I started VOLLEY with a question: how much of a satellite's initial orbital distribution can be handled by the system that releases it, before the satellite has to carry propulsion of its own?
The project investigates provider-hosted control of individual payload departure conditions: relative release velocity, direction and timing across one spacecraft or a manifest. Keeping the spacecraft mechanically and electrically unmodified is the design objective; the interfaces still have to earn that claim.
Computational engineering programme. Nothing has been built, fired, measured, qualified or flown.
Interactive project site · Documentation portal · Current work · Evidence · CAD · Open problems
Concept geometry, not manufacturing-ready parts. The current direction is a motor-charged mechanical release cell. Lunar carrier and larger-payload layouts are future concepts; no lunar lifetime or flight readiness is established.
Explore the 3D models and lunar orbits · CAD downloads and model guide · Engineering review
 |  |
| MC-L2 mechanical cell: 80 mm working-stroke target. Component envelopes await detailed CAD. | C0-S2 two-body release screen. Orbital lifetime, uncertainty and disposal remain open. |
| Concept assembly | STEP | STL |
|---|---|---|
| Mechanical cell | STEP | STL |
| Four-cell bank | STEP | STL |
| Lunar carrier | STEP | STL |
| Larger-payload pallet | STEP | STL |
 |  |
| A machine that actually existed in the engineering record. Gen5 is the frozen electromagnetic comparator, not the next design. | The idea that survives the machines. Same host, individually controlled departure conditions, then the spacecraft is on its own. |
A secondary spacecraft begins with the position and velocity its launch service gives it. Its separation mechanism adds another initial condition. VOLLEY asks when controlling that condition more deliberately is worth the hardware, energy and operational burden.
The host supplies navigation, attitude authority and any permitted orbital manoeuvres. The deployment system supplies the relative release condition and mechanical interface. After separation, the spacecraft follows the resulting trajectory and whatever capabilities it carries itself.
VOLLEY cannot provide continuing stationkeeping, collision avoidance or formation maintenance after contact ends. Release changes velocity at the release position; it cannot place a propulsion-less spacecraft into an arbitrary different circular orbit by itself.
The comparison therefore includes ordinary springs, release timing, host manoeuvres, cooperative spacecraft interfaces and payload propulsion wherever they can do the same mission. Programmability has to beat a competent baseline, not a deliberately weak one.
The project began in 2021 around a coilgun and a dedicated free-flyer. In 2023, learning from the spent-upper-stage platform idea represented by POEM reframed the host as an active post-primary delivery platform. By mid-2025 the coilgun had become a linear synchronous motor because commanding release velocity mattered more than simply ejecting a payload. In 2026 the stage itself became part of the machine.
 | |  |
| Gen4. Last hand-modelled generation. Detailed, but no committed STEP export and known geometry/analysis disagreement. | Gen5. Frozen electromagnetic baseline. Reproducible generated CAD and the manuscript comparator. The mass comparison failed. | Existing Gen6. Stage-integrated gas-guide investigation. Contact/release physics and accommodation remain unresolved; trim is suspended. |
The mission evidence no longer supports treating maximum release velocity as the design objective. The reference carried forward is an independent retained cell with a motor-charged mechanical accumulator, independent latch, short guided pusher and local catcher. It is pictured functionally because spring form, latch geometry, guide, catcher, motor and structure have not earned detailed geometry yet.
P92 reference-architecture screen · Full generation record · Decision lineage · Visual machine evolution
At the frozen Gen5 model point the calculated release velocity is 16.029 m/s, peak acceleration is 10.07 g, dry mass is 126.6 kg, and net electrical-to-payload efficiency is 18.8%. Those are outputs of one frozen configuration, not product requirements.
The mission studies are already showing why this matters. In one common-energy campaign the 11.8, 16.029 and 29.009 m/s release-authority screens all avoid host corrections. S4 sharpens the point: on its fine grid, the best tested BOLLEY, Gen5 and existing Gen6 campaigns all use the same 4.569852 m/s first release and the same 2.777987 kg ideal host-fuel result. Extra release authority above that does not improve this tested campaign.
That is not a universal optimum. It is a reason to stop treating maximum release speed as the product score. Timing, direction, terminal-state error, host manoeuvres, installed burden and controllability matter alongside raw velocity.
For a 4 kg payload, the S4 study point is only 41.77 J of ideal payload energy. At a 10 g constant-acceleration screen it corresponds to 106.5 mm of ideal stroke and 46.6 ms of acceleration. By contrast, even at the project's 25 g study ceiling, 16.029 m/s needs about 0.524 m of ideal stroke and 29.009 m/s needs about 1.716 m. Every extra metre and joule now needs a mission reason.
Latest review, 16 September: combined release errors now cover 512 conditional event corners. The reference-cell mechanics screen retains all 144 spring/pusher cases, including the 96 that fail its acceleration/contact assumptions. The verification matrix connects each result to the closure it still needs. Neither study closes the full campaign or P92.
Review and restart record · Installed burden and failure consequences · BOLLEY website
| Study | Cases | What it adds | What remains outside the result |
|---|---|---|---|
| S1 · Departure-state trade | 210 | Finite host recoil and assumed release authority | Complete campaign and installed-system benefit |
| S2 · Sequential campaign | 180 | Host propagation, remaining manifest mass and propellant allocation | Full position/velocity targets and useful constellation performance |
| S3 · Terminal-state timing | 300 | Identical terminal position/velocity target and release-time search for one payload | Coupled multi-payload optimization |
| S4 · Two-payload manifest timing | 100 | Shared-host continuity, two target orders, coupled release timing and full terminal-state acceptance | Clearing/settling intervals, release/navigation error, installed burden and complete-manifest closure |
S4 accepts 44 of 100 tested schedule/order/authority cases. Failed searches remain in the JSON; missing roots are not treated as proofs of infeasibility. P113/E5 therefore remain open.
The browser mission sandbox is deliberately not part of this evidence chain. It is a simple two-body visualisation for intuition. The validated studies above remain authoritative.
Provenance · Run sheets · Figure index · Visual evidence portal
The failures are not a footnote because several of them changed what VOLLEY is allowed to claim.
Gen5 loses its 3U dispenser mass comparison. The original unique constellation-phasing advantage was withdrawn after release timing proved to be a real competitor. The long gas guide exposed contact/release and model-form defects. The proposed electromagnetic trim path did not survive its own checks and is suspended.
The point is not to collect red boxes. The point is that a failed calculation must be able to change the machine.
Open problem register · Kill criteria · Computational closure
The redesign separates two decisions that earlier generations tended to collapse: payload arrangement and release mechanism.
Payload arrangement: independent retained cells are the current reference; small banks and the shared magazine remain charged alternatives.
How the release is produced: a motor-charged mechanical accumulator is the current reference; direct short-stroke electromechanical and compact gas pushers remain backups. The existing 8 m gas guide and Gen5 LSM remain historical comparators, while BOLLEY remains the cooperative-interface path.
The reason for the independent-cell reference is fault topology, not aesthetics. A blocked pusher should not mechanically strand unrelated payloads. Shared power, command and host services are still common-mode risks and are shown as such. The reason for storing release energy mechanically is equally narrow: it lets a small actuator charge slowly before release instead of asking the host electrical bus to source the full short-duration mechanical pulse.
This does not close P92. Installed mass, release repeatability, latch shock, pusher friction, tip-off, catcher loads, cycle life, host power/thermal limits and provider accommodation still have to decide whether the reference survives.
The existing reference is 4 kg / 3U. From there the work moves to adjacent payload sizes and 2-, 4- and 12-payload arrangements. Very small spacecraft are sensitive to adapter overhead. Larger spacecraft bring stronger interfaces, distributed force application and larger host reaction demands.
A programmable ground separation-test system is an adjacent application worth keeping: controlled push profile, instrumented surrogate, measured release velocity, tip-off and repeatability. It could answer release-physics questions before a flight product exists, provided gravity and test-support effects are accounted for.
1. Try to kill the reference early. Freeze coupon bands for accumulator force-displacement, preload repeatability, latch shock, pusher friction, exit velocity/tip-off and catcher load before detailed CAD.
2. Charge the installed burden. Compare independent cells, banks and the shared path using real structure, actuators, controls, energy, envelope and failure consequences per successfully delivered payload.
3. Put the missing mission physics back in. Add release/navigation uncertainty, clearing/settling time, host attitude recovery and complete-manifest effects to the mission comparison.
4. Close P92 or change the machine. Only then freeze the release cell, scale 2/4/12-cell arrangements and produce the drawings, BOM, assembly/inspection and instrumentation package for a named test article.
Programme execution · Mission and redesign · Engineering closure · Visual current-work portal
| Concept | Evolution | Mission sandbox | Evidence | Documentation |
| Mission boundary and real host context | Five years of machine changes and failures | Interactive release velocity, timing and propagation | Studies, run sheets, provenance and visual outputs | Mission, CAD, validation, build readiness and programme map |
The interactive project site is the visual front door. The repository remains the engineering record.
Use Python 3.12 or newer and a full-history clone. The companion provenance check needs the recorded export commit in local history.
python3 -m venv .venv
. .venv/bin/activate
python -m pip install -r requirements.txt -r requirements-dev.txt
bash tools/verify_all.shRun from a clean committed checkout. The command checks repository consistency and the numerical checks it lists; it does not rerun every external solver or certify the design. Missing checks remain visible.
| Repository | Role |
|---|---|
| VOLLEY | Flagship engineering record and unmodified-payload design objective |
| BOLLEY | Sister project: passive spacecraft interfaces in exchange for different launcher machinery |
| VOLLEY-paper | Gen5 manuscript and exported reproducibility evidence |
| VOLLEY-thesis | Gen5 submission material and exported evidence |
| VOLLEY-lab | Stopped alternatives and conditions for reopening them |
Adityavardhan Mishra · Mechanical Engineering, Symbiosis Institute of Technology, Pune
Project begun April 2021 · adityavardhanmishr@gmail.com
I welcome independent reproduction, design review and prototype collaboration. Please identify the configuration and run when reporting a discrepancy.
CC BY 4.0. See NOTICE, LICENSING.md and CITATION.cff for attribution and scope. Historical snapshots retain the terms under which they were released.