The Poking Ape and The Ham Radio Crashed Course

Two hams, one Baofeng Mini, and a TinySA walk into a “purity test”... and everyone leaves with strong opinions and weak traceability.

By ON6URE (RF.Guru) — certified in sarcasm, not ISO 17025
(Measured with adapters, assumptions, and the ancient RF proverb: “Close enough for YouTube.”)

Josh — Ham Radio Crash Course

The Smoking Ape

The TinySA Purity Trials: “NIST Traceable by Vibes” Edition

So here we are: two hams, one Baofeng Mini, and a TinySA — a device that is absolutely wonderful at what it’s meant for… and absolutely not the thing you use when you want to speak with the booming certainty of an accredited EMC lab.

And yet, that’s exactly the energy:

“Your radio is dirty.”
“No, YOUR test is dirty.”

All being argued while the measurement chain is held together by adapter stacks, wishful calibration, and finger-tight SMA... the most powerful random number generator known to amateur radio.

The first mistake: treating a TinySA like a court of law

A TinySA is a pocket spectrum analyzer-ish tool. Great for:

• Finding signals
• Checking relative changes (“did my filter help?”)
• Quick sanity checks
• Learning RF fundamentals

Not great for:

• Declaring compliance
• Arguing pass/fail like you’re reading a calibrated report
• Pretending it has the same confidence level as a proper analyzer with preselection and a known uncertainty budget

The real comedy is this: someone will say “It’s not lab grade” out of one side of their mouth… and out of the other side they’ll say:

“This radio is definitely clean/dirty.”

Guys. You can’t do “definitely” with “probably.”

“Purity test” is not a vibe... it’s a procedure

What these tests are actually about: harmonics and spurious emissions relative to legal limits, using a defined method.

What YouTube turns it into: a morality play where harmonics represent sin, and the TinySA is the confessional.

The problem isn’t that measuring harmonics at home is impossible. The problem is that method matters so much (RBW, detector behavior, calibration, attenuation correction, coupling method, load conditions, and dynamic range) that you can make the same radio look “clean” or “dirty” just by shifting one knob or swapping one adapter.

Which is exactly what happens when people argue screenshots instead of uncertainty.

The TinySA’s dirty secret: it also makes spurs

Yes, internal spurs exist. “Spur removal” exists. Attenuation steps (10–20 dB) can help verify whether a spur behaves like something real or like analyzer artifacts. That sanity technique is... genuinely reasonable.

But here’s the punchline:

The TinySA is part of the experiment.

So when you see spurs, it might be:

• Your radio
• Your attenuator chain
• Your coupler (and its ripple and leakage)
• Your cables/adapters
• Your measurement settings
• The TinySA itself
• Or “all of the above, with bonus mystery resonance”

When your tool can generate artifacts, you don’t get to argue “truth.” You get to argue “clue.”

The great attenuator fantasy: “It’s just a pad, bro”

Attenuators are essential. A good passive pad, within spec, doesn’t magically create harmonics.

But also... not all attenuators are equal:

• Some are frequency-lumpy
• Some have poor return loss (hello, standing waves)
• Some are not rated for the power people throw at them
• Some are mislabeled (especially cheap ones)
• Some are “30 dB” in the same way my dog is “mostly a wolf”

And even if the attenuator is perfect, the system still isn’t: connector repeatability, adapter stacks, coax loss vs frequency, mismatch, coupler ripple, leakage, near-field pickup... it all piles up.

So when someone says “I used 70 dB of attenuation” like that number descended from a mountain on stone tablets, the correct response is:

“Cool. What’s the tolerance? Flatness? Return loss? Power rating? Frequency range? Verified with what?”

The “lab environment” myth: your bench is an RF haunted house

A real compliance lab spends absurd money to reduce surprises.

A typical ham bench has switching supplies, USB cables acting like antennas, laptop noise, LED lights, random metal objects, and grounding that’s “generally in the area.”

So you can absolutely end up measuring:

• Radiated junk coupling into your setup
• Analyzer overload artifacts
• Cable shielding imperfections
• Your own noise floor pretending to be spurs

Which makes the online fighting extra funny because it’s like:

“My haunted house says your ghost isn’t real.”
“Wrong, my haunted house says it IS real.”

Dynamic range is not a feeling

Here’s why both sides can “prove” their point:

• Too little attenuation → overload risk, internally generated junk → radio looks “dirty.”
• Too much attenuation → harmonics vanish into the noise floor → radio looks “clean.”

So the conclusion can become a function of input level, analyzer nonlinearity, noise floor, RBW/VBW, gain/attenuation mode... and whether your measurement chain is behaving today.

That’s not science. That’s metrology roulette.

Bonus debunk: the math got “spurious”

There’s a moment in these debates where a limit gets translated into dBm... and the numbers wobble.

25 µW is not -6 dBm.
It’s about -16 dBm.

Because:

• 1 mW = 0 dBm
• 25 µW = 0.025 mW
• dBm = 10 log10(0.025) ≈ -16.0 dBm

If your spurious limit shifts by ~10 dB mid-sentence, that’s not “nuance.” That’s the kind of thing that makes metrology people start stress-eating calibration stickers.

The “dBc stability” trick: useful... not a magic wand

The idea that real spurs remain at a similar dBc while internal distortion changes with attenuation... that’s a real diagnostic concept.

But the trap is that weak spurs can “move” because you’re bouncing around the noise floor, detector behavior, and sensitivity changes. Internal gain changes shift the apparent floor, instrument response at higher harmonics can be weird, and “spur removal” has limits.

So yes: good diagnostic. No: not a compliance certificate.

The “over-the-air analyzer next to the antenna” moment

Putting a handheld analyzer inches from a transmitting handheld and declaring “See! Fifth harmonic!” is like:

• Putting a stethoscope on a subwoofer
• Hearing distortion
• Then declaring the song is recorded badly

At that distance and power level, you’re extremely likely to be overdriving the front end, generating distortion in the receiver, and measuring near-field coupling weirdness. That “test” is an anecdote generator.

What both sides keep forgetting

Not that the TinySA is limited... they say it. They just forget what that means emotionally.

A limited instrument means limited certainty. And limited certainty means:

• Don’t declare absolutes
• Don’t turn borderline plots into verdicts
• Don’t fight like you’re defending a certified report
• Don’t let your narrative outrun your measurement chain

Because the real “dirty emission” in these videos isn’t the harmonic... it’s confidence leaking into frequency bands it’s not licensed for.

The honest conclusion

The TinySA is a flashlight. It can absolutely show you where the mess might be.

But both sides keep using the flashlight like it’s a forensic lab, a judge, and a courtroom stenographer... and then arguing about who’s holding it at the correct angle.

The “how to do it less wrong” checklist (still at home)

• Use a dummy load rated for the power
• Use a directional coupler or known safe sampling method
• Use known-good attenuators (ideally verified, or at least characterized)
• Keep the analyzer out of compression (prove it by stepping attenuation and seeing linear tracking)
• Show the noise floor and document the settings
• Use consistent RBW/VBW and state them
• Use a known low-pass filter as a sanity check (and understand mismatch effects)
• Treat the result as screening, not certification
• If it’s close to the limit... stop performing YouTube law and get a better measurement

Appendix: ISO 17025 YouTube Lab Report (unofficial, yet emotionally binding)

Laboratory Name: “Certified Comment Section Metrology”
Calibration Status: “I said ‘calibrated’ out loud.”
Traceability: “NIST... Traceable... by vibes.”
Uncertainty Budget: “± one comment section.”
Attenuator Verification: “Amazon listing says DC–6 GHz, probably.”
Connector Torque: “Finger tight, then one more inspirational twist.”
Environmental Controls: “LED strip on ‘rainbow’ mode for stability.”
Pass/Fail Criteria: “If the screenshot looks clean, it is clean.”
Witness Signature: “Eric said it’s fine.”

Interested in more quiet RF heresies?
https://shop.rf.guru/pages/subscribe

Questions or experiences to share? Feel free to contact RF.Guru.