Skip to content

Product path

What would turning this from a graduation deliverable into a product actually take?

The reality is: the gap is not the inverter — the as-built hardware/firmware is small and well-characterised. The gap is everything that surrounds a product.

Engineering work

In rough order of cost (cheapest first):

Track Effort estimate (engineer-months) Notes
LC output filter 1 Hardware + characterisation. Required before any non-trivial load.
Closed-loop control 2 Software + tuning. Requires AC voltage sense (hardware respin) or feedback reconstruction (software-only).
Thermal enclosure 2 Mechanical + EMI rework. Required for any enclosed deployment.
Dashboard productisation 1–2 Auth, multi-unit support, telemetry persistence, alerting. Currently single-instance dev tool.
Field test rig 1 A safe way to run the inverter on real load before deployment. Includes ground-fault detection on the load side.
Manufacturing-test fixture 1 A bed-of-nails or pogo-pin jig + a structured test plan for each fab batch.
Grid tie 4–6 PLL, anti-islanding, compliance testing. The hardest single track and the longest calendar wait (for utility witness).

Compliance work

Standard Required for Where
IEEE 519-2022 Any deployment with non-trivial harmonic interaction (grid tie, shared bus) Accredited test lab, ≈ 1 month wait
IEEE 1547-2018 Grid-coupled deployment Accredited test lab + utility witness
IEC 62109-1 / -2 PV inverter safety Accredited test lab
CE / UKCA / TSE marking Regional regulatory marks for retail product Notified body in Türkiye
Operational manual Real one, with safety notices and a service procedure Tech-writing engagement

Compliance testing typically costs ≈ 50,000 TL per standard at a Turkish accredited lab (numbers vary; check current EPDK-approved labs). Total compliance budget: 200–400 k TL for a grid-tied PV inverter.

Minimum-viable product paths

Three plausible MVP scopes, smallest first:

MVP-1 — bench instrument for teaching (smallest)

No new engineering, no compliance work. Package the as-built into a kit form for university labs:

  • Two pre-populated single-bridge PCBs.
  • Pre-flashed STM32 Nucleo-F303RE.
  • Pre-installed dashboard on a Raspberry Pi.
  • Step-by-step exercise sheets (bring-up → STAIR → STAIR_ALT → PSC).
  • Safety enclosure for the supplies + load (the bridges themselves can stay open for visibility).

Price point: ≈ 3000 TL per kit. Market: Hacettepe + other Turkish university EEE departments running power-electronics courses.

MVP-2 — off-grid PV+battery hub (mid-scale)

LC filter + thermal enclosure + closed-loop control (so ≈ 5 engineer-months), no grid-tie work. Sell as a residential / small-commercial PV-storage unit:

  • Built-in MPPT (which is a separate addition the project hasn't started).
  • Battery management.
  • Off-grid output to local load (no utility involvement → no IEEE 1547 → no anti-islanding → much shorter compliance loop).
  • IEC 62109-1/-2 still required for safety.

Engineering effort: ≈ 6–8 engineer-months total. Compliance budget: ≈ 100 k TL.

MVP-3 — grid-tied PV inverter (full product)

Everything in the roadmap, plus the compliance work, plus the utility relationship.

Engineering effort: ≈ 18–24 engineer-months. Compliance budget: ≈ 400 k TL. Calendar: ≈ 18 months from start to first utility-approved unit.

Realistic recommendation

If the team or a follow-on group wants to keep this alive as a product candidate:

  1. Pick one application — single-phase off-grid PV (MVP-2) is the most natural fit for the topology and the team's existing knowledge.
  2. Build the LC filter first (cheapest single track, unblocks everything).
  3. Build the closed-loop control in parallel with the application-specific work.
  4. Defer grid tie + compliance to a phase 2 where there's funding and a lab partner.

For a research / educational use, MVP-1 is essentially free — the project is already useful as a teaching tool for CHB modulation, gate-drive isolation, and bring-up workflow. The documentation here supports that use case directly.

What kills product attempts at this stage

Three failure modes, in order of likelihood:

  1. Underestimating compliance. The compliance budget often exceeds the engineering budget; founders running out of money at the certification step is the most common failure.
  2. Trying to skip the LC filter. "We'll do it in v2" is how a product never reaches v2.
  3. Over-scoping the first release. MVP-1 is shippable in months. MVP-3 is shippable in years. Picking the wrong starting point usually means picking the harder one.

Where this leaves the graduation work

This project is the inverter component of any of those three MVPs. The architectural decisions (CHB topology, TLP250 + B0515S isolation, PSC modulation, modular per-bridge boards) all hold up — they were each justified through the design notes and survived the bench-validation. The graduation work is the foundation. The product path is what someone builds on top of it.