Osprey 75 2-Stage post-build thread

[woferry]

To start at the end, I flew my first 2-stage rocket last Sunday at TCC's October Skies and it was a success. J400SS to I212SS, booster only went to ~1000' and the sustainer to ~3000', but I wanted a low & slow flight for my first attempt at staging. As it was I lost visual on the sustainer shortly after burn-out, only finding it again when I saw the tracker said it was ~150' from the ground and scanning the horizon to see it just before it landed. But I did manage to get the separation charge on video as well as sustainer ignition (though it was just above the frame when the motor actually lit). I have the video online, the Google albums I post to don't seem to allow direct-linking, so I'll just point to my signature and the October Skies 2017 album, the last 9 images & video are all related to the 2-stage flight.

I figured I'd document the main details of the build even though it's after the fact. I'm also working on a Double Shot using the same basic build techniques, since the Osprey 75 booster was largely done I finished the Osprey first, the Double Shot sustainer is finished but the booster has a few steps left to finish (just rail buttons, aft CR and motor retainer). With any luck the Double Shot will fly next month.

 

[woferry]

I originally bought the Osprey 75 kit for Black Friday 2015. I was laying fillets to finish the airframe in my RV during LDRS 2016, but wasn't happy with my electronics setup so I didn't end up flying it there. After sorting out the electronics my first flight was at TCC's August 2016 launch, on an I285R and everything was successful. Second flight was at XPRS 2016 on an I223SK, but I had a complete brain-fart and failed to properly arm the Eggtimer TRS (it was sitting at the dro/main screen, I didn't do the second long-press to actually arm it), so the rocket came in ballistic. After spending over an hour digging it out of the playa with a hand shovel the nose cone had split, and the base of the nose cone and the avionics bay had been crushed into the upper ~7" of the airframe, had to hacksaw the airframe to get the pieces separated. The rest of the airframe/fincan was intact, so I figured right away that I'd be rebuilding the rocket and flying it again.

By the time I got around to ordering the replacement pieces from Madcow, they effectively weren't making this kit anymore, apparently they were working to standardize some of their parts and I guess the nose cone was in limbo during this time. Of course what I needed to re-build was the nose cone, the 7" coupler/av-bay, 1" vent band, plus a second coupler and 8" of airframe to replace what I had to cut off. After talking to Madcow about ordering the items that weren't listed on their site (NC and odd coupler length), it became clear that ordering the pieces was going to cost as much as ordering a whole new kit. This got me thinking that I could turn it into a 2-stage rocket, building the new kit as a sustainer. So instead I ordered the full kit, a length of 3" orange tube (36", no cuts), a 6" coupler (for the booster extension) and a 9" coupler (for the ISC).

So all the booster needed was to be restored to its original length. That took the 6" coupler and an 8" cut from the tube I bought (and some epoxy, obviously). For the 2-stage configuration I decided to go with a drogue as well as a Chute Release'd main.

 

[woferry]

The sustainer was obviously where most of the work needed to be done. First was notching the fins so that a coupler could go 3" into the end of the airframe. The Osprey 75 has very long fins, so there was still plenty of length for the attach. I used a scroll saw to cut the notches, with a guide clamped to the fins to ensure I followed the desired line.

I'm not keen on modifying motors for top-ignition or the like, especially since I fly at launches that are NAR or don't do TRA Research. I planned to use an RRC3 in the sustainer av-bay to do the sustainer motor ignition, so I had to run a lead down between the 54mm motor mount and 3" airframe for the ignitor. I opted to go with 1/8" brass tubing from my local hardware store, whatever length I bought (12" I think) was basically perfect for the distance between my two centering rings. For the wire to run down this conduit I went with a 3' extension for a 2.1mm barrel jack, as commonly used with security cameras. This was just the right diameter to fit in the conduit, and as I was looking for a good break-away connector for the sustainer av-bay connection I concluded that this would be a good connector designed to handle a few amps while having decent retention force (not likely to come out under acceleration) but non-locking so that it could unplug when the sustainer drogue charge fired. I have some electrical tape wraps around the cord to limit how far up/down it can travel, I considered notching the insulation and adding an epoxy blob, but this seems to be working well so far.

For the fins, I decided that I wanted to try internal dams on this build. The original Osprey 75 was my first attempt at injected fillets, and that was pretty much a disaster. I drilled 3 holes per-side along each fin, and injected epoxy but what I used (30 minute Z-poxy) basically was too viscous, it resulted in 3 blobs that went from wall to MMT, but didn't spread out any wider than maybe 1/4". I think at least one of these broke off when the booster lawn-darted, as I can hear something rattling around in the fin-can area now. To do the dams, I decided to use my 3D printer. And I didn't just print the dams, I printed the centering rings as well, including the cut-out for the 1/8" tube, the shock cord loop that would be epoxied to the MMT, and attachment points for the dams. The fins were long enough that I couldn't print a single-piece dam, so I printed a pair of interlocking pieces, 12 pieces total. I used orange PLA on the printed parts to reduce their visibiity through the semi-transparent tube, on the original booster it's easy to see where the CR's are epoxied, especially since I had used Proline 4500.

So I got the CR and dams to be a perfect fit. Of course 3D-printed centering rings wouldn't be nearly as strong as FG. Given the length of epoxy that would be surrounding the fins and connecting the MMT to the airframe, I didn't feel that I needed any structural support through the CR's themselves, they were just for holding the MMT and conduit in the right place during the build. However I was a bit concerned about the exposure to burning BP at the upper CR, so I made a wall with blue-tape and poured a layer of JB Weld on top of the upper CR to offer a bit of a thermal barrier. So here's how the MMT assembly wound up before being slid into the airframe. I had also added some short lengths of kevlar to help protect the shock cord loop where it could otherwise rub against the top edge of the MMT (after seeing someone else do that on this forum, seemed like a good idea).



Once the MMT assembly was in the airframe, it was a pretty simple matter of pouring in a measured amount of epoxy to ensure the entire dam area would be filled, then inserting the fin and letting some epoxy squeze out, then cleaning the excess up for later fillet laying (the airframe was masked like crazy to ensure none of the Proline would leave black marks on other areas of the airframe). Once that was done the exterior fillets were added, using a piece of 3/4" PVC (1.05" OD) for the shaping, setting a radius that was basically 4% of the root length. I put strips of blue tape inside the airframe to seal the portion of the fin slots that extended below the lower CR, to ensure that the epoxy would fill this area but not interfere with the ISC (I had a piece of coupler in place during the fillet pouring to ensure the tape stayed in place). It took some work to get the tape out after the fact (tough to work between the motor retainer and airframe as the gap is small), but I did eventually get it all.

The replacement nose cone turned out to be a bit of a challenge compared to the original. The original Osprey 75 NC ended flat, with a metal tip that attached using a 1/4"-20 screw and washer. In my original design I replaced this screw with a 1/4"-20 forged eye-bolt, and attached my shock cord to this eye-bolt. The replacement nose cone simply had a black FG tip epoxied in place, so there was no provision for shock cord attach (seems like a bad idea for a HED rocket, but I guess they hadn't sorted-out the NC yet when I ordered the not-on-their-website kit). I wound up kludging together a bunch of bits, namely a screw, 2 different washer sizes (I think with a nut in-between) and a coupler nut, then pouring epoxy into the nose tip and dropping this mess in, along with two pairs of holes drilled through the NC and brass rods inserted, the washers end up between the brass rods and the end of the NC, so the epoxy should be holding everything in place along with the rods, which were cut just a bit shorter than the length needed in the NC so that I could 'cap' the ends with Proline. This was obviously harder than using longer lengths and cutting them flat after the epoxy set, but I did manage to pull it off, actually by temporarily lengthening the rods with more rod and CA, then snapping the CA joint as soon as the pins were inserted and carefully doing the final positioning with toothpicks. The coupler nut was so that the eye-bolt could be removed should it be necessary to replace the shock cord, it's thread-locked in place right now.

For the avionics bay, as mentioned previously I'm using an RRC3 for its 3 outputs (airstart on AUX, drogue and main), I have it paired with an RTx for tracking and telemetry downlink. I designed a single 3D printed sled that holds both parts and their batteries (9V for the RRC3 as recommended by MW, 950mAh 2s LiPo for long battery life on the RTx), as well as a pair of Featherweight Magnetic Switches, one for each battery. The sled has the battery wells on one side, the RTx on an adjacent side and the RRC3 on the next side, the 4th side hugs the coupler wall so there's no room for anything there. This sled was designed to go in the 54mm coupler of the Double Shot sustainer, and works in the Osprey 75 coupler by use of an extra 'extender' part, which both moves the magnetic switches closer to the coupler wall and also shifts them up, to deal with the different positions of the vent band holes on the Double Shot (centered along bay length) and Osprey 75 (much closer to the top since the coupler only goes about 1.7" into the nose). In the 54mm bay the batteries are captured purely by the fact that the coupler wall gets in the way, for the 3" bay the extender also covers the battery wells to hold the batteries in place. The sled also has some wire routing features, though I need to improve these for a clean fit into the Double Shot. There is a coupler nut buried inside the sled that holds the sled in place (both vertically and rotationally) as well as connecting the top eye-bolt to the lower eye-bolt. The RTx/RRC3 combo requires the RTx to be powered-on first, then the RRC3 (and then the base station, and they all have to be powered-on within about a minute of each other or they won't sync-up), so I first swipe the magnetic below the vent hole to activate the RTX, then swipe over the vent hole to activate the RRC3, then flip the switch on the base unit. It took a while to get this all sorted out, but mostly because I didn't have the switches set right on the two RTx boards.



For recovery, it turned out my original main from the crashed second flight was fully-intact, so I re-used that along with its chute protector, the folded drogue and its protector had been punctured all the way through by something during the crash so they needed to be replaced.

All told the sustainer turned out to be 1.3 oz heavier than the booster (3 lbs, 8.1 oz vs. 3 lbs, 6.8 oz), that's like-for-like, no recovery except for the epoxied shock cord loops and no electronics, just the extension cable for the sustainer ignitor.

 

[woferry]

While the bulk of the work was on the sustainer, the single most complicated part has to be the ISC. I wanted to fit two pieces of electronics in here, an Eggtimer Quantum to handle the booster/sustainer separation charge as well as the apogee drogue deployment (Chute Release for the main as mentioned earlier), and an Eggfinder TX for locating the booster. I decided to throw a tracker in both after watching several of David Robb's 2-stage flights, where he's usually asking for help keeping a watch on the booster for recovery, I figured I'd use trackers in both. I'm hopeful that some day Cris will have an Eggtimer that combines the TRS's GPS tracking with the Quantum's airstart capabilities (and safer FET enabling so I don't feel the need to use a WiFi or Magnetic Switch on the TRS deployment power), that would make things a bit simpler, but for now its two separate boards.

The real challenge for the ISC is the space, or specifically the length. When I do the Double Shot, I think that ISC will be a bit easier, even though it's narrower its a lot longer and I think I'll be able to get the Quantum and EF TX to fit pretty easily, each alongside its battery (effectively two separate sleds, stacked on top of each other along a central rod that connects the upper bulkhead to the lower one).

But for this bay, I had a 9" coupler to work with (probably could have custom-ordered something longer from Madcow, but I figured I had a way to make this work). 3" of that goes into the sustainer, and then I need space below that for the motor retainer and the nozzle, which for some 54mm motors can extend quite a bit. So that basically leaves me ~4.5" for the bay length. My batteries and the Quantum fit easily in this space, but the Eggfinder's antenna is an issue. I just use the stock brass rods on my Eggfinder TXs, having found them to offer better range than the SMA-style antennas I've tried. Since the rod itself (and provided heat-shrink tubing) is quite narrow, I decided to extend it into the ISC area, it can easily fit between the Aero Pack motor retainer and the coupler wall so long as it starts out close enough to the edge. So my sled holds the EF TX at an angle to help get the antenna up against the coupler wall in a hurry. The sled holds 3 batteries, my go-to 950mAh 2s LiPos for the TX power (runs it for several hours) and the Quantum primary power (again to run the WiFi module for several hours, no need to worry about replacing batteries during a typical day), and a third smaller battery (500mAh 1s LiPo) for the Quantum deployment battery, to take it a bit easier on the deployment FETs (the 950mAh's are 25-50C discharge, so they could easily fry the original TRS's FETs, though Cris' later designs have been much more robust in this regard).

So the ISC is printed in two pieces. The lower piece holds the 3 batteries and the Eggtimer Quantum, as well as having several cable management features (including capturing the balance-charging leads). The upper piece forms the upper bulkhead, and has an integrated charge well for the separation charge, as well as a hole for the EF TX antenna, and holds the EF TX in place, and also provides for 'caps' for the lower piece's battery wells, which are top-loading.



The upper piece is designed to 'float', it does not attach to anything inside the coupler, in case I need to adjust it for different motor nozzles. Instead the two printed pieces are secured to the eye-bolt that the shock cord attaches to (again there's a central captured coupler nut), so all that leaves is the actual 9" coupler and 3" vent band. For these I epoxied a pair of #10-32 coupler nuts to the inside of the coupler, using a 3D-printed part to position the nuts and form the walls to pour the epoxy into (this piece had to be chiseled out, leaving the epoxy & nuts behind), and a second part to properly position the holes in the Madcow FG stepped bulkplate to line up with these nuts. One might notice that the sustainer av-bay bulkplate pictured above has epoxied holes, I drilled the wrong bulkhead the first time. There is a spacer and washer visible at the end of the lower bulkplate eye-bolt, this secures the sled in a fixed position relative to the bulkplate, but the spacer was just one I had handy (the washer is needed because the spacer would fit inside the hole made for the coupler nut), I've already proven that some motors will require a shorter spacer (I have a K535W DMS motor loaded and ready to fly for a different rocket, its nozzle is too long to fit in the space I currently have), so I'll probably print a shorter spacer to allow the sled to be brought much closer to the bulkhead, I do need some space for the bulkhead charge wires but that should be it. On the right angle (and provided I forgo one of the screws for the Quantum) the electronics just manages to fit between the epoxied coupler nuts, as shown in the 3rd picture above.

Ground/flight testing of this has exposed one main issue, the seal between the sustainer coupler area and the bay. It was printed to be a tight fit, and was reasonably air-tight when blowing into the sustainer end of the coupler. But it's tight enough that once the BP fouls the interior walls of the coupler it's very difficult to get the assembly out. I've had to thoroughly wipe-down the inside walls first before I could remove the sled. So my plan is to reduce the OD of the lid somewhat, and embed a ridge for an O-ring, and attempt to make the seal using a greased O-ring instead. This was actually my original thought, but I couldn't find a source for O-rings of the right size, so I gave up on the idea. But I have on order the bits necessary to make my own custom O-rings, so I'm going to give that a shot. Reviewing the data from the first flight also shows a significant altitude drop upon firing of the separation charge that took a while to recover (probably a combination of the charge firing plus the exposed coupler 'scooping' air post-separation), so I want to try a bit harder to seal-off this end, including giving more thought to the hole the EF TX antenna sticks through, which I had overlooked before the first flight. I may just use a dob of hot-glue or silicone caulk for this hole. Need to check the seal on the other screw holes as well.

 

[woferry]

I described the electronics above (ISC: Eggtimer Quantum & Eggfinder TX, Sustainer: RRC3 and RTx), so that wasn't so much the purpose of this post. I more wanted to comment on these electronics and my initial experiences with them. Finding the right power-up order for the RRC3/RTx combo took a bit, but has been solid since I got it all working. The base unit makes so many different beeps it seems to have its own language, the occasional odd beep in the middle of a sequence of 'normal' beeps is definitely off-putting but seems to be normal (my first flight had the rocket on the pad for over 30 minutes with me standing next to it, so I kept hearing the different beeping patterns when range was definitely not an issue). I had performed successful ground testing with both sets of electronics, proving that I could fire 2 e-matches with each output since I got burned once by a "successful" TRS test that, while it fired the ground-test charge, also destroyed the FET such that the first in-flight deployment failed, so I now do every test twice (not two loaded charges, but two e-match firings) to ensure that it's repeatable.

To flight-test the RRC3/RTx I first put up the booster only with the sustainer's recovery train, on an I140W (sim'ed to 1300'). There was something weird with the RRC3, it seemed to be about 100' off in its altitude measurements. The flight started at 0, but apogee was 100' less than what the AltimeterThree that also went along for the ride reported (and I've found my A3 to match my other altimeters [multiple TRSs, multiple SLCFs, EasyMini] within a few feet). And when the rocket landed the RRC3 reported the altitude as -100', but it's a very flat field, again the A3 reported 0 at start and landing. This repeated on the second flight as well. So that seems odd to me, not sure if that's just a 'feature' of the RRC3 since it's the first time I've flown it, I will reach out to MissileWorks, though if anyone reading this has compared RRC3 and other non-MW altimeter results I'd be curious to know if you've seen this also.

It seems to me that the RRC3 detects launch delayed, which I guess may be what leads to the 100' offset. For both flights it reported both apogee and the flight duration about 1-1.5 seconds less than the A3 did, and the firing of the airstart was later than I'd expected as well, having specified a time from launch. And by firing I mean when the video showed light smoke starting to come from the nozzle, which was still quite a bit before the Smoky Sam motor came up to pressure and finally lit with dark smoke.

The RTx was really slick, being navigated to the rocket without having to pull out my phone was nice (looking forward to the equivalent feature in the Eggfinder LCD). The only gripes I ran into there were that once you were on the right track and getting close the directional arrows seem hyper-sensitive, I could see that I was walking straight towards the rocket but the arrows were saying to turn right, which would have led me in the wrong direction. It seems to me instead of just "^^" (straight) and ">>" (turn right) there should also be a "^>" (straight, maybe a bit to the right) when you're close to the right heading. And one nice feature of the EF LCD/ET TRS combo is that after the flight it displays the maximum altitude, the RRC3 downlink on the RTx base unit simply goes to all 0's after landing, you have to connect to a PC (or count-out the RRC3's beeps) before you can get any basic flight data.

The 2-stage flight was also my first flight of the Eggtimer Quantum. I had used one for a few ground tests in the past, but it actually failed during a ground test (did the test successfully, then never connected on WiFi again, instead the WiFi module got very hot to the touch, so something obviously failed inside the module, though removing the shield I couldn't see anything bad, it was the chip itself getting hot). This one flew just fine, though when I downloaded the data the log stopped before it reached the ground. So the TRS generally reports flights as being longer than they really are, because it takes a few seconds after landing for the FVeloc reading to go to 0, but in this case the FVeloc was still -46 fps (and FAlt 100') at the last reading. The flight status was Landed so I don't see any reason to think it lost power, etc., it just seemed to stop recording prematurely, or when replaying the flight data left the final rows out. Will send the data to Cris to see what he thinks. It was the only altimeter I had in the booster so I don't have anything to compare the data against, I should have at least thrown my AltimeterTwo in for that flight to get a second opinion.

The A3 flew in the sustainer, so it captured the 2-stage flight, though the A3 app doesn't properly analyze 2-stage flights so far (I seem to recall John mentioning a while ago that he'd like to add this). And strangely if I enable charting of the accelerometer data the app instantly crashes, so I have data from this flight to send him as well.

The Eggfinder TX had flown before, and worked perfectly as always. Altitude reporting with these units has always been terrible (basically doesn't change for the entire flight), the RTx's ublox receiver did a much better job of tracking the sustainer's altitude. Unfortunately I had built an EF TX using the pin-compatible Maestro A5135H instead of the supplied A2235H, which is a SirfStar V chip instead of SirfStar IV (does GPS and GLONASS, and I'm hoping maybe does a better job of tracking altitude), but it wasn't the EF TX I flew on this flight. I was planning to fly it on the Double Shot, where I'll actually have 3 trackers on-board for the first flight, an A2235H-based EF TX in the booster, the RTx in the sustainer's av-bay, and the A5135H-based EF TX in the sustainer's nose cone, so that I can compare the RTx and A5135H. I have one EF RX that I hacked to support programmable frequencies and one stock unit that runs at a fixed frequency, so I can collect two streams at the same time, at least provided the fixed-frequency one isn't conflicting with anyone else's Eggfinder.

 

[woferry]

So this post isn't really anything specific to the 2-stage rocket, more of something else that went along with this flight that I was surprised worked as well as it did. If I think of anything else I may add it here as well.

I've loved Doghouse's charge wells with integrated terminals. You can see two of the 1.5g wells in the shots above. Unfortunately, when Binder picked up Doghouse's products, they've decided not to list these on their site. Instead there's a 3D-printed 'sleeve' that goes around his traditional charge well, which to me is not the same thing since it requires a third hole for attaching the well itself, and the sleeve itself seems worthless to me, I could easily use the screw posts without that, which would reduce the accuracy needed to drill holes to work with everything. But while I had a few extras of the 1.5g well sets, I was totally out of the 0.5g wells. And for the small nose cone volume in the Osprey 75 I wanted a small well, just no need for a bigger one. So I decided to try 3D-printing my own well, a direct copy of the Doghouse part.

I didn't feel all that good about the robustness of what I printed, as I prefer to print 6 walls (3 inner and 3 outer) on my 0.4mm nozzle to get what I consider to be a solid part, but the plastic part has thinner walls than this. Using my desired wall thickness would either increase the OD and interfere with the screws (if kept in the same place as the Doghouse wells), or reduce the ID to next to nothing, probably too small for the e-match to fit and reducing the amount of BP it could hold also. So I printed the walls using the thickness of the Doghouse part, but feeling concerned that the walls would be weak.

So I basically figured that I'd print a bunch of these, and it would probably break every time I used it, so I'd be replacing the part every flight. Well, the well survived the first (booster-only) flight without any issue, and on the second flight even face-planted in the dirt, so I had to scoop dirt and grass from in/around the well, and it feels just as strong as it did when printed. So maybe this small of a part will work better than I had expected it to. Still, the focus on this particular print was to match the dimensions of the stock 0.5g well, the code I wrote to print this can also print wells that are more optimized for my rules (I can even specify the exact volume I want), which would not use the same hole pattern but if I'm designing a new rocket and have a bit more space I could just use my dimensions instead of a direct clone.

 

[woferry]

So, having read bunches of threads around here on staging, there were several things I figured I'd "play it safe" on for my first attempt. I'd heard that CTI motors are easier to light (and had proven during ground testing that my RRC3 with a new 9V battery could NOT light an AeroTech FirstFire ignitor, I know that I'll have to do more to stage to an AT motor), and that certain propellants like Smoky Sam were even easier(/faster). I also knew that I wanted to keep things pretty low the first flight, both so that I could see everything and because it was fairly windy so I didn't want things blowing too far. Fortunately, Bay Area Rocketry had just gotten their first post-fire shipment from CTI a month ago, and was at least somewhat re-stocked on 38mm motors (not all of the propellants, but Whites and Smoky Sams at least seemed to have one or two in each size). When I saw that I could pick up the J400SS (Pro38-6G) and the I212SS (Pro38-3G) and the OR sim suggested ~1200' for the booster and ~3800' for the sustainer I figured I had a good set of motors. Of course there was no need for the booster to also be a SS since it was being lit by the pad electronics, but I figured I'd use the same look for both stages.

The rocket worked out to 14 lbs, 15 oz fully-loaded. Cg was 49.75" compared to an OR-calculated Cp of 53.395" (1.17 cal). Would have preferred a bit higher margin on the Cg/Cp, as it was I used a short A3854 adapter in the booster and my older, larger A3854 in the sustainer since that was more weight (barely) ahead of the Cp. I opted for a single #2-56 shear pin on the nose and a single pin on the ISC-booster coupling, no pins (or even holes for pins) on the ISC-sustainer coupling or the av-bay drogue separation point (I do have vent holes in each airframe section). The ISC-booster coupling also needed shored-up with a bit of blue tape to make for a tighter fit, to reduce the wobble of the sustainer as I held the rocket by the booster. Here's the OR model with the weight/Cg dialed-in to actual measurements, and the pre-flight sim plot based on that model.



Going by the nice write-up on staging Cris has in his Eggtimer Quantum Airstart Manual, I used the expected burn time of the J400SS (1.8s) to pick a separation time on the Eggtimer Quantum of 2.2s. I added 0.6s to set the RRC3 AUX channel firing delay (2.8s). The RRC3 was also configured with the altitude lockout, set to 500' since the simulation showed the rocket would be at ~557' at 2.8s (the sims actually said more like 600' when I programmed the RRC3, once I had the final weights it went down a bit, so I was realizing that I was pushing it when I went out to the pad, though I figured I'd rather fail-safe and not light the sustainer than lower the limit since it can only be programmed in units of 100').

Out at the pad things turned into a bit of a disaster. Fortunately I was put on the away pad (even though it was only a complex J, I didn't mind as I wanted it called as a heads-up flight anyway) and wasn't in competition with anyone else, as I tied up one of the two pads for a good hour or so. When I connected the battery to my Eggtimer Quantum at camp it didn't emit its usual beeping, which on my Quantum's speaker is more of a gurgle, and I've noticed times where I didn't hear it beep before but things seemed to work fine, so I didn't give it a second thought. Well, when I got out to the pad my phone wasn't even seeing the Quantum's WiFi network. So I realized that I needed to open-up the ISC and power-cycle the Quantum, but I didn't bring out the hex wrench I needed to undo the lower bulkplate screws, the screwdriver to remove the ISC shear pin, or the stick that I generally need to push the rather tight lid out of the coupler. So there was the first trip back to my camp to grab tools that I'll need to bring with me on future 2-stage launches just in case, but had never needed before. After opening the bay and power-cycling the Quantum it came up normally and I was able to connect. But for reasons I still don't quite understand it failed to arm, reporting lack of continuity on one of the channels (I forget which). Of course I'd already re-assembled the bay at this point, so here I was taking it apart again. The Quantum had gone unresponsive, I couldn't re-load the main page after the continuity failure.

So once again I opened up the ISC, and power-cycled the Quantum again. I had also re-checked the connections, but this time connected from my phone with the bay open. I figured I'd test continuity one charge at a time, so I had the lower bulkhead connector unplugged the first time I checked (only the separation charge, on the main channel, connected). I was incredibly lucky that things had gone so sideways that I wound up here, because something I had totally not noticed up to this point was that I had paired a 4-wire Doghouse header built by Bill (previous owner) with a 4-wire Doghouse header built by Binder (new owner), and the orange/brown wires were swapped between the two headers! So while I expected the main to show continuity and the drogue to show open, I saw the opposite on the Quantum's screen. It was immediately obvious what was wrong looking at the mated connector, and I had actually run into the same issue the day before while ground testing a different av-bay I'd built for a rocket that didn't make its first flight at October Skies due to my work on this 2-stager. I had ground tested the ISC as well, but I guess wasn't thinking about which match should have fired when as I did the testing, it didn't seem wrong at the time but it must have been backwards then as well.

At first I figured I could just re-configure the Quantum to make the 'main' channel handle the apogee charge and the 'drogue' channel handle the separation, but while the drogue channel could be programmed to the 2.2s delay, the main channel didn't offer nose-over as one of its options. So I needed to swizzle the pins on the 4-pin header before I could fly. Needless to say I didn't have the tool for that at the pad, so now I was carrying the ISC assembly back to my camp to do a quick pin-swap. That took all of a few seconds after the long walks back to camp (fortunately I was about as close to the LCO/RSO tent as one could park, but the ground was like walking on sand so it was slow going), and had verified continuity on both channels (with the right assignments) before leaving camp.

So after re-assembling everything it was all finally working as expected. The Quantum definitely got funny after arming it for flight, the load of the status page seemed to hang for a minute or two, and I thought I'd be tearing things apart yet again, but suddenly it started refreshing and seemed happy. I had left the sustainer's RTX/RRC3 (and the A3) running during all of this, since I wasn't concerned about battery life and didn't want to disturb things, though I hadn't considered that the RTx only records the GPS track for so long on a given flight (thankfully I hadn't used up all of the time yet). So I was listening to the seemingly-random squawks coming from the RTx base unit this entire time. Bruce, the club's Prefect, had come out to offer any help a bit earlier, of course it was mostly me bumbling around and 2 trips back to camp so there wasn't much for him to do, he did take a picture with me rather than the one I took of the rocket alone earlier.

I guess I should have been more shaken after getting so far before realizing that the ISC connector pinout was wrong, but it was the only thing that hadn't been flight-tested up to this point, so I was glad that I caught that before the flight (even if it was by complete accident) and was ready to fly.

Honestly, having seen so many 2-stage flight attempts where the sustainer failed to light, I was already assuming that's what was going to happen to me, either due to pushing it on the lock-out altitude or just because there's a lot of things that can go wrong in an airstart. I've always enjoyed the excitement of that moment where you're waiting to see if the second stage lights, such that I knew I'd want to build 2-stage rockets myself. I didn't realize how much more exciting it is when its your own rocket, after the sustainer lit I was euphoric the rest of the day.

Anyway, here's the actual flight plot, combining the Eggtimer Quantum and RRC3 data. To get the separation events aligned I had to add 0.55s to all of the RRC3 times. And, going by the Quantum's clock, separation occurred as intended at 2.2 seconds, sustainer ignition occurred at 4.85s instead of the intended 2.8s. So I updated the simulation for ignition at 4.85s instead of 2.8. Things match pretty well, though there are a few things I still need to correct. The apogee for both flights wasn't as high as expected, they're (barely) within the 20% motor tolerances and probably more due to some extra angling into the wind, I've never really played with the wind values in my sims but might give it a shot here to see how much it affects things. I also never loaded a second chute into the booster in the sim, though the descent rates between my 18" drogue and 36" main aren't all that different, I clearly could use to reduce the size of my drogues in this rocket for future flights.

 

[kevinkal]

I'll be looking forward to your updates and video Will. I enjoyed watching your flight last Saturday too.