Resistor Combination Finder

Ever tried to hit a weird target resistance and spent ages juggling E-series parts in series/parallel? This app does the book-keeping and instantly computes either (1): every valid combo within your tolerance (Path A) or (2): the Top-K# closest matches (Path B).

Quick start

git clone https://github.com/4rvyn/resistor_finder.git
cd resistor_finder
pip install -r requirements.txt
python main.py

Tested on Windows 10+, Python 3.13+.

Using it

Enter target + tolerance, choose E-series, decades, N (1–3), and topologies. Pick a Compute core:

• Path A: exhaustive within ±tolerance (deterministic & complete).
• Path B: best-first global search; returns exactly K#-number of best matches.

Then Run and (optionally) sort by Deviation (%) or Config Type.

Tip

Use Path B (Top-K) to scout great candidates in milliseconds, then —if needed— switch to Path A for a complete set within your tolerance.

Warning

Path A with wide ranges (E96 + wide decades + N=3 + Mixed) will generate millions of results and can crash the app. Use Top-K to explore safely.

Compute cores (what actually happens)

flowchart LR
  I[Inputs: target, tolerance or K, E series, decades, N, topologies] --> C{Choose core}
  C -- Path A --> A1[Path A step 1: binary search in sorted arrays]
  A1 --> A2[Path A step 2: band windows for R and G]
  A2 --> A3[Path A step 3: enumerate N equals 1 and 2 direct; N equals 3 via per k windows]
  A3 --> A4[Path A result: emit ALL solutions and display]

  C -- Path B --> B1[Path B step 1: seed streams S1 S2 P2 S3 P3 SP PS at target]
  B1 --> B2[Path B step 2: global min heap by absolute deviation]
  B2 --> B3[Path B step 3: pop best then emit then advance]
  B3 --> B4[Path B result: stop after K and display]

Loading

In short: Path A slices the sorted pair tables by your tolerance band to guarantee completeness. Path B runs a multi-stream best-first search (one global heap, no heuristic bias) that always yields the true Top-K fastest.

Precompute

Binary tables live in assets/:

• rcf_e12.bin, rcf_e24.bin, rcf_e96.bin: sorted singles and pair sums used by both cores.

Regenerate / validate (optional):

python precomp.py
python precompval.py

Note

If you want to change E-series or decade coverage, re-run precomp.py with your values so both cores stay fast and correct.

License

MIT: see LICENSE.