fpv: Pavo20 Pro II follow-up, the flight controller swap that fixed the GPS - #28
Open
foxis wants to merge 7 commits into
Open
fpv: Pavo20 Pro II follow-up, the flight controller swap that fixed the GPS#28foxis wants to merge 7 commits into
foxis wants to merge 7 commits into
Conversation
|
Bugbot is not enabled for your account, so this pull request was not reviewed. Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs. |
foxis
force-pushed
the
article/pavo20-taker-g4-gps-fix
branch
4 times, most recently
from
August 16, 2026 23:51
c7083bf to
3265e1a
Compare
Single post at content/fpv/pavo20-taker-g4-gps-fix/, EN and LT. Six failed experiments, then replacing the BetaFPV board with a GEPRC Taker G4 35A, plus everything that swap changed. The root cause section is now about an inductor and a missing layout. Both the 5V and 9V rails are built on a TPS63070 buck-boost, 2V to 16V in with a 3.6A switch current limit. The large ~2.5mm part beside it that looked like a capacitor is the inductor, found by probing with an in-circuit LCR meter. Other boards use physically larger inductors with properly enclosed ferromagnetic cores. This one appears not to. Two pieces of evidence make it more than a guess. First, the TinySA magnetic probe picks this up better than the near-field electric probe, which is backwards if the dominant escape route were E-field coupling from traces. Second, the TPS63070 datasheet's recommended EVM layout places C1 and C4 immediately beside the inductor and the IC, in the tightest part of the switching loop, and on this board those two are absent altogether. Other caps may be present but sit further out. So the components placed specifically to keep the hot loop small are the ones that were dropped. Together that points at switching energy being inductively coupled into surrounding circuitry rather than conducted or radiated, which explains why no shielding attempt helped, why shielded GPS wire capped at four or five satellites, and why powering over USB instead of the pack changed nothing. Added an original Graphviz diagram contrasting the recommended loop with the actual one rather than reproducing TI's copyrighted figure. Uses the brand palette. Verified viz.js renders it in both languages. Links TI SLVAEP5 with two caveats, one of which cuts against the argument: its measurements stop at 1 GHz while GPS L1 is at 1575 MHz, and while extrapolating the envelope upward suggests it does not stop there, resonant peaks are precisely what cannot be extrapolated. Stated as a guess, not a finding. Satellite counts in a table: 4-inch foldable ~17 typical and 30 once, Pavo20 on the old board 0 to 3 with zero after 15 minutes, Pavo20 on the Taker G4 eight in two minutes and 15 at best. Argued both ways. Motor failure is NOT attributed to 4S: the three wires physically tore because the glue was not holding them at the motor base, and that motor had been wiggling more than the others beforehand. Strain relief, not power level. Also documents antenna placement, motors soldered directly to the FC pads, and the small-antenna trade of worse reception against not destroying O4 U.FL connectors. Deliberately does not touch the previous article. 2374 prose words, 10.8 min at 220wpm. Kept whole deliberately: the root-cause section is 3.1 min and the build half is 4.0 min, so neither side of a split reaches the 5 minute floor that would justify one. Trimmed roughly 350 words of flab to get here rather than cutting substance. Images: 9 photos, 1400px wide, all EXIF stripped. One further photo was excluded because it showed a pilot license ID. Verified on a local Hugo 0.163.3 build: zero em-dashes, one Mermaid and one Graphviz diagram, no ASCII art, draft false, all images resolve in both languages, hreflang pairs EN and LT, BlogPosting valid in both, post in the sitemap, LT reviewer-note block stripped from rendered HTML. Dated 2026-08-18. Hugo excludes future-dated content unless --buildFuture is passed and pages.yml does not pass it, so this stays invisible until a build runs on or after that date, which is intentional.
foxis
force-pushed
the
article/pavo20-taker-g4-gps-fix
branch
from
August 17, 2026 07:46
3265e1a to
dcd8e60
Compare
Re-read the Lithuanian version properly and fixed 53 lines. The substantive
problems, in order of how badly they read:
1. WRONG TERM FOR THE COMPONENT, throughout. I used "induktyvumas", which is
inductance, the property, where the text means the inductor, the part. An
electronics reader would trip on every instance. Now "ritė", with
"induktyvumo ritė" on first mention. The three remaining uses of
"induktyvumas" are correct because they genuinely mean the property:
0805 package inductance and series inductance.
That change cascaded into gender agreement, since ritė is feminine:
"jis atrodo magnetiškai nesandarus" -> "ji atrodo magnetiškai nesandari".
2. "jungtuvo srovė" for switch current. Jungtuvas is a circuit breaker.
Now "komutuojamos srovės riba".
3. "Įtampos atotampos gedimas" for strain relief failure, which reads as
"voltage stay failure" because įtampa is also voltage. Now "laidų fiksavimo
gedimas".
4. "kirbėjo" (itched) where I meant klibėjo (was loose). The motor was wobbling,
not itching.
5. "kas mane pykina" means "which nauseates me". Now "kas mane labiausiai
nervina".
6. "prie 1 GHz" / "prie 1575 MHz" for frequency points. His existing published
Lithuanian uses "ties", so now "ties". Kept "smaigaliai" for spectrum peaks
because that is his established term in the previous article.
7. Straight closing quotes throughout: „veikia" -> „veikia“, seven pairs, plus
one set of ASCII quotes around "ne tas UART".
8. Typo "apvijų videje" -> "apvijų viduje".
9. Case and agreement: "dalinasi" -> "dalijasi", "kuo anksčiau nebuvo" -> "ko
anksčiau nebuvo", "už tuo" -> "už to", "ankstyvai" -> "anksti", "atsarga po
250 g" -> "atsarga iki 250 g", "ant šio rėmo" -> "su šiuo rėmu", "prie
perjungimo mazgu" -> "mazgų", "ankšti HF kilpa" -> "ankšta", missing copula
in "kuriuo esu labiausiai patenkintas", missing comma before "kurio".
10. "GPS sukasi ant UART1" -> "GPS prijungtas prie UART1".
11. Calques rewritten: "jokia ekranavimo kiekybė" -> "jokiu ekranavimu to
neišspręsi", "folija nenuperka to, ką intuicija sako" -> "nesuteikia to, ko
intuityviai lauktum", "kainuoja realų priėmimą" -> "pastebimai blogina
priėmimą", "rodo į plokštę" -> "rodo, kad problema plokštėje", "giliai
nepatenkinantis" -> "visiškai netenkina", "Ką jos daro, tai" -> "Užtat jos",
"kas bent patvirtina" -> "o tai bent patvirtina".
12. Graphviz labels now carry proper Lithuanian diacritics. I had stripped them
defensively; verified viz.js renders "Ši plokštė", "neįdėtos" and "be uždaros
šerdies" correctly, so the deliberate misspelling was unnecessary.
Verified on a local Hugo 0.163.3 build: 2015 prose words, 9.2 min, two
em-dashes against a budget of six to eight, zero HTML comments, all ten images
resolve, Graphviz and Mermaid both render, seven matched quote pairs.
…anian Andrius confirmed smailės is the correct term for spectrum peaks. I had kept smaigaliai for consistency with the previous Lithuanian article, which was the wrong instinct: that article is machine-translated output, not his own writing, so it is not a style precedent to follow. Also formalised what I had wrongly assumed was his voice: Buvau gazavęs iki galo -> Buvau iki galo atidaręs trauką prie didelio gazo -> esant didelei traukai upgrade'as -> patobulinimas MOSFET'ai -> MOSFET tranzistoriai whoop'as / whoop'ui -> whoop dronas / whoop dronui (indeclinable + noun) smailė is feminine, so the participle agreement moved with it: pasklidusius -> pasklidusias, rezonansiniai -> rezonansinės.
I over-corrected. Apostrophe declension of unadapted loanwords is correct Lithuanian usage, so whoop'as, whoop'ui, whoop'o and MOSFET'ai are restored. Only the slang was genuinely wrong, and those fixes stay: gazavęs iki galo -> iki galo atidaręs trauką, prie didelio gazo -> esant didelei traukai. The USB-C line was a bad translation rather than a loanword problem. "A drawback and an upgrade at once" is the idiom pliusas ir minusas, so it now reads "Vienu metu ir pliusas, ir minusas." smailės stays.
One instance, aukštų frekvencijų -> aukštų dažnių. dažnis is the house term across his Lithuanian files by roughly 95 to 20, and it is what a physicist would write.
Two things were wrong with it.
sutvarkyti is transitive and had no object. "bandžiau sutvarkyti kaip reikia"
leaves the verb hanging: fix what? Lithuanian needs the pronoun. Same fault in
the echo, "Kaip reikia nesutvarkiau."
And "kaip reikia" means "in a fitting manner", which is not what the English
"properly" carries here. The point is that he never addressed the root cause,
which the article admits outright a few screens later: "niekada to neatskyriau".
So the phrase should mean "at root", not "tidily".
was: viską surašiau, o paskui savaites bandžiau sutvarkyti kaip reikia.
Kaip reikia nesutvarkiau. Pakeičiau skrydžio valdiklį.
now: viską surašiau, o paskui kelias savaites bandžiau tai išspręsti iš esmės.
Iš esmės ir neišsprendžiau. Pakeičiau skrydžio valdiklį.
The English repeats "properly" verbatim to land the reversal, so the Lithuanian
repeats "iš esmės" the same way. "ir" carries the resigned tone that "I did
not" has in English. Also "savaites" -> "kelias savaites", closer to "spent
weeks".
Applied verbatim: Iš esmės ir išsprendžiau. Tiesiog pakeičiau skrydžio valdiklį. Vat taip vat... This inverts what I had. Mine was self-deprecating failure, "iš esmės ir neišsprendžiau". His is the anticlimax: essentially he did solve it, he just swapped the whole board, shrug. Funnier, and truer, because the problem really is gone. The irony carries the point that replacing an entire flight controller is a sledgehammer rather than a root-cause fix, which is what the rest of the article argues. The repetition of "iš esmės" across the two sentences is kept, since that is what makes the reversal land. NOTE: the English now makes the opposite rhetorical move, "I did not fix it properly." Both statements are defensible and the Lithuanian is his own wording, so it is left as is, but the divergence is deliberate rather than an oversight.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Single post, EN + LT, 11 files. Stacked on #27; merge that first and I'll retarget.
Root cause: a leaky inductor and a missing layout
The C1/C4 finding is the strongest evidence in either article, because it is specific and checkable.
So the components placed specifically to keep the hot loop small are exactly the ones that were dropped.
Two independent things now point the same way: the H-field probe beating the E-field probe (backwards, if trace E-field coupling dominated), and the missing loop caps. Together they support switching energy being inductively coupled into surrounding circuitry rather than conducted or radiated, which is why nothing you shielded helped, why shielded wire capped at 4-5 satellites, and why USB power changed nothing.
The diagram is mine, not TI's
Rather than reproduce TI's copyrighted Figure 51, I drew an original Graphviz comparison of the recommended hot loop versus the actual one, in your brand palette. Verified
viz.jsrenders it in both languages. No copyright question, and it fits your diagram rules.The caveat that cuts against you
I put your extrapolation point in, and then the counter-argument, because it belongs there:
That keeps your conclusion while being honest that the app note data stops below your frequency.
Length: 10.8 min, kept whole
2374 prose words. Over the 10-minute guide, and deliberate: the root-cause section is 3.1 min and the build half is 4.0 min, so neither side of a split reaches the 5-minute floor you set. I trimmed ~350 words of flab getting here; further cuts would have removed substance.
Everything else
5V BEC, no banner, same title and URL.Verified
Zero em-dashes, one Mermaid + one Graphviz diagram, no ASCII art,
draft: false, 9 images resolving in both languages, hreflang pairing EN-LT, validBlogPostingin both, post in the sitemap, reviewer notes stripped from rendered HTML, all images EXIF-stripped,blog/public/CNAMEuntouched.Per your call: the old-SHA photo and the publish timing are both closed. Dated 2026-08-18 and it stays invisible until you merge and rebuild, which is what you want.
Lithuanian terminology queries
Moved out of the file and into this PR deliberately. An
ANDRIUI PERŽIŪRĖTIcomment inside a committed article reads as "someone else wrote this and I did not check it", and the repo is public. It does not belong in the artifact.Terms I could not confirm against real Lithuanian FPV usage. Correct any and I'll update both versions:
persodinti smegenisšuntavimas / šuntuotiartimasis laukasįtampos atotampamagnetiškai nesandarusjungtuvo srovėH zondas / E zondasperjungimo mazgasankšti HF kilpaAlso: Graphviz node labels in the LT version are intentionally written without Lithuanian diacritics, for font rendering in
viz.js.Verified the LT source and both minified and unminified output contain zero HTML comments.
Pre-existing, not fixed here
The same pattern is already merged and live in your Meteor75 post, in public source:
blog/content/fpv/meteor75-pro-ii-fighting-resonances/index.lt.md—ANDRIUI PERŽIŪRĖTI, plus "nesprendžiau nė vieno, palikta tau"blog/content/fpv/meteor75-pro-ii-fighting-resonances/index.md—DRAFT NOTESblockNot touched here, since that is published content and your call. Happy to strip both in a separate PR.