25 of 35 Fails by One: Two Scripted Runs of the 2026–2030 FCC Technician Mock and What the Subelement Grid Shows
Originally published on DEV.to — tested with MorsePass. The FCC Technician exam has 35 questions and you need 26 right. So you can miss nine. Miss ten and you fail, and the score alone won't tell you which ten were the
Originally published on DEV.to — tested with MorsePass.
The FCC Technician exam has 35 questions and you need 26 right. So you can miss nine. Miss ten and you fail, and the score alone won't tell you which ten were the expensive ones.
I wanted to see what a free practice exam actually shows you when that happens. So on Monday, November 16, 2026, I sat the free FCC Technician practice test on MorsePass twice: once answering every question from the official answer key, and once missing ten questions on purpose. Both runs used the pool that has been in effect since July 1, 2026.
Disclosure: I build MorsePass. A Playwright script in headless Chrome entered the answers, not my fingers, so the timings below are machine times and not human ones. Here is how the answers were picked. In run A, every answer came from the answer letters in the NCVEC's public pool document. In run B, the script picked the next letter after the correct one for every question in subelements T5 and T6, plus the first two T7 questions it met. Everything else in run B came from the official key.
First, the two numbers that matter (checked against the sources)
- Pass line: 47 CFR § 97.503(a) says Element 2 has "35 questions… The minimum passing score is 26 questions answered correctly" (eCFR, read November 16, 2026).
- Pool: NCVEC released the 2026–2030 Technician pool on December 18, 2025, and issued errata on February 19, 2026. The errata reworded four questions: T1C01, T5A05, T7A09 and T0A10. ARRL's write-up says the revised 2026–2030 Technician pool is valid for exams from July 1, 2026 through June 30, 2030.
I parsed the Feb 19, 2026 pool PDF and counted 409 questions across ten subelements. MorsePass says the same thing at the bottom of the exam page:
The site advertises "1,431 FCC questions." Its question sitemap adds up to that: 409 Technician + 423 General + 599 Extra = 1,431. Only the 409 Technician questions matter here. One more check: question T7A09 was in my run B exam, and the site shows the corrected February wording.
Step 1: open the exam page (no account)
There's no sign-up and no email. The Start exam button builds a 35-question set right away. The ham radio practice test hub lists the three classes side by side: Technician 35/26 (pool 2026–2030), General 35/26 (pool 2023–2027) and Amateur Extra 50/37 (pool 2024–2028).
Step 2: answer blind
When you click a choice, it just gets a highlight. There's no green tick and no red cross, which is how the real exam works. Each card shows the pool ID (T1C11 here), so you can look the question up later. In all 32 questions I could compare word for word against my pool parse, the A–D order matched the pool document. The script couldn't cleanly line up the other three.
The "Auto-advance after each answer" box starts out checked. My script unticked it and pressed Next question itself, so the pace stayed predictable.
Step 3: move through the palette
The 35-square palette lets you jump around and see what's still empty. Score official exam is already clickable on question 1. Every set follows the official structure of one question from each of the 35 groups. That means the subelement weights below are fixed and never random.
Run A: the answer-key run, 35/35
The result card said "Passing score · 35 of 35 correct · You cleared the 26 question passing line," with every tile full, from T0 3/3 to T9 2/2. Above the card: "Exam readiness: 100%." Every one of the 35 answers came from the NCVEC key, and the site marked all 35 correct. So for that set, the site's key agreed with the official one on every question. The whole script finished in under 23 seconds, page load included.
Run B: ten deliberate misses, 25/35
The script started run B at 10:02 (UTC+8). It was a fresh browser with a new random set, the same rules, and 10 wrong answers placed on purpose. From the click on Start to the result took 30.1 s, including four screenshots.
- Headline: "Keep drilling weak groups · 25 of 35 correct · You need 1 more correct answers to pass."
- Grid: T0 3/3 · T1 6/6 · T2 3/3 · T3 3/3 · T4 2/2 · T5 0/4 · T6 0/4 · T7 2/4 · T8 4/4 · T9 2/2
- Readiness: 71%, with the note "Readiness turns green after three attempts at 74% or better."
I also had the script compare all 35 questions in this set against the official key. There were 0 mismatches.
Here is the same run laid out against the syllabus weights:
This is why the grid matters more than the score. T5 Electrical Principles, T6 Components and T7 Practical Circuits are only three of the ten subelements, but together they put 12 questions on every exam. A learner who skips "the electronics part" has already spent more than the nine-miss budget.
Step 4: read "What to review next"
Under the grid, MorsePass sorts the misses into its own buckets: Equipment and circuits 6 · Calculations 3 · Schematics and figures 1. Then it links three lessons: Components and schematics, Electricity basics, and Ohm's law and power math. The page says outright that this is "guidance, not an AI-generated answer." These buckets are the site's own groupings. They aren't NCVEC subelements, so use the grid for "where" and the buckets for "what kind of thing."
Step 5: open the review book
Review missed questions goes to the local review book. It listed all ten misses: T5A11, T5B09, T5C03, T5D14, T6A11, T6B06, T6C04, T6D02, T7A09 and T7B04. Each one shows its correct answer, for example "3 dB" for the 5 W → 10 W question (T5B09), "Carbon-zinc" for the non-rechargeable battery (T6A11), and "High SWR" for low RF output (T7B04). The page says that questions "leave this list after you answer them correctly in a later exam," and that the history stays on your device. To dig into one subelement, the T5 question review shows all 50 T5 pool questions with their answers.
The free limit you'll hit
The free tier on the pricing page says "1 complete 35-question simulated exam per day." I hit that card while I was setting up the script. A later start that morning still went through, which is how I got run B. Plan on one scored mock per day unless you pay. The question browser, the review book and the subelement pages kept loading normally.
How I'd use this as a learner
- Take one blind mock a day. Write down the grid, not just the score.
- Any subelement under half is your next study block. Under the syllabus T1 has 6 questions, T5 to T8 have 4 each, T0/T2/T3 have 3, and T4/T9 have 2.
- Clear the review book before the next mock. Misses only drop off when you get them right later.
- For calculation misses (dB, Ohm's law), drill the formulas with the Ohm's law calculator or the Technician cheat sheet. Don't keep re-reading answer letters.
Numbers at a glance
| Run A (answer key) | Run B (10 deliberate misses) | |
|---|---|---|
| Score | 35/35, pass | 25/35, fail by 1 |
| Headline | "Passing score" | "Keep drilling weak groups" |
| Weak tiles | none | T5 0/4, T6 0/4, T7 2/4 |
| Readiness | 100% | 71% |
| Site key vs NCVEC key | 35/35 agree | 35/35 agree |
| Machine time | < 23 s whole script | 30.1 s Start → result |
| Account | none | none |
Pool: NCVEC 2026–2030 Technician, Feb 19, 2026 errata release, 409 questions, effective 2026-07-01 to 2030-06-30. Pass line: 26/35 under 47 CFR 97.503(a).
Earlier in this series: Koch Morse from zero in a browser tab covers the CW side of MorsePass (Koch levels and Farnsworth spacing). If your other exam is a quant-trading mental-math test, QuantSprint runs the same kind of timed browser drill for that.
— mengyuxuan
Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.








