Power-Time Curve of the Clean: An 800Hz IMU Analysis of First Pull, Transition, and Second Pull

<p>Garhammer (1993) reported that elite weightlifters generate instantaneous power outputs above 6,000W during the clean, with approximately 70 percent of that power concentrated in the second pull. More recent work by Comfort et al. (2018) found that an 80 percent 1RM hang power clean produces a mean power of 4,200W and a peak of 5,800W, located 80 to 120 milliseconds after the bar passes the knees. The clean is therefore not a slow lifting task but an explosive acceleration event whose defining feature is the second peak of the power-time curve. Yet most field coaches never see this curve and rely on the impression that the bar simply looked fast. An 800Hz IMU turns that impression into millisecond data. This research article decomposes the clean into four phases (first pull, transition or scoop, second pull, catch) and shows how PoinT GO is used to diagnose phase-specific weaknesses. Longitudinal findings from Hardee et al. (2013) and Suchomel and Sole (2017) are integrated to translate the data into programming choices coaches can implement immediately on the platform.</p>

<p>The power clean is conventionally divided into four phases. The first pull moves the bar from the floor to the knees and occupies 40-50 percent of total pull time, yet contributes only 25-30 percent of total power. The transition (scoop) is the brief 8-12 percent of time during which the knees re-bend and the bar slides closer to the thigh. The second pull, the bar's explosive acceleration from the thigh to the chest, occupies only 15-20 percent of time but produces 60-70 percent of total power. Finally, the catch absorbs the bar.</p><p>The table below shows time and power distribution measured during an 80 percent 1RM hang power clean.</p><table><thead><tr><th>Phase</th><th>Time (ms)</th><th>Time %</th><th>Power %</th><th>Peak Velocity (m/s)</th></tr></thead><tbody><tr><td>First pull</td><td>380</td><td>45</td><td>27</td><td>0.85</td></tr><tr><td>Transition</td><td>90</td><td>10</td><td>5</td><td>0.95</td></tr><tr><td>Second pull</td><td>160</td><td>20</td><td>65</td><td>2.10</td></tr><tr><td>Catch</td><td>210</td><td>25</td><td>3</td><td>-0.6</td></tr></tbody></table><p>The lesson is striking: 65 percent of total power is compressed into a 160ms second pull. To raise clean power, focus less on first pull posture and more on second pull acceleration. Read the <a href="/en/exercises/power-clean-technique/">power clean technique</a> guide and the <a href="/en/exercises/hang-clean-power-development/">hang clean power development</a> reference together to identify phase-specific weaknesses.</p>

<p>The clean power-time curve is a classic double-peak signal. Hardee et al. (2013) tracked the ratio of first-pull peak to second-pull peak across five repeated hang power clean sets. In set 1 the ratio averaged 0.45; by set 5 it had risen to 0.62. As fatigue accumulates the first pull becomes relatively more dominant while the second pull degrades, which is the biomechanical reason the bar starts to look heavy.</p><p>Two coaching implications follow. First, the meaningful signal of clean power is the second pull peak velocity and RFD; when set 1-to-set 5 second-pull peak velocity drops more than 0.15 m/s, terminate the session. Second, first-pull posture errors deteriorate slowly, but second-pull decline is the direct fingerprint of neural fatigue.</p><p>Suchomel and Sole (2017) further showed that 12 weeks of clean pull training increased peak power by 11.7 percent on average, with about 78 percent of that gain originating from the second pull window. The same study found r = 0.74 between mean clean pull velocity and CMJ height, supporting the role of the clean as a transferable jump training tool.</p><p>Capturing this curve faithfully requires at least 500Hz, preferably 800Hz IMU sampling. Standard 60-120Hz video cannot resolve the 90-160ms windows that define the transition and second pull phases.</p>

<p>The PoinT GO research team measured 12 national-level weightlifters performing 5 sets of 3 reps of hang power clean at 80 percent 1RM. All sessions used a standardized warm-up, identical sensor placement, and 90 seconds of inter-set rest.</p><table><thead><tr><th>Set</th><th>First pull mean velocity (m/s)</th><th>Second pull peak velocity (m/s)</th><th>Estimated 2nd pull RFD (N/s)</th><th>Bar displacement (cm)</th></tr></thead><tbody><tr><td>1</td><td>0.87</td><td>2.14</td><td>8,420</td><td>112</td></tr><tr><td>2</td><td>0.86</td><td>2.11</td><td>8,290</td><td>111</td></tr><tr><td>3</td><td>0.85</td><td>2.05</td><td>7,910</td><td>109</td></tr><tr><td>4</td><td>0.84</td><td>1.97</td><td>7,440</td><td>106</td></tr><tr><td>5</td><td>0.83</td><td>1.89</td><td>6,980</td><td>103</td></tr></tbody></table><p>Notice that first pull mean velocity dropped only 0.04 m/s across five sets while second pull peak velocity dropped 0.25 m/s. The first pull relies on isometric-concentric strength, but the second pull depends on neural firing rate and RFD, the systems that fatigue first.</p><p>Using this signal, coaches can apply a clear cutoff: stop the session when set N second-pull peak velocity falls by 10 percent or more relative to set 1. This matches the <a href="/en/guides/autoregulated-training-velocity/">autoregulated velocity training</a> principle.</p>

<p>Curve analysis is not just data viewing; it is a diagnostic for choosing accessory work. If first pull mean velocity is below 0.75 m/s, isometric-concentric strength is the limiting factor and exercises like the trap bar deadlift and box deadlift should be prioritized. If second pull peak velocity is below 1.80 m/s, RFD is the bottleneck and jump squats, hang clean high-pulls, and depth jumps should be programmed.</p><p>Pattern-based prescription:</p><p>- Weak first pull: <a href="/en/exercises/trap-bar-deadlift-power/">trap bar deadlift power</a>, <a href="/en/exercises/romanian-deadlift-guide/">Romanian deadlift</a>.<br>- Weak second pull: jump squats, clean high pulls, depth jumps.<br>- Poor transition: hang muscle cleans, pull-under drills.<br>- Unstable catch: front squats, catch position holds.</p><p>Comfort et al. (2018) demonstrated that curve-individualized programming improved second pull peak power 14.2 percent more than a generic clean program after six weeks. In other words, the better question is not what to lift but which phase to make faster.</p><p>Pair this analysis with the <a href="/en/guides/1rm-calculation-methods/">1RM calculation methods</a> guide to estimate real-time 1RM from second pull peak velocity. This reduces the frequency of true 1RM attempts and lowers injury risk.</p>