Follow-up Q19: Tesla's "stout copper bars" (the "hairpin circuit")

Date: 2026-10-08. Role: STEP F follow-up investigator. Web time: about 5 minutes (Chrome, date +%s 1791507868 to 1791508168). Stopped early because the question was answered as far as the online record allows; the open items are listed at the end.

Question (Q19)

Tesla's 'stout copper bar' experiments, which newer researchers call the 'Tesla hairpin' circuit: lamps short-circuited by a thick bent bar light from a spark discharge (1891 AIEE lecture, Inventions, Researches and Writings pp.194-195, Figs. 128-129), and lamps of different voltages burn properly in parallel on two stout bars (On Light, 1893, pp.36-37, Fig. 19a). Experimenters have lit nightlights and larger bulbs, even bulbs submerged in water. Is this 'adaptation to the load' (compare Moray's 'the draft of the current is adapted to the load')? Who named it the hairpin (Lindemann?) and what do modern replications measure? Does the underwater result survive in salt water, where mains AC would short?

What was searched

Saved pages (evidence trail, not for publication): research/pages/q19-*.html and .txt (Waveguide articles, EEVblog thread, Manning Substack, Atlantis Rising No. 77 text). Saved file: research/files/q19-ar77/AtlantisRising77.pdf (Wayback copy, 10.9 MB).

What Tesla wrote (the inventor's own words)

• 1891 AIEE lecture [read: IRW printed pp.194-195, Figs. 128-129; same text in _text/tesla-1891-05-20-lecture, Fig. 32]: "if a thick copper bar be bent, as indicated in Fig. 128, and shunted by ordinary incandescent lamps, then, by passing the discharge between the knobs, the lamps may be brought to incandescence although they are short-circuited." With a large induction coil "it is easy to obtain nodes on the bar", which are "simply maxima and minima of potentials along the bar", which he attributes to "the irregularity of the arc". "The thicker the copper bar ... the better". With straight-filament lamps and "smaller frequencies or smaller impulsive impedances" the filament can be brought "to any desired degree of incandescence"; with higher impedance the gas in the bulb glows and the filament stays dark.
• T. C. Martin's summary in IRW [read: IRW printed p.127]: "the impedance of a copper bar is so great that a potential difference of several hundred volts can be maintained between two points of a short and thick bar", enough to keep an ordinary lamp at full candle power with its terminals "no more than a few inches apart".
• On Light, 1893 [read: On Light pp.36-37, Figs. 19a-c]: bars B and B1 joined at the top by a low-voltage lamp, then a 50-volt lamp, a 100-volt lamp lower, and an exhausted tube lower still: "By carefully determining the positions of these devices it was found practicable to maintain them all at their proper illuminating power. Yet they were all connected in multiple arc to the two stout copper bars and required widely different pressures. This experiment requires of course some time for adjustment". In Fig. 19b the 50-volt lamp burns and the 100-volt lamp is dark; joining the bars with a thick cross bar B2 (Fig. 19c) reverses it. Tesla's own reading: "should not be considered to be due exactly to frequency but rather to the time rate of change".
• 1898 electro-therapy paper [web: quoted in Kraakman 2017, https://waveguide.blog/brief-history-tesla-hairpin-circuit-stout-copper-bars/, citing tfcbooks 1898-11-17]: the same principle as a variable-voltage tap: "By varying the distance between these contacts, any difference of potential, from a few volts to many thousands, is readily obtained". Tesla calls high-frequency currents ones that "pass with difficulty through stout conductors of high self-induction".

Tesla attributes the effects to self-induction (impedance) of the bar under a fast discharge. He does not invoke a special kind of energy in these passages.

Moray's phrase, for comparison

[read: Morray-TH followup Q16 note, citing Radiant Energy pamphlet ("1926" file) p.19]: "It makes no difference whether one fifty-watt lamp is used or whether 50 are connected to the machine, the draft of current is adapted to the load." Moray describes a source that delivers whatever the load draws, with no retuning.

Who named it the "hairpin"

• Karl Palsness, December 2008. The earliest use found is his Energetic Forum thread "Tesla's Stout copper bars 'Hairpin Circuit'", first post 27 Dec 2008 (copper bars 3/8 in x 4 ft, 10 kV transformer, homemade spark gap) [web: https://www.energeticforum.com/forum/energetic-forum-discussion/renewable-energy/3546-tesla-s-stout-copper-bars-hairpin-circuit].
• Atlantis Rising No. 77 (2009), Jeane Manning's column on the 2009 AEPC meeting in Maryland [web: Wayback copy of https://www.debunkingskeptics.com/AtlantisRising77.pdf, magazine p.58 area]: Palsness showed "what he calls the Tesla's Hairpin Circuit. Tesla only refers to it as the 'Two Stout Copper Bars circuit' which Karl views as not a very elegant name. So he prefers the Hairpin Circuit". This credits Palsness with choosing the name, but does not say he coined it.
• Peter Lindemann posted in the same thread on 27 Dec 2008, explaining why Palsness was not shocked (the body acts "more of a capacitor than a resistor" in a circuit "already in a 'short circuit' condition") [web: thread 3546, p.1]. Atlantis Rising quotes him. No source found shows Lindemann coining the name. On the record found, he endorsed and publicised it.
• Nick Kraakman (2017): "Nikola Tesla himself never used the term 'Hairpin circuit' ... Wherever the name Hairpin originated, it's catchy" [web: waveguide.blog history article]. He does not know the origin either.
• A Google search limited to pages dated before 2008 found no use of "hairpin circuit" for Tesla's bars. That search is weak, because forum and video dates before 2008 are poorly indexed.
• [memory: medium] "Hairpin" was already ordinary radio-engineering slang for a short U-shaped shorted parallel-line section (hairpin loops and tank circuits in VHF/UHF oscillators). The name is a natural description, so it may have been used independently before 2008.

What modern replications measure

• Palsness (2008-2009) [web: thread 3546]: qualitative. Lamps lit, the bars touched by hand, and in a video of 27 Feb 2009 a lamp dipped in "regular TAP water". The lamps were 120 V 100 W and 60 W halogens, the supply a 10 kV furnace-ignition transformer, and the gap magnetically quenched. He says "The Longitudinal wave does not see the water as a path to follow." He gives no voltage, current, frequency or power measurement. In a 23 Sep 2009 post (p.5) he says "we can not meter this energy".
• Kraakman (Waveguide, about 2018) [web: https://waveguide.blog/tesla-hairpin-circuit-stout-copper-bars-replication/]: the most instrumented replication found.
• Setup: a 15 kV 30 mA 50 Hz neon-sign transformer, two 40 kV 2000 pF doorknob capacitors (1000 pF in series), an air-quenched spark gap, and 22 mm copper bars (2 x 100 cm with a 20 cm shunt).
• Lamps: 12 V 10 W, 50 V 40 W, 110 V 40 W and an 8 W fluorescent tube. All four were lit at full brightness at once "after some careful adjustment", with the lowest-voltage lamp at the top as Tesla placed them; "If you change the order around, not all lamps will light up."
• No nodes: "it simply showed a maximum voltage, minimum current near the bottom, and maximum current, minimum voltage near the top of the bars, which is to be expected if you compare the Hairpin to a shorted transmission line."
• Measurements:
  • a frequency counter read 45-60 kHz;
  • an LCR meter gave the bar inductance as 182.23 µH, from which he calculated a resonance much higher than the counter reading (373 kHz with 1000 pF, my arithmetic);
  • a scope probe placed near the bar showed the 50 Hz wave with a high-frequency wave on top.
• He did not measure input power, lamp power, lamp voltage or current, or the water.
• A 34 kHz neon-sign transformer failed: the capacitors passed the input and the gap never fired.
• Liberty University poster, 2018 (Esteban Alvarez-Cespedes, undergraduate) [web: https://digitalcommons.liberty.edu/research_symp/2018/Posters/19/]. The Google snippet lists the parts: transformer, two high-voltage doorknob capacitors, spark gap, two copper pipes, copper shunt. The poster's content was not read: the PDF had no text layer in Chrome and returned 403 to curl.
• EEVblog forum, 2023 [web: https://www.eevblog.com/forum/projects/tesla-hairpin-circuit-curious-phenomenon/]. An engineer's reading: "varying cases of transmission line, transformer or inductive loop circuits". Points made:
• the spark gives wideband excitation;
• tungsten's hot-to-cold resistance ratio is about 8, so the effective circuit changes as the lamps warm;
• "It is doubtful that any of the lightbulbs are near nominal power";
• nerves do not respond above about a kHz, but RF can still heat tissue.

No measurements were posted. - Energetic Forum thread 11044 (2013): a member (Pedro73) proposed measuring whether the water heats or cools as a test [web: https://energeticforum.com/forum/energetic-forum-discussion/renewable-energy/11044-...]. No result was posted on the page read.

No replication found reports an energy balance (electrical input against lamp output) or a measured lamp voltage at each position.

Salt water

• Palsness's own argument (23 Sep 2009, thread 3546 p.5): "If this was skin effect you would not be able to put the bare wires into water and light a bulb...the salt water would short out the skin effect." He states this as an argument. The post does not describe a test in salt water with a stated salt concentration.
• Davy Oneness, reported by Jeane Manning (Substack, 22 Jun 2026) [web: https://jeanemanning.substack.com/p/the-courage-of-a-whistle-blower]: a photo caption describes "'Tesla Hairpin' experiments, lighting a fluorescent bulb in series with a 12-volt incandescent bulb. He's holding it under salt water with his bare hands without getting any electrical shock." This is one published photo report, from about 2016 by the article's timeline. No second source was found. It gives no salt concentration, does not say whether the lamp leads were bare in the water, and reports no brightness comparison with fresh water. Manning is an advocate writer, and the report is hers and Oneness's.
• A YouTube channel ("Efficient Energy Transformations") lists a video with "salt water, and a 200 W light bulb" in its title or description (Google snippet only, not viewed).
• No controlled comparison (fresh water against salt water, lamp brightness or power measured) was found.

Reasoning (inferred; for the reviewers to weigh)

1. What the bars are, in textbook terms. The capacitors discharge through the gap into a loop made of the two bars and the top shunt. A fast discharge makes the current change rapidly, and a bar of inductance L between two clamps then has a voltage V = L dI/dt (or V = ωLI for a sinusoid of angular frequency ω) across it, even though its resistance is tiny. That is Tesla's "impedance" explanation in modern notation.

Worked numbers [calc]: - A pair of 22 mm bars 20 cm apart has about 1.2 µH per metre: L ≈ (μ0/π) ln(d/r) per metre, with μ0 = 4π × 10⁻⁷ H/m, d = 0.2 m and r = 0.011 m. - At 5 MHz, a plausible ringing frequency for a few µH with a 1 nF capacitor, that is ωL ≈ 36 Ω per metre. - The peak discharge current from 1 nF charged to 15 kV through about 2 µH is I ≈ V√(C/L) ≈ 15,000 × √(10⁻⁹ / 2×10⁻⁶) ≈ 340 A. - So tens to hundreds of volts between clamps a short distance apart, as Tesla and Martin report ("several hundred volts").

Kraakman's 182 µH reading is about 100 times what this geometry gives. It may be a meter-range error, and it is unresolved.

2. "Adaptation to the load"? Two readings.

• (a) No, not in Moray's sense. On the bars, each lamp sees the voltage of the stretch of bar (and shunt) it bridges, and that voltage depends on where it is clamped. Tesla matched each lamp to its position: 50 V and 100 V lamps at different heights, "carefully determining the positions", "some time for adjustment". Kraakman confirms the order matters. This is matching by tap position, an inductive voltage divider or a shorted transmission line. Tesla's 1898 paper uses the same arrangement as an adjustable tap giving "from a few volts to many thousands".
• (b) Yes, in a weaker sense. A source that adapts to its load is one whose output voltage stays put as load is added. Moray says this ("one fifty-watt lamp ... or 50") of his device, with no retuning. Tesla's bars are the opposite: a soft source (internal impedance of tens of ohms) whose voltage distribution shifts as lamps are added, so the positions have to be re-adjusted. The resemblance is in the words "adapted to the load". The physics differs: Tesla's bars are fitted to each load; Moray claims a source that is indifferent to the load.

Reading (b) has one real point of contact with Moray's Q16 discussion: the bars are a resonant, impulse-excited circuit whose stored energy is shared among the loads. A load placed at a matching point extracts power efficiently, which is ordinary impedance matching. The record found contains no measurement showing more power to the lamps than the transformer supplies.

3. Underwater, and salt water. - Tap water. Tap water conducts weakly: about 0.005-0.05 S/m [memory: high]. Between bare leads about 1 cm apart, each about 1 cm² in area, that is a resistance of the order of 2-20 kΩ. A hot 120 V 100 W lamp is about 144 Ω. So the water carries only a small part of the current, at 50 Hz mains or at radio frequency alike. - Mains in tap water. A lamp on 120 V mains with bare leads in tap water also stays lit; it is dangerous, not dark. The tap-water demonstration therefore does not separate "longitudinal" energy from ordinary current. - Sea water. Sea water is about 5 S/m [memory: high]. The same leads give about 20 Ω, now comparable to or below the lamp. Ionic conduction holds well into the MHz range. The skin depth in sea water at 1 MHz is about 0.2 m (δ = 1/√(π f μ0 σ)), far larger than the gap between the leads, so the skin effect gives no protection. - Mainstream prediction for salt water on the bars. The bars are a soft source, so a 20 Ω water path in parallel would roughly halve the lamp voltage and cut its power several-fold. The lamp dims or goes out, unless the leads are insulated or the lamp is over-driven. - Mains in salt water. The owner's premise needs one correction. On mains (a stiff voltage source of well under 1 Ω), salt water does not darken the lamp. It draws a large current through the water: heat, gas, a tripped breaker, and a shock hazard. So "mains would short" is right about the current but not about the lamp. The contrast to test is lamp brightness, and the fresh/salt difference should be larger on the bars than on mains. - What a test needs. Oneness's salt-water photo would bear on this only if the leads were bare and the brightness was compared. The report says neither.

4. Not being shocked. This is not part of the question, but it is in the same record: Palsness and others hold the bars; Lindemann says the body is "more of a capacitor". The textbook reading is that the voltage between two hands on the same bar is small (the bar's ωL over the hand spacing), and that nerves do not respond to currents above about 10 kHz [memory: high]. Heating is still possible (EEVblog; Kraakman's electrosurgery warning).

Open (what would settle it)

• The salt-water question. One controlled bench test would settle it. The same lamp, bare leads at fixed spacing, is immersed in distilled water, tap water and 3.5% NaCl, first on the bars and then on mains through an isolation transformer and a current limit. Measure lamp brightness (a lux meter or photodiode) and the RMS current in the lamp lead (an RF current probe) in each case. No such test was found.
• "Adaptation". A test would add lamps one at a time at a fixed clamp position and record the voltage on each lamp with a high-voltage RF probe, against the input power at the transformer primary. Tesla's text predicts that positions must be re-adjusted; Moray's phrase would predict no change.
• The name. Pre-2008 Usenet, KeelyNet and Tesla Coil Mailing List archives, and Palsness's own YouTube descriptions (2008), would show whether "hairpin" for these bars predates his thread.
• The Liberty 2018 poster. Its measurements, if any, are unread. Obtain the PDF, which needs a text-capable viewer or OCR.
• Oneness's experiment. Its details (concentration, whether the leads were bare) are not in the Substack post.