The Great Watts Rip-Off: Right Diagnosis, Wrong Verdict?

A station is not a radio. That is the valuable message in Dave G3LRC’s video The Great Watts Rip Off / Scam. Transmitter power, feedline, matching system, antenna, return paths, local noise and operating choices work together. I agree with the diagnosis. I do not agree with every step from that diagnosis to the verdict.

Too often, we buy a beautifully specified black box and treat everything outside its antenna socket as an accessory. Then we spend more to overcome a loss, a poor pattern or a noise problem that the new radio cannot fix. Dave is right to challenge that habit.

But neither an imperfect antenna nor a 100 W transceiver establishes that somebody was deceived. A compromise may be exactly what the installation needs. The engineering question is whether its limitations are understood and whether it does the job. Buy the station you need, not the largest number on the front panel.

Dave’s point is real: watts buy logarithmic steps

For powers measured on the same basis, the change is 10 log10(P2/P1). Doubling is about 3 dB; ten times the power is 10 dB.

Dave acknowledges that extra power can matter in a pile-up or contest. It can also make an unreadable signal readable. Calling the increase “only” 10 dB must not hide that benefit: the useful margin depends on the path, interference, mode and receiving station. Nor does a power ratio guarantee a particular movement on an uncalibrated S-meter.

My practical agreement is about priorities. Recovering an avoidable 3 dB loss doubles the power available beyond that loss without doubling transmitter output. Improving the antenna pattern toward the wanted path may help too, by a different mechanism. Compare those opportunities with an amplifier rather than assuming either purchase always wins. More power remains a legitimate next step, within the station’s ratings and operating limits.

The coax example is conditional, not a default HF loss

At about 16:27 in the video, Dave starts with 100 W, subtracts 3 dB for the feedline and another 3 dB for a short loaded vertical relative to a dipole. The arithmetic gives roughly a quarter of the starting power in that comparison. Later, he describes the antenna term as inefficiency. That change of meaning matters.

First, the coax figure must belong to an actual cable and installation. Belden’s 8219 data gives nominal attenuation of 1.3 dB per 100 ft at 10 MHz and 3.1 dB per 100 ft at 50 MHz. A 100 ft run is about 30.5 m. Those figures illustrate frequency dependence; they do not make 3 dB the normal loss of a short HF feedline.

The cable label alone is not a verdict either. Length, frequency, construction, condition, connectors and load mismatch determine the result. Choose the feedline for the installation instead of treating every RG-58 run as either harmless or hopeless.

Second, mismatch and dissipation are different. Reflection is power travelling back along the line, not automatically heat. Standing-wave voltage and current can increase real cable loss, while a tuner can give the transmitter a good match without removing those standing waves from the antenna-side line.

Weaker in one direction does not mean those watts became heat

Antenna gain combines radiation efficiency and directivity: G = η × D, using consistent linear quantities. Efficiency describes the fraction of accepted power radiated. Directivity describes its angular distribution. The MIT antenna notes separate these quantities explicitly.

So a loaded vertical can be down in a chosen direction because of soil or coil dissipation, a different elevation pattern, or both. Polarisation mismatch can further reduce the received signal without turning that missing link power into heat inside the transmitting antenna.

This is not a defence of inefficient installations. If the coil or matching transformer dissipates power, that is a real penalty. But changing height or current distribution may improve the wanted path by redirecting radiation, not by recovering the same number of watts from a resistor.

The same care applies to common-mode current. Current on the outside of the coax can change radiation, noise pickup and interaction with the installation; it is not intrinsically a loss resistor. Where that current flows determines which effects occur. The right question is not simply how much power the coax “stole.” It is what reaches the wanted path, and what limits the return path?

A compromise is not the same thing as a false promise

Dave’s criticism becomes too broad when an end-fed wire or a quick-deploy loaded antenna is enough to condemn the purchase. A portable operator may need low weight and fast setup. A balcony operator may have no room for the antenna that wins an unconstrained comparison. A temporary antenna can be a good engineering decision without being the most efficient antenna imaginable.

That does not excuse miracle claims. Low SWR is not proof of high radiation efficiency. A string of distant contacts is not a measured pattern. An antenna sold across many bands still has matching, material, current-path and pattern limits.

A single high-ratio transformer serving an 80–10 m EFHW faces a much wider frequency span and changing loads than a design confined to one band or a bounded pair. That is a reason to demand an honest account of transformer loss, voltage stress and operating limits—not a reason to declare every end-fed antenna fraudulent. Narrowing the design range can make those burdens easier to manage; it does not make the rest of the installation disappear.

I would rather put a suitable antenna at a useful height, give it a sound return system and control unwanted feedline participation than buy more transmitter power to cover an avoidable weakness. Where space prevents that, I want the compromise stated plainly. The problem is not compromise. It is selling compromise as though nothing was traded away.

Reciprocity is real; the receive-SNR shortcut is not

At about 26:35, Dave takes improved antenna gain into a claim about an improved receiver. A reciprocal antenna has corresponding transmit and receive patterns at the same frequency and configuration, including when the antenna has passive loss. Reciprocity is not restricted to a perfect, lossless antenna.

But received signal power is not signal-to-noise ratio. Noise also enters through the antenna pattern, and the receiver adds noise of its own. The same MIT notes describe received noise through the angular response, not merely gain toward the wanted station.

Consider an antenna change that raises both the desired signal and dominant external noise by 3 dB, with receiver noise negligible. The signal is stronger, but their ratio is unchanged. Alternatively, a pattern null aimed at local interference can improve readability while the wanted signal stays the same—or even falls a little. In a receiver-noise-limited setup, reducing antenna or feedline loss can indeed improve SNR. The outcome depends on the noise terms, not on a slogan.

That is why I take receive-only antennas seriously. A loop, Beverage or phased receive array may be useful because of where it listens and what it rejects, not because it would make a better transmitting radiator. Common-mode control and placement may keep local noise out of the receive system. None of that contradicts reciprocity.

Receiver tables are useful tools, not shopping commandments

Dave is right that a receiver ranking can become a trophy. If the station is dominated by external noise, buying extra sensitivity may achieve little. But it does not follow that better receiver performance is generally wasted.

Sherwood Engineering’s table states its sorting basis: narrow-spaced third-order dynamic range, or reciprocal-mixing dynamic range when phase noise is limiting. That answers a strong-signal question under particular test conditions. It is not a universal ranking of complete stations or operating value.

A busy contest band, a nearby transmitter and a quiet-band weak-signal path do not demand the same things. Filtering, overload behaviour and close-in phase noise can matter even when the external noise floor is high. Conversely, adding preamplifier gain where it is unnecessary can reduce strong-signal headroom. That is a gain-setting problem, not proof that a good sensitivity specification is harmful.

My conclusion is not “ignore the laboratory.” It is use the laboratory results to solve your station’s problem. Buy the receiver strengths you can use, rather than its position in a table.

FT8 success is not an antenna-efficiency certificate

Toward the end of the video, Dave argues that weak-signal modes can make an inefficient station appear satisfactory. There is a useful warning there: a contact log does not explain how much margin the station has left.

FT8 combines structured messages, narrow signalling and forward error correction. Its developers document both the protocol and the effect of propagation conditions on decoding in The FT4 and FT8 Communication Protocols. The WSJT-X guide defines SNR reports against a 2,500 Hz reference noise bandwidth, not simply the occupied bandwidth of the FT8 signal.

At about 33:36, the video then links questionable SNR reports to reciprocity and antenna inefficiency. That does not follow. If an inefficient transmitting antenna delivers a weaker signal to the distant receiver, a lower reported SNR can reflect a genuinely weaker signal there. It is not made inaccurate merely by reciprocity. The local receive report has its own noise environment and cannot be substituted for the distant one.

Reports still have estimation limits, fading and operating-setting dependencies. A single FT8 QSO cannot separate antenna efficiency, directional gain, power, propagation and the other station’s noise. Repeated controlled comparisons can be useful; uncontrolled reports are not a calibrated efficiency measurement. FT8 proves that communication worked under those conditions. Enjoy the contact without pretending it measured every component.

A wattmeter can locate part of the problem—not deliver the whole verdict

To turn Dave’s missing-watts argument into a useful diagnosis, declare the measurement plane. At a real reference impedance, forward and reflected powers measured in the same stable state give Pnet = Pforward − Preflected. Thus 100 W forward and 10 W reflected means 90 W net crossing that plane; |Γ| = √(10/100) = 0.316, return loss is 10 dB and VSWR is about 1.92:1. It does not mean 90 W radiated.

Bird’s Model 43 instruction book gives the same forward-minus-reflected relation. Meter position matters: a transmitter-socket reading includes different downstream losses from a reading at the antenna feedpoint. Subtracting values taken at different planes or operating states is not that calculation.

The instrument also sets the uncertainty. Finite coupler directivity leaks forward response into the reverse channel, where the samples add vectorially. Bird specifies more than 25 dB directivity and CW accuracy of ±5% of full scale for the Model 43: on a 100 W element, that basic accuracy term is ±5 W, not ±5% of a small reading. Select suitable elements and ranges within their ratings.

Match the detector to the signal. CW average power, AM carrier power, digital-waveform average power and SSB peak-envelope power are different quantities. The standard Model 43 is not automatically a PEP meter; the 43P adds peak-reading circuitry with its own response limits. A slowly moving pointer does not establish speech PEP. For an output check, use the manufacturer’s prescribed test signal, a suitably rated dummy load and the appropriate sensor.

For feedline loss, compare Pnet,in and Pnet,out in the same stable state: LdB = 10 log10(Pnet,in/Pnet,out). Include instrument insertion effects. For a line with negligible radiation, their difference measures dissipation; if exterior feedline current radiates, the difference is not a pure heat measurement. The same-plane mismatch term −10 log10(1 − |Γ|²) describes accepted versus incident power, not dissipation.

A low-power VNA can characterise small-signal match and transmission; it does not by itself reproduce hot ferrite, arcing or high-power drift. Record calibration, reference planes, frequency, waveform, duty cycle, load, temperature, raw readings and uncertainty. Keysight’s power-measurement guidance and NIST’s calibration work explain why mismatch, detector response and measurement setup remain part of the answer.

These checks serve the station question. They tell us whether there is avoidable loss to fix. They do not turn a wattmeter reading into an antenna pattern, a receive-SNR measurement or a judgement on every antenna of that type.

The real lesson is to build the station, not worship the box

A 20 W station can beat a 100 W station if its antenna-system advantage toward the path exceeds the roughly 7 dB power deficit. A 100 W station with a suitable antenna can be excellent. An amplifier can be a rational upgrade once the limiting problem is insufficient transmitted signal rather than avoidable loss or local receive noise.

I agree with Dave that the antenna system deserves serious attention and a fair share of the budget. I disagree with turning every constrained installation into evidence of a scam, treating gain as heat, or using reciprocity to promise a fixed SNR improvement.

Fix real dissipation when it is the weakness. Improve height or pattern when radiation is going to the wrong place. Improve noise rejection when that is what stops the contact. Choose receiver performance for the signals it must handle. Then decide how much transmitter power is worth having.

Not anti-power. Not anti-radio. Not anti-EFHW. Not anti-compromise. Just engineering. The black box is one part of the station. That is Dave’s strongest point—and the one worth acting on.

Questions, antenna-factor records or height trials to share? Contact RF.Guru.