Tangerine Tiger: The Engineering Question

The Tangerine Tiger series follows Russ Sadler as he removes the conventional glass CO₂ tube and high-voltage power supply from a low-cost 500 × 300 mm Chinese laser and replaces them with a sealed RF CO₂ laser source. The point of the project was not simply to complete the conversion. It was to find out, experimentally, what the RF source genuinely improves, what it does not improve, and which apparent advantages are really caused by optics, mechanics, airflow, control strategy or material behaviour.

Across 21 videos the project evolves from a mechanical and electrical retrofit into a much broader comparison of RF vs glass/DC CO₂ laser behaviour. Russ repeatedly proposes an explanation, tests it, and sometimes has to revise it. That progression is one of the most useful parts of the series, so this page preserves the distinction between what was measured, what was observed, what was derived, and what remained a working hypothesis.

Tangerine tiger rf co2 laser conversion — The Tangerine Tiger RF CO₂ conversion project

Executive Verdict

The completed series does not show that RF CO₂ is simply better than a glass/DC tube. It shows something more specific:

RF provides a genuine rapid-modulation advantage. That matters most in high-speed engraving, timing-sensitive marking and controlled 3D relief work.
RF did not demonstrate a meaningful inherent cutting advantage at matched optical output. In the best direct comparison, the glass system was marginally faster in 5 mm extruded acrylic.
Beam delivery matters. The tested RF source had a small exit beam and high published divergence, so a beam expander became an important part of the optical design.
Mechanical dynamics limit useful engraving speed. Very high commanded scan speeds do not automatically reduce cycle time because acceleration and overscan become dominant.
Process engineering can outweigh source technology. In one plywood test, changing nozzle standoff from about 7 mm to 4 mm improved cutting speed from roughly 12 mm/s to 22 mm/s without changing the laser source.
Photographic grayscale and 3D grayscale are different problems. Continuous-tone photo engraving on wood proved difficult because the material response was nonlinear; depth-map engraving produced one of the strongest RF results in the entire series.

What the Project Actually Established

1. The RF conversion worked — but it was not plug-and-play

The conversion required mechanical rework, a new mounting arrangement, 48 V power infrastructure, wiring changes, controller configuration, beam-height correction, alignment and additional beam-conditioning optics. The project therefore provides a useful reality check for anyone imagining an RF source as a simple replacement for a glass tube.

2. RF beam divergence was a real integration issue

The reference specification discussed in the series gives the RF source an exit beam of roughly 1.8 mm and a divergence of about 7.5 mrad. Russ also observed the untreated beam growing to around 10 mm over the long beam path in this machine. The exact 7.5 mrad figure is therefore a published/reference specification, not a direct divergence measurement made by Russ, but the observed beam growth clearly supported the need for beam expansion and better collimation.

Russ tested both 3× and 2× expansion and ultimately found the 2× arrangement the more useful compromise in this particular machine. Some of the Gaussian-beam explanations used in the videos are simplified; the robust experimental finding is that the expander increased beam diameter and reduced divergence enough to improve the fit through the machine optics.

3. The installed RF source appeared to deliver more optical power than its nominal label suggested

Russ’s homemade calorimetric tests repeatedly indicated roughly 35–38 W of optical output, even though the source is described at different points in the series as 20 W or 30 W. Those measurements are useful for relative testing but they are not laboratory-grade calibration. The safest conclusion is that the source delivered substantially more than the nominal numbers Russ was initially using, while the exact rated specification remains ambiguous within the series.

4. Controller-side PWM behaviour was directly observed

Oscilloscope tests show the controller changing duty cycle while the logic-high level remains approximately constant, and changing the layer frequency changes the carrier period. This is strong evidence for the control waveform being sent to the RF source. It does not by itself prove that the instantaneous optical output perfectly mirrors an ideal electrical square wave.

Episode 21 adds an important final correction: controller percentage is not the same thing as percentage of measured optical watts. Russ’s own calibration indicates roughly 25 W at 50% command versus around 38 W near full command. The material therefore responds to the actual optical energy delivered, not to the number shown on the controller.

5. Spatial PWM behaviour was validated experimentally

At 1,000 mm/s and 500 Hz, Russ produced a spatial modulation period of about 2 mm, corresponding to roughly 1 mm on and 1 mm off at 50% duty. Reducing speed reduced the spatial spacing accordingly. This is one of the clearest demonstrations in the series that the RF system can respond rapidly enough for useful high-speed marking behaviour.

6. High-speed engraving is where RF makes the strongest practical case

The machine was pushed into the 1,000–1,400 mm/s region, with very high acceleration settings. The tests show that RF can support rapid modulation at those speeds, but they also show why headline speed is not the same as productive throughput. On shorter scan jobs, extra acceleration and deceleration distance increases overscan, and Russ found little or no cycle-time benefit when moving from around 1,000 mm/s to 1,400 mm/s on some examples.

The useful conclusion is not “RF equals 1,400 mm/s engraving”. It is that RF removes one temporal limitation, after which machine acceleration, available optical power, overscan, spot size and the material response become the next constraints.

7. Photographic grayscale exposed a material problem, not simply a switching problem

Episode 19 shows why fast modulation alone does not guarantee smooth grayscale photography. At 254 ppi and 500 mm/s, pixel timing is not the main obstacle. The difficulty is that wood does not convert digital power commands into a stable, linear set of visible tones. As exposure rises, the process moves through colour change, pyrolysis, charring and then material removal.

This means that 256 grayscale values in an image do not translate into 256 repeatable shades on the material. The material transfer function is nonlinear and depends on species, density, grain, moisture, resin, focus, speed and thermal history.

8. 3D grayscale engraving is one of the clearest positive RF results

Episode 20 uses grayscale as a height map rather than as a photographic tone map. That changes the problem completely: the required output is variable material-removal depth rather than a precise visible shade. Using a long-focus lens and repeated passes, Russ produces a convincing 3D relief result with relatively little charring. The demonstrated four-pass job takes roughly 12–14 minutes.

This is a much stronger application of RF modulation because the process is tolerant of some nonlinear material behaviour. The output variable is depth, which can be accumulated over successive passes.

9. RF did not show a special cutting advantage at matched optical output

Episode 18 is the strongest direct cutting comparison in the series. Russ adjusted the glass system to approximately match the RF system’s indicated optical output, used the same 2.5-inch plano-convex lens and cut the same 5 mm extruded acrylic. The glass machine cut through at about 10 mm/s; the RF machine managed about 9 mm/s.

The difference is small and the two systems are not a laboratory-matched pair, so it should not be turned into a universal efficiency claim. What the test does show is that there was no dramatic RF cutting advantage at similar measured watts.

10. Nozzle geometry and air assist can outweigh source choice

In Episode 16, reducing nozzle-to-work distance from roughly 7 mm to 4 mm while retaining focus improved 3 mm plywood cutting from about 12 mm/s to about 22 mm/s in Russ’s setup. That is a machine-specific result, but it is an important one: airflow delivery, kerf clearing and nozzle geometry can have a much larger effect than changing the laser source.

11. PWM frequency did not reveal hidden cutting power

Episode 21 closes one of the series’ most interesting loops. Extremely low PWM frequencies produced deeper isolated marks and briefly suggested a possible start-up overshoot or “magic” cutting regime. Russ then tested the idea more carefully. The individual pulses could be deeper, but the effect disappeared as a useful advantage once the marks had to join into a practical continuous cut.

At 200 mm/s and 50% command, 25 kHz and 1 kHz produced about 0.95 mm and 0.94 mm depth respectively in poplar. Near full command, the same comparison produced roughly 1.47 mm and 1.434 mm. Within this useful cutting/engraving range, frequency alone had very little effect on depth.

What the Series Does Not Prove

It does not establish a universal speed at which all glass/DC CO₂ systems stop forming useful dots. Any threshold depends on the particular tube, PSU, controller, optics, speed and material.
It does not directly measure the RF source’s optical rise time, fall time or any transient start-up overshoot. Those would require suitable high-speed optical instrumentation.
It does not prove that RF output is a perfect fixed-height optical pulse whose power never changes during the “on” interval.
It does not prove that fourth-corner alignment faults are normally optical rather than mechanical. Episode 17 demonstrates one effective long-path alignment refinement on this machine.
It does not establish universal optical-loss percentages for RF beam expanders or combiners. Russ’s measurements are approximate and setup-specific.
It does not demonstrate that photographic grayscale is impossible on RF systems. It demonstrates that the tested wood/material/process combination did not produce a reliable continuous-tone result.

How to Read the Evidence

Throughout the rebuilt Tangerine Tiger pages, conclusions are presented using a simple evidence hierarchy:

Measured — a value was directly measured or instrumented.
Observed — the behaviour was clearly visible in the experiment.
Calculated — arithmetic or geometry derived from stated values.
Derived — a reasonable engineering conclusion from the evidence.
Working hypothesis — Russ’s explanation at that point in the investigation, awaiting further testing.
Correction / later finding — a later experiment changes the interpretation of an earlier result.

The 21-Part Tangerine Tiger Series

The videos are best understood as five phases of one engineering investigation rather than 21 unrelated demonstrations.

Phase 1 — Machine rebuild and RF integration

01. Tangerine Tiger Begins: Assessing the Chinese Laser Platform — donor-machine inspection and project baseline.

02. Rebuilding the Motion System Before the RF Upgrade — mechanical platform, mirror support and motion-system work.

03. Preparing the Chassis for the RF Conversion — enclosure, fume isolation and ancillary-system preparation.

04. RF Laser Integration: Wiring, Power and Machine Rework — 48 V power, wiring, extraction and air-assist integration.

05. First RF Laser Pulse: Alignment, Beam Growth and a False Start — first light, alignment and the discovery of substantial beam growth.

Phase 2 — Beam delivery, divergence and PWM control

06. RF Beam Divergence and Why a Beam Expander Matters — reference beam specifications and 3× expander testing.

07. RF Beam Expansion and Power: 2× vs 3× Optics — comparative expander testing and approximate calorimetry.

08. RF PWM Control: What the Oscilloscope Actually Shows — duty cycle, frequency and controller-side waveform evidence.

09. RF PWM, Speed and Spatial Pulse Spacing — spatial pulse-period tests and correction of earlier assumptions.

Phase 3 — High-speed engraving and image formation

10. High-Speed RF Engraving: When ‘Hyperdrive’ Stops Helping — acceleration, overscan and real cycle time.

11. RF Engraving Timing: Pixels, PWM and Dot Formation — pixel timing versus carrier timing.

12. RF Photo Engraving: Timing, Pixel Pitch and a Better Workaround — revised interpretation and improved dither results.

13. RF vs Glass CO₂: A Provisional Engineering Comparison — an intentionally early comparison later refined by more testing.

14. Photo Engraving Resolution: PPI vs Physical Spot Size — why digital resolution and physical mark resolution are not the same thing.

15. High-Speed RF Engraving: Speed vs Real Throughput — practical material engraving and the limits of headline scan speed.

Phase 4 — Cutting and alignment

16. RF Laser Cutting: What Can 20 Watts Actually Do? — plywood cutting, lens choice, beam expansion and nozzle standoff.

17. Fixing the Fourth-Corner Alignment Error — long-path verification and residual angular error.

18. RF vs Glass CO₂ Cutting at Matched Optical Power — the strongest direct RF/glass cutting comparison in the series.

Phase 5 — Grayscale, 3D engraving and the final PWM tests

19. Grayscale Photo Engraving: RF Speed Meets Material Limits — fast modulation meets nonlinear wood behaviour.

20. 3D Grayscale Engraving: Where the RF Laser Delivers — successful variable-depth relief engraving.

21. RF Laser PWM and Cutting: What the Tests Actually Show — duty cycle, real optical output, low-frequency anomalies and the final correction.

Important Technical Corrections Across the Series

The experiments remain valuable even where the original spoken explanation is incomplete. The rebuilt editorial notes on each episode preserve the historical transcript while adding modern technical context.

Beam optics: a beam expander does not simply remove “rogue rays” or flatten a Gaussian beam. Its useful function here is to increase beam diameter and reduce divergence.
PMMA/acrylic: laser cutting is not well described as the material behaving like ice, water and steam. PMMA absorbs 10.6 µm radiation strongly and undergoes heating, melting and significant thermal decomposition/depolymerisation.
Wood: colour and material removal involve absorption, heating, pyrolysis, volatile release, charring, oxidation and ablation rather than a single fixed-temperature transition.
PWM: average delivered energy matters in thermal processing. Peak output, pulse duration and repetition rate can also matter, but “full instantaneous power during each on pulse” does not make average power irrelevant.
Fourth-corner alignment: the long-diagonal method is a valuable refinement, but fourth-corner errors can still have mechanical or structural causes.

RF vs Glass CO₂: Practical Buying Decision

RF makes most sense when:

The work genuinely benefits from rapid modulation.
Fine, fast raster or vector marking is commercially important.
Controlled depth modulation or 3D relief engraving is part of the workflow.
The host machine has sufficient stiffness, acceleration, beam-path quality and thermal management to exploit the source.
The higher capital cost is justified by the application rather than by the expectation of more cutting watts.

A glass/DC system remains the stronger value proposition when:

The machine is used mainly for cutting acrylic, plywood and similar sheet materials.
Very high engraving speed is not commercially important.
The budget would be better spent on air assist, optics, alignment, cooling or mechanical improvements.
Low replacement cost and simple servicing matter more than rapid modulation.

Engineering FAQ

Does an RF CO₂ laser cut better than a glass tube at the same wattage?

Not in the Tangerine Tiger tests. At approximately matched indicated optical output, the glass system cut 5 mm extruded acrylic at about 10 mm/s and the RF system at about 9 mm/s. The result is close enough that it should be read as “no special RF cutting advantage”, not as a universal 10% superiority claim for glass tubes.

What is the clearest RF advantage demonstrated in the series?

Rapid modulation. It becomes most valuable in high-speed engraving and in controlled depth modulation. The successful 3D grayscale engraving in Episode 20 is one of the strongest positive RF demonstrations.

Does PWM frequency change cutting power?

Within the practical 1–25 kHz range tested in Episode 21, changing frequency alone had very little effect on measured depth. Extremely low frequencies produced deeper isolated marks, but that did not translate into a useful continuous-cut power multiplier.

Is controller percentage the same as optical power percentage?

No. In this setup, Russ measured roughly 25 W at 50% command and about 38 W near full command. The relationship was not linear, so percentage settings should not automatically be interpreted as optical watts.

Why did the RF source need a beam expander?

The RF source had a small exit beam and relatively high divergence. Over the long beam path in the machine, the untreated beam grew enough to threaten mirror and aperture utilisation. Expanding the beam reduced divergence and produced a more usable beam path.

Why did 3D grayscale work better than photographic grayscale?

Photographic grayscale requires stable visible tones. Wood does not respond linearly enough for that to be easy. A 3D height map only needs different levels of material removal, so repeated passes can accumulate depth even when the visible colour response is nonlinear.

Final Takeaway

The Tangerine Tiger project is most useful because it replaces a simple “RF is better” narrative with a system-level engineering picture. RF solves a genuine temporal-control problem and can deliver excellent high-speed engraving and 3D depth-modulation performance. It also adds cost, beam-delivery requirements, thermal load and integration complexity.

For ordinary cutting, the series does not show a meaningful advantage from the RF source itself at matched optical output. In several tests, machine dynamics, nozzle geometry, air assist, optics and alignment mattered as much as — or more than — the choice of laser source.

The strongest reason to choose RF is therefore not “more cutting power”. It is rapid controllability where the application can actually exploit it.

RF vs DC CO₂ Laser Upgrade: What the Tangerine Tiger Experiments Actually Show