillation in the lateral axis data, up to the moment of airframe failure? Very interesting.
Yes. People sometimes ask why I bother with recording the lateral accel axis when it's not used in any deployment logic, and it's for cases like this.
The baro "sees" the ejection because when the payload bay is separated from the booster the movement causes an atmospheric pressure decrease then increase as it rebounds from the shock cord's elasticity. If you have a baro-only recording unit, this is a good way to tell when your ejection occurred if it's being done by the motor and not the electronics; you'll see multiple increasing-decreasing velocity spikes dampening over a second or so.
It is stuff like this that is the reason I started this discussion. I really want to know how to dissect these data files. There's so much useful info in there. It is a chance to discover what our rockets *actually did* as opposed to what we think they did. It is a direct path to better practices and better designs.
Thanks Adrian and everyone. Keep 'em coming!
Rocket- Formula 54 with 38mm engine tube. Converted Nose cone coupler to av bay. 1st flight on Estes F50-4T, 290ft, 2nd flight on Cesaroni H123 Skidmark w/motor ejection and apogee and 700' charges [L1 cert] 2300' =/-, 3rd fl Cesaroni H123 Skidmark with apogee and 700' charges 2300' =/-, 4th fl Cesaroni H123 Skidmark with apogee and 700' charges 1900' =/- [bad spin during flight]. 3rd and 4th flights had forward and aft av bay plates sealed against charge gasses.
NAR 6306 L1
Tri 14074 L1
View attachment Number 1.FIPa
View attachment L1 Cert.FIPa
View attachment 10-14-12A.FIPa
View attachment 10-14-12B.FIPa
Replace "700" with "main" in the above post.
This flight was 'interesting'. 2 J500, 3 H178, 2 H73 cluster.
View attachment Umax09.FIPa
I didn't get much video from the ground but I had a keychain cam aboard[YOUTUBE]Jxgc-HIaAkM[/YOUTUBE]
Sorry I haven't been on this for a couple days; been busy. (and hey any and all of you are most welcome to comment, this isn't supposed to be my personal thing LOL)
On number one it looks like both the apogee and main channels went off at the same time, was this intentional or inadvertent? And yeah the baro trace has a great big whack in it right at that moment; about 5.4 seconds in. Looks like AV bay leakage to me. Something else happens before that however... at about 3.4 seconds there is activity on baro & accel traces. Any idea what happened there? Motor ejection charge, maybe?
"Level 1 flight" looks like the Raven was along for the ride (no BP charges) and the motor ejection did all the work?
3rd & 4th flights have that classic dual deploy look, but the 4th especially has the roughness on the baro data that I mentioned earlier in this thread. Interesting you note the flight had a lot of spin, was that on the way up or down? If it was on the way down maybe this data roughness relates to a spinning airframe?
And it'd be great if you share what you did to further seal your bay. I've had this problem and I'm sure others would also like to know the tricks that work. Thanks.
Man, I don't even know where to begin on this one. So many accel spikes! Were all the motors supposed to light on the ground? The video (thanks) suggests 2 tiers of ignitions but the accel data seems to show many staggered ignitions (I think?). Was the rocket OK?
View attachment spare parts j316.FIPa
Nice flight. Awfully big motor for an awfully small rocket! You actually hit 50 G's there for a bit. And you were over 30 for almost two full seconds. Wicked.
Looks to my eye that everything worked normally? What e-matches were you using, especially at apogee? The volts apogee data is ratty after the charge lit, which shows partial conductivity still exists after firing (which is harmless but makes the data look cluttered).
I just noticed the temperature data is noteworthy: It varies over a 20 degree range during the course of the flight (which is a lot compared to any other file in this thread). I wonder why it swings from 80 to 100 degrees? Most of the flights here vary maybe 3 or 4 degrees.
Separate names with a comma.