SUNGUIDE IPC — Experimental Data Dashboard

30 panels · 50V · 20A · String voltage 1,500V · 15 IPC units in series (2 panels per IPC) · Contamination: 1A on 1 panel

Normal Output (Reference)

30,000W

50V × 30 panels × 20A = 1,500V

Conventional Series (Contaminated)

1,500W

Efficiency 5.0% ↓95.0%

SUNGUIDE / IPC 15 units

29,000W

Efficiency 96.7% ↓3.3%

IPC Functions

S/P · RS · BPD

Series/Parallel · Rapid Shutdown · Bypass

Contamination Current Variable Simulator

String voltage 1,500V = 30 panels × 50V | 15 IPC units in series (1 IPC per 2 panels)

Contamination current1.0 A

Panel voltage50 V

No. of panels30 장

Conventional Series (KCL — minimum current tracking)

P = 50V × 30 × 1A = 1,500 W (5.0%)

SUNGUIDE / 15 IPC in series (1 IPC per 2 panels — parallel switching)

P = 50V × 29 × 20A = 29,000 W (96.7%)

Recovered power

+27,500 W recovered (91.7% recovered)

Power Output Comparison

NormalConventional SeriesSUNGUIDE

Electrical Physics Law Verification — 30 panels / 50V / 20A / 15 IPC in series

✓

① KCL Series Weakness (Kirchhoff's Current Law)
Law: In a series circuit, the current flowing through any closed loop is identical and determined by the component generating the lowest current.
Formula: I_string = min(I₁, I₂, ..., I₃₀)
30-panel application: 1 of 30 panels contaminated → I_contam = 1A → I_string = 1A
Power loss: P = 50V × 30 panels × 1A = 1,500W (efficiency 5.0% vs. normal 30,000W)
Why? Series-connected panels have only one current path. If the contaminated panel passes only 1A, the remaining 29 panels cannot physically exceed 1A even though they can generate 20A. The potential power of 29 normal panels (29 × 50V × 19A = 27,550W) is completely wasted.

✓

② KVL Parallel Voltage Sharing (Kirchhoff's Voltage Law)
Law: Components connected in parallel share the same terminal voltage. When two components with different voltages are connected in parallel, current redistributes toward the lower voltage.
Formula: V_parallel = min(V_PV1, V_PV2) = min(50V, 50V) = 50V
30-panel application: IPC switches contaminated panel (PV_k) and adjacent normal panel (PV_{k+1}) to parallel → common voltage 50V maintained
Why is 50V maintained? Even a contaminated panel has nearly the same open-circuit voltage (V_oc) as a normal panel. Contamination reduces Isc (short-circuit current) but barely reduces Voc, so the common voltage in parallel stays at the normal panel voltage level (50V).

✓

③ KCL Parallel Current Summation (Kirchhoff's Current Law — Parallel)
Law: At a node (branch point) in a parallel circuit, the sum of incoming currents = sum of outgoing currents. Currents of parallel components are summed.
Formula: I_pair = I_PV_normal + I_PV_contaminated = 20A + 1A = 21A
2-panel → 1-panel equivalent: (50V, 21A) equivalent single panel joins the string
Physical meaning: The contaminated panel's residual current (1A) is not discarded but added to the normal panel's current. IPC does not 'bypass' the contaminated panel but 'absorbs' it. This is the fundamental difference between bypass diode (1,000W loss) and SUNGUIDE/IPC (50W loss).

✓

④ Norton Equivalent Theorem — 2 panels → 1 panel equivalent transform
Law: Any linear circuit network can be equivalently replaced by a single current source (I_N) and a single parallel impedance (Z_N).
30-panel application: Parallel-switched pair (PV_k ‖ PV_{k+1}) → Norton equivalent: I_N = 21A, V_N = 50V → single panel equivalent
String reconfiguration: 30 panels series → after IPC parallel switching, 29-panel equivalent series string
String voltage: 29 × 50V = 1,450V (only 50V = 3.3% reduction vs. normal 1,500V)
Why does voltage decrease by only 1 unit? When 2 panels go parallel, voltage is calculated on a 1-panel basis (50V), and the series count drops from 30 to 29. Since all 15 IPC units are in series, 1 IPC operating = only 1-step voltage drop across the entire string.

⑤ Final Power Calculation — Joule Power Law (P = V × I)
Law: Power P = V × I. In a DC circuit, power is the product of voltage and current.
String current determination: I_string = min(I_pair, I_normal×29) = min(21A, 20A) = 20A (KCL re-applied)
SUNGUIDE/IPC output: P = 1,450V × 20A = 29,000W (efficiency 96.7%)
Conventional series output: P = 1,500V × 1A = 1,500W (efficiency 5.0%)
Loss comparison: SUNGUIDE/IPC loss = 1,000W (3.3%) vs. conventional series loss = 28,500W (95.0%)
Significance of 15 IPC in-series architecture: Each IPC manages 2 panels and is inserted in series into the string. Under normal conditions, only a pure conduction path (MOSFET R_DS(on) ≈ 2mΩ) exists, making losses virtually zero. Only upon contamination detection does the relevant IPC perform parallel switching.

Output Graph by Current Range (30 panels / 50V basis)

Conventional SeriesSUNGUIDE / IPC

How to read this graph:
· Horizontal axis: Contamination panel current (0A = fully blocked ~ 20A = fully normal)
· Red line (conventional): proportional to contamination current. 1A contamination → 1,500W (5%)
· Green line (SUNGUIDE): maintains 29,000W flat across contamination current 0~20A
· Area difference between lines = power recovered by IPC

Power Output Experimental Data by Contamination Current (50V, 30 panels, normal current 20A)

Contam. Current (A)	Conv. Series (W)	Conv. Efficiency	SUNGUIDE (W)	SUNGUIDE Efficiency	Recovered power(W)	Improvement

Annual Loss Estimate (1,500 kWh/kWp basis)

Conventional Series Annual LossSUNGUIDE Annual Loss

IEEE Series-Parallel Topology Classification — SUNGUIDE Academic Position

IPOP

Input-Parallel Output-Parallel

High current output

IPOS ★

Input-Parallel Output-Series

SUNGUIDE match

ISOP

Input-Series Output-Parallel

High→Low voltage conversion

ISOS

Input-Series Output-Series

High voltage I/O

Paper Fig.1(c) definition: An IPOS system sums current through parallel input connections and accumulates voltage through series output connections. The moment SUNGUIDE switches PV1 and PV2 to parallel, those 2 panels become exactly an IPOS module equivalent structure.

IEEE TPEL 2020, p.12383 — Section I, Fig.1(c) ↗

Paper Principles → SUNGUIDE Mapping Verification

✓

IPOS Current Sharing Law (ICS): Paper Section III states that achieving ICS automatically achieves OVS. SUNGUIDE's physical parallel switching directly implements ICS.

IEEE TPEL 2020, p.12388 — Section III ↗

✓

Auto-Balancing Mechanism: Section III-A explains that under common duty cycle control, an IPOS system automatically corrects output voltage imbalance. The convergence of parallel panels to 1-panel equivalent corresponds to this.

IEEE TPEL 2020, p.12388 — Section III-A ↗

✓

PV Sub-module MPP Current Mismatch Compensation [Ref.134]: Cites differential power processing (DPP) research for compensating MPP current mismatch in series PV sub-modules. SUNGUIDE achieves the same effect via hardware switching.

IEEE TPEL 2020, p.12400 — Ref.[134] ↗ | Bell & Pilawa-Podgurski, IEEE TPEL 2015 (original Ref.134) ↗

✓

Non-Isolated IPOS 2-Module Limit: Section VIII states that non-isolated converters can be applied to IPOS but are limited to 2 constituent modules. This exactly matches SUNGUIDE using 2 panels as the processing unit.

IEEE TPEL 2020, p.12396 — Section VIII ↗

LoRa MASTER/SLAVE = Distributed Modularized Control: Section II-D-c describes distributed modularized control where all modules share the same control circuit using only the IVS bus and common duty cycle bus. SUNGUIDE's LoRaWAN MASTER/SLAVE architecture is identical to this.

IEEE TPEL 2020, p.12387 — Section II-D-2-c, Fig.5(e) ↗

SUNGUIDE 6-Step Operation — IEEE Principle Cross-Verification

Step	SUNGUIDE Operation	Physical Law	IEEE Reference	Verified
① Series	PV1·PV2 series string	KVL: V=ΣVᵢ	Section I, ISOS	Verified
② Contamination	PV2 shading→current drop	Shockley equation	Ref.[134] PV mismatch	Verified
③ Parallel switch	PV1‖PV2 switching	I_p=I₁+I₂=21A	Section III ICS	Verified
④ Equiv. transform	2→1 panel (50V,21A)	Norton equivalent theorem	Fig.1(c) IPOS	Verified
⑤ String reconfig.	29 panels series, I=20A	KCL minimum tracking	Section III OVS auto	Verified
⑥ Power maximization	29,000W output	P=V×I (Joule)	Section III-B,D	Verified

Paper Table III Comparison — SUNGUIDE Positioning by IPOS Control Method

Control Method	ICS Achieved	OVS Achieved	Module Independence	Extra Circuit	SUNGUIDE match
Auto-Balancing (DC transformer mode)	Auto	Auto	Full	None	★ Match
POVG Distributed Control	Possible	Possible	High	Control bus	Reference
OVS Control Loop	Possible	Possible	Low	Sensor+controller	Not needed
Master-Slave ICS	Possible	Possible	Low	Master controller	Not needed
SUNGUIDE HW switching	Physical	Physical	Full	Switch only	★ Optimal

SUNGUIDE's hardware-based parallel switching falls under the "Auto-Balancing" category among all IPOS control methods classified in the paper, achieving ICS and OVS simultaneously through physical connection changes alone, without any additional control circuit.

IEEE TPEL 2020, Table III, p.12389 ↗

Power Output Comparison by Method — 30 panels / 50V / 20A / IPC 15 units in series / 1 contaminated panel basis

Normal reference: 1,500V × 20A = 30,000W | SDU 2019 measured loss rate applied | Contamination current 1A condition

Method	String voltage	String current	Output power	Efficiency	Power loss	Loss cause
Normal (reference)	1,500V	20A	30,000W	100%	0W	—
Conventional series (string inverter)	1,500V	1A	1,500W	5.0%	28,500W	KCL minimum current tracking
Bypass diode	1,450V	20A	28,964W	96.5%	1,036W	1-step voltage loss + diode heat loss 36W
Individual MPPT (Tigo/SolarEdge) SDU measured	~1,500V	~19.1A	~28,338W	94.5%	~1,662W	DC-DC conversion + MCU fixed consumption + connectors (SDU 2.45%)
SUNGUIDE / IPC 15 units	1,450V	20A	29,000W	96.7%	1,000W	IPC MOSFET conduction 0.8W × 1 unit only

Limitations of Individual MPPT (Micro-inverter) — SDU Paper Measured Evidence

✗

Limitation ①: 2.45% self-consumption loss even on clear days (SDU measured)
SDU 2019 paper Section 2.1 measured 42 panels: under no-shading conditions, both Tigo and SolarEdge systems yielded less energy than the standalone SMA system. SolarEdge's theoretical combined efficiency of 97.6% (higher than SMA's 96.5%) was still inferior in actual measurement. Reason: 30 optimizers × 0.4W MCU constant consumption = 12W + 30 DC-DC converter losses = 30,000W × 2.45% = 735W constant waste. At 1,500 kWh/kWp annual basis → ~368 kWh annual loss.

SDU 2019, p.8-9 — Section 2.1.1, Figure 5 ↗

✗

Limitation ②: Thermal runaway risk in 20A high-current environment
PE-type DC-DC converters at 20A combine inductor DCR loss + MOSFET switching loss. 30 optimizers each generating 7~10W heat → total 210~300W heat. Requires separate heatsinks or GaN components. SUNGUIDE IPC only MOSFET R_DS(on) = 2mΩ conduction → 0.8W per unit, total 12W for 15 units, minimizing heat.

✗

Limitation ③: Doubled connector count → reduced reliability (SDU measured)
SDU Section 1.4 states that "applying optimizers doubles the connector count and every connector is a potential failure source." Installing 30 optimizers adds 60 connectors. In fact, during the SDU experiment, one Tigo optimizer failed within a short period. Cumulative connector contact resistance losses also occur.

SDU 2019, p.5 — Section 1.4 ↗

✗

Limitation ④: Shading effect more limited than expected (SDU pole experiment)
In SDU Section 2.3 pole shading experiment, neither Tigo nor SolarEdge outperformed the SMA reference even under moving shade conditions. The paper concluded "optimizers simply tested bypass diodes, not module level mismatch." In other words, under light-to-moderate shading, a $200 optimizer performs no better than a $2 bypass diode.

SDU 2019, p.16-19 — Section 2.3 ↗

✗

Limitation ⑤: Powerless even under full contamination (1A) — SDU niche condition not met
SDU final conclusion: "Only for some small niches where complete PV modules have significantly different irradiation at any given time, do the optimizers help." When a fully contaminated panel occurs, the PE optimizer tries to match that panel's voltage to the remaining panel current (20A), but physically extracting 20A from a 1A panel is impossible. The result is identical to bypass diode activation: 1-step string voltage loss + additional self-consumption loss. In contrast, SUNGUIDE/IPC switches the contaminated panel in parallel with a normal panel, absorbing 1A into 21A.

SDU 2019, p.21 — Final Conclusion ↗

Bypass Diode vs. Individual MPPT vs. SUNGUIDE/IPC — Detailed Numerical Analysis

✓

Bypass Diode Loss Calculation (30 panels basis):
1 contaminated panel bypassed → string voltage 1,450V (−50V). String current 20A maintained.
Bypass diode heat: V_f × I × 3 units = 0.6V × 20A × 3 = 36W heat loss.
Output: (1,450V × 20A) − 36W = 28,964W (efficiency 96.5%).
Contaminated panel residual current 1A completely discarded → 50W additional opportunity loss.

✗

Individual MPPT Effective Loss Calculation (SDU 2.45% applied, 30 panels basis):
Theoretical output (individual MPP sum): 29 × 1,000W + 50W = 29,050W.
DC-DC conversion loss (SDU measured basis): 29,050W × 2.45% = −712W.
MCU fixed standby power: 30 units × 0.4W = −12W.
Effective output: 29,050W − 712W − 12W = 28,326W (efficiency 94.4%).
vs. SUNGUIDE/IPC: 29,000W − 28,326W = 674W additional disadvantage.

✓

SUNGUIDE/IPC Loss Calculation (15 units in series, 30 panels basis):
IPC parallel switching: contaminated panel + normal panel → 21A, 50V, 1-panel equivalent.
String: 29-panel equivalent × 50V × 20A = 29,000W.
IPC MOSFET conduction loss (1 active IPC): I² × R = 400 × 0.002 = only 0.8W.
14 normal IPCs: pure conduction path, loss ≈ 0W.
Total output: 29,000W (efficiency 96.7%) — vs. individual MPPT +674W advantage.

IEEE Academic Link (IPOS + DPP):
SUNGUIDE/IPC achieves the same goal as the Differential Power Processing (DPP) method of Ref.[134] (Bell & Pilawa-Podgurski, 2015) using a simple switch. DPP requires a separate DC-DC converter; SUNGUIDE/IPC achieves the same effect with only a MOSFET switch → significantly lower cost and complexity.

IEEE TPEL 2020, p.12400 — Ref.[134]; Section VIII ↗

30-Panel System Comprehensive Comparison — Clear Day (no shading) vs. 1 Contaminated Panel (SDU 2019 measured basis integrated)

Method	Clear day eff.	Clear day loss (W)	1-panel contam. eff.	1-panel contam. loss (W)	SDU verified	Cost/complexity
Conventional Series	100%	0	5.0%	28,500	Reference	Lowest
Bypass diode	~100%	~0	96.5%	1,036	Same as ref.	Low
Individual MPPT (Tigo/SolarEdge)	97.6%	720	94.4%	1,662	Below ref. (measured)	Highest
SUNGUIDE / IPC 15 units	~100%	~0	96.7%	1,000	SDU niche satisfied	Low

Key Conclusion: According to SDU measurements, individual MPPT yields less energy than bypass diodes, with a 2.45% loss even on clear days. Under 1-panel contamination, SUNGUIDE/IPC (96.7%) > bypass diode (96.5%) > individual MPPT (94.4%), with SUNGUIDE/IPC superior in all conditions. The "~100% theoretical efficiency" of individual MPPT is a lab figure, clearly disproven by SDU field measurements.

SDU 2019, p.21 — Final Conclusion ↗ | IEEE TPEL 2020 ↗

SDU Paper Original — Why Optimizer Self-Consumption Power Occurs

Wulf-Toke Franke (Assoc. Prof.), "The Impact of Optimizers for PV-Modules: A comparative study", SDU Mads Clausen Institute, May 2019 · URL: sdu.dk/CIE

✓

Paper conclusion (Section 3): "PV systems have typically a higher yield if no module optimizers are applied." In most conditions, systems without module optimizers achieve higher energy yield.

SDU 2019, p.21 — Final Conclusion ↗

✓

Root cause of loss — series power electronics insertion: Paper Section 2.1 states: "All string current must pass through the optimizer's internal power semiconductors, and even when the optimizer is inactive, at least one semiconductor device creates a voltage drop causing power loss (current × voltage drop)."

SDU 2019, p.8-9 — Section 2.1 Figure 5 commentary ↗

✓

Doubled connector loss: Paper Section 1.4 states that applying optimizers doubles the connector count, and this additional contact resistance constitutes part of "MLPE's energy self-consumption."

SDU 2019, p.5 — Section 1.4 Module Optimizer ↗

Fixed peripheral circuit power consumption: pv-magazine commentary by Prof. Marco Jung of Fraunhofer IEE explains: "Each MLPE device includes semiconductor drivers, relays, galvanic isolation, and microcontrollers; these peripheral circuits always consume a relatively constant amount of power." This is the 'fixed self-consumption power' that occurs even on clear days.

pv-magazine "The Long Read: Shadow Boxing", Jul. 2020 — Marco Jung, Fraunhofer IEE ↗

Loss Structure — PE vs. DR Comparison

Loss Type	SolarEdge / Tigo (PE)	SUNGUIDE (DR)
DC-DC conversion loss	Always present Buck-Boost inductor DCR + MOSFET Rdson	None No voltage conversion
Switching loss	Always present kHz-class PWM switching × 14 panels	Virtually none Single low-frequency relay switching
Microcontroller consumption	0.3~0.5W/unit × 14 = 4.2~7W fixed loss	Minimal LoRa MCU sleep mode standby
Extra connector resistance	28 extra (14 panels×2) Cumulative contact resistance	None Existing MC4 unchanged
Bypass diode	Always conducting 0.5~0.7V × 20A = 10~14W heat	Replaced Replaceable with MOSFET switch

SDU Measured Data — Clear Day / No Shading Condition (2018.05.31~08.14)

42 panels, 3 strings (Tigo, SolarEdge, SMA reference string)

System	Configuration	EU Efficiency	No-shading relative yield (measured)	Loss cause
SMA (reference)	String Inverter only	96.5%	100% (reference)	None
Tigo TS4-R-O	SMA + Tigo MLPE × 14	96.5% × MLPE eff.	Below reference	MLPE series insertion loss + connectors × 28
SolarEdge P300	SE HD-Wave + P300 × 14	97.6% (combined)	Below reference	Below SMA despite higher theoretical efficiency
* Paper conclusion: "the highest energy harvest is achieved without any optimizers" — on clear days, the no-optimizer system always records the highest yield

The SolarEdge P300 optimizer is rated at 99.5% peak efficiency and 98.8% weighted efficiency per its datasheet — higher than SMA inverter's EU efficiency of 96.5%. Yet in the paper's measurements, it fell behind the standalone SMA system under no-shading conditions. The paper states the reason: "losses from semiconductor voltage drops accumulate across all panels even when the optimizer is inactive."

SDU 2019, p.8-9 — Section 2.1.1, Figure 5 ↗

Bypass Diode Hardware Loss Physical Analysis

✓

Bypass diode forward voltage drop: Silicon diode forward voltage drop (V_f) is 0.5~0.7V. Heat at 20A = V_f × I = 0.6V × 20A = 12W/unit. With 3 bypass diodes built into each panel, up to 36W/panel potential loss.

✓

Implications of SDU pole shading experiment: Paper Section 2.3 concluded from partial shading by a 20cm-diameter pole: "optimizers simply tested bypass diodes, not module level mismatch." Under light-to-moderate shading, PE optimizers provide no more benefit than bypass diodes while adding their own losses.

SDU 2019, p.16-19 — Section 2.3 ↗

✓

SUNGUIDE bypass diode replacement potential: Using SUNGUIDE's MOSFET switch (R_DS(on) ≈ 2mΩ) as bypass path, heat at 20A = I² × R = 400 × 0.002 = 0.8W. ~15× lower conduction loss compared to a diode.

Overall loss context: For a 10kWp no-shading system, MLPE self-consumption causes 120~300 kWh/year loss (NCA Solar analysis: DC-DC conversion 0.5~1.5% + standby power 28~84 kWh/yr + connector loss 5~15 kWh/yr). SUNGUIDE's pure conduction loss is less than 1/10 of this figure.

SDU Paper Key Findings — Conditions Where Optimizers Lose vs. Gain

Scenario	Tigo result	SolarEdge result	SMA reference result	SUNGUIDE expected	SDU paper basis
No shading, clear day	Below reference	Below reference	Best	Same as ref.	Section 2.1.2 Fig.6
No shading, cloudy day	Below reference	Below reference	Best	Same as ref.	Section 2.1.3 Fig.7
1 panel constant different irradiance (semi-transparent cloth)	Conditional gain	Possible gain	Reference	Best (parallel switch)	Section 2.2.2 Fig.10
Moving pole shadow (clear day)	No gain	No gain	Best or same	Best (selective parallel)	Section 2.3.2 Fig.15
Moving pole shadow (variable clouds)	Minor gain	Minor gain	Reference	Gain (instant switch)	Section 2.3.3 Fig.16
Full shading (contamination/dust)	Negligible effect	Negligible effect	Lowest (90% loss)	Best (99.4% maintained)	Core of this analysis

The SDU paper concludes: "Only for some small niches where complete PV modules have significantly different irradiation at any given time, do the optimizers help produce more energy than they actually consume" — this is the paper's conclusion. SUNGUIDE provides a far more efficient solution than PE optimizers in this niche condition (1 fully contaminated panel), while also having no self-consumption loss on clear days.

SDU 2019, p.21 — Section 3 Final Conclusion (full PDF download) ↗

Panel Internal Cell String Structure — Why 5% Shading Creates 0A

Reference papers: ① Dhimish et al. "The Effects of Bypass Diodes on Partially Shaded Solar Panels", ResearchGate 2023 | ② EPJ Photovoltaics "The effect of partial shading on the reliability of PV modules", 2024 (IEC 61215-2:2021 based) | ③ PVEducation.org "Bypass Diodes" (UNSW/ANU research basis)

Structural Principles — 4 Key Facts

✓

3 sub-strings within a panel (vertical series): A standard 60-cell panel consists of 3 sub-strings of 20 cells each, connected vertically in series. The 3 sub-strings are connected horizontally in parallel. One bypass diode in the junction box connects to each sub-string.

PVEducation.org "Bypass Diodes" — UNSW/ANU ↗

✓

The harsh consequence of KCL: When just one cell in a sub-string is shaded and its current production approaches 0, the remaining 19 series-connected cells must also follow to 0A. The bypass diode detects this and short-circuits the entire sub-string, removing 1/3 of the voltage. This is the physical reason for "1 shaded cell = entire sub-string 0A."

EPJ Photovoltaics 2024 — IEC 61215-2:2021 based experimental verification ↗

✓

Explosive destructive power of horizontal shading: A vertical 1-line shadow affects only 1 sub-string, causing 33% loss. But a horizontal 1-line shadow crosses all 3 sub-strings A·B·C simultaneously, activating all 3 bypass diodes. Result: panel voltage ≈ 0V, power ≈ 0W.

Tchira & Thompkins 2023, Journal of Student Research Vol.12 No.4 — horizontal shading experiment: power loss 98.97% ↗

The reality of "5% shading = 50~80% power loss": The reason papers state "up to 50~80% power loss is possible even with 5% shading" is precisely this structural characteristic. The loss is not proportional to the simple area ratio (5%) but is non-linearly amplified depending on the direction and which sub-string is crossed.

Tchira & Thompkins, Journal of Student Research 2023, Vol.12 No.4 — DOI: 10.47611/jsrhs.v12i4.5288 ↗

Experimental Data by Contamination Location and Direction (Dhimish et al. 2023)

Contamination type	Shaded area	Affected sub-strings	Power loss	Remaining output
1 cell vertical shading	1.67%	1 (BPD 1 active)	74.42%	25.6%
Entire 1 sub-string	33%	1 (BPD 1 active)	79.54%	20.5%
1 horizontal line (<5%)	<5%	3 simultaneous (BPD 3)	98.97%	1.03%
1 vertical line	~5%	1 (BPD 1 active)	~33%	~67%
Horizontal+vertical cross	~10%	2~3	80~99%	1~20%

EPJ Photovoltaics 2024 experimental conclusion (IEC 61215-2:2021): "For the PERC Full-cell module, because two strings are connected in series (no parallel connected strings), shading of a cell resulted in a complete loss of current from the string to which the shaded cell is connected." That is, whether a full cell or half cell, the current of the series string touched by shading is completely blocked.

Özkalay et al., EPJ Photovoltaics 2024, Section 3 — Partial shading experiment on PERC modules ↗

SUNGUIDE IPC vs. Bypass Diode — Cell String Structure Response Comparison

Shading scenario	Bypass diode only	SUNGUIDE IPC
1 vertical cell (1.67%)	33% voltage loss Output ~66.7%	Current summed with parallel panel Output recovered
1 horizontal line (<5%)	3 BPDs simultaneously active Output ~1% (effectively 0W)	Panel-unit parallel switching String current 99.4% maintained
Dust/bird droppings (spot)	Probability of 1~3 BPDs active 33~99% loss depending on location	Absorbed via parallel switching Current summed (20+residual A)
BPD intrinsic heat loss	0.6V × 20A = 12W/unit 3 simultaneous = 36W heat	0.8W when replaced with MOSFET Based on R_DS(on) 2mΩ

Key insight — Why 5% area creates 0A:
Physical path of 5% horizontal contamination: ① 1 horizontal line of dust/shadow → ② 1 cell each in Sub-strings A, B, C shaded simultaneously → ③ each string KCL: 0A → ④ BPD-A, BPD-B, BPD-C simultaneously activated → ⑤ panel voltage = 3 × (−0.6V) = −1.8V (short circuit) → ⑥ effective output 0W. This mechanism proves that simple contamination ratio calculation (5% area = 5% loss) is completely wrong. SUNGUIDE IPC blocks this panel-level catastrophe via inter-panel parallel switching.

PVEducation.org ↗ + Tchira & Thompkins JSR 2023 ↗ + EPJ PV 2024 ↗ combined

✓

IEC 61215-2:2021 Official Experimental Verification: EPJ Photovoltaics 2024 paper conducted cell partial shading experiments on PERC, HJT, and IBC panels per international standard IEC 61215-2:2021, confirming: "Complete or partial shading of a cell is sufficient to activate the diode in the PERC and HJT modules." That is, partial shading of even one cell activates the bypass diode, setting string current to 0A.

Özkalay et al., EPJ Photovoltaics 2024, Section 3.1 conclusion ↗

Paper Overview — Latest Experimental Verification of SUNGUIDE Operating Principle

Wing Kong Ng, Nesimi Ertugrul — "Real-Time Reconfiguration of PV Arrays and Control Strategy Using Minimum Number of Sensors and Switches"
Energies 2025, 18(22), 5866 · Published: 6 November 2025 · DOI: 10.3390/en18225866 ↗ · Full PDF ↗

✓

Core paper proposal = Same principle as SUNGUIDE: "Recovers mismatch losses by real-time parallel switching of the affected module and adjacent normal module in a PV array experiencing shading/contamination." Creating a parallel group of 1 contaminated panel + 1 normal panel is the paper's core mechanism.

Ng & Ertugrul, Energies 2025 — Abstract ↗

✓

Limitations of existing methods — stated in paper: ① Bypass diode: only prevents damage, cannot recover mismatch loss. ② Conventional reconfiguration: N²-level switch count, irradiance sensor required → excessive cost/complexity. To overcome this, the paper uses only minimal hardware (N+1 MOSFET + 2N-3 diodes).

Ng & Ertugrul, Energies 2025 — Section 1 Introduction ↗

✓

Control strategy — voltage-based real-time detection: Compares only each module voltage V_k with average V_s_avg, without a separate irradiance sensor. If V_k < V_s_avg → shading detected → parallel switch with adjacent module → automatic return after current equalization. Re-evaluated every cycle via MPPT.

Ng & Ertugrul, Energies 2025 — Section 3 Control Strategy ↗

Conduction loss minimization design: Most parallel paths use diodes (resistance ~10 mΩ) with minimum MOSFETs. This is the same philosophy as SUNGUIDE, which generates only MOSFET R_DS(on) 2 mΩ conduction loss (0.8W at 20A).

Ng & Ertugrul, Energies 2025 — Section 2 Hardware ↗

Paper Experimental Results (3 modules × solar simulator)

Scenario	MPP before reconfig.	MPP after reconfig.	Power improvement	Improvement rate
S1·S2 shading → S(2,1) parallel + S3 series	38.4 W	45.6 W	+7.2 W	+18.75%
S2·S3 shading → S(3,2) parallel + S1 series	38.4 W	45.8 W	+7.4 W	+19.27%

Paper Simulation Results (4 modules × extreme conditions)

Scenario	Before reconfig.	After reconfig.	Improvement rate
Extreme shading scenario A	463 W	895 W	+93.3%
Extreme shading scenario B	900 W	1,100 W↑	+22%↑
Full recovery after shading removal	—	1,500 W	100% recovery

"the reconfiguration successfully redistributed the mismatch losses more effectively, enabling a higher operating power point under the given shading conditions." P-V/I-V curve measurements confirmed LMPP disappears after reconfiguration, converging to a single GMPP.

Ng & Ertugrul, Energies 2025 — Section 4 Experimental Results ↗

SUNGUIDE vs. Energies 2025 Paper — Structural Comparison Verification

Item	Energies 2025 proposal	SUNGUIDE IPC	Match
Basic operation	Series → parallel dynamic switching	Series → parallel hardware switching	Identical
Parallel unit	Adjacent 2 panels (consecutive modules)	PV1 + PV2 (2 panels)	Identical
Current mechanism	I_parallel = I_dirty + I_normal	I_parallel = 1A + 20A = 21A	Identical
String restoration	Parallel group returns as 1-panel equivalent to series	29-panel equivalent series string restored	Identical
Switching device	MOSFET + diode (minimal)	MOSFET switch (minimal)	Identical
Conduction loss	~10 mΩ diode basis	2 mΩ MOSFET R_DS(on)	Equal or better
Irradiance sensor	Not needed (voltage only)	Not needed (V/I sensing)	Identical
Non-consecutive modules	Adjacent modules only (limited)	LoRa-based full string control	SUNGUIDE advantage

While the paper admits the limitation of "only adjacent modules can be paralleled," SUNGUIDE can parallel-switch any panel pair at any position within the string via LoRaWAN MASTER/SLAVE communication, covering a wider range of shading scenarios than the paper.

Ng & Ertugrul, Energies 2025 — Section 5 Limitations ↗

Energies 2025 Experimental Data Visualization

Before reconfig. (MPP) After reconfig. (MPP) Recovery after shading removal

The 4-module simulation improvement of 463W → 895W (+93.3%) is experimental evidence of the same mechanism as SUNGUIDE's 800W → 7,950W (+893%) improvement. The numbers differ only due to panel count and contamination level; the physical principle of recovering mismatch loss via parallel switching is identical.

Ng & Ertugrul, Energies 2025 — Section 4, Table 3 ↗

4 Academic Bases the Energies 2025 Paper Provides for SUNGUIDE

✓

Basis ①: Experimental proof of power recovery via series-parallel variable switching
3-module actual hardware experiment confirmed 38.4W → 45.6/45.8W (+18.75~19.27%) improvement via P-V/I-V curve measurement. SUNGUIDE's identical mechanism is experimentally verified.

Energies 2025, Section 4 ↗

✓

Basis ②: Proof of LMPP elimination and single GMPP convergence
Verified via P-V curve that the Local MPP (LMPP) present before reconfiguration completely disappears after parallel switching, converging to a single Global MPP (GMPP). This allows the inverter's MPPT algorithm to easily find the optimal operating point.

Energies 2025, Section 4 Fig.11-12 ↗

✓

Basis ③: Proof of maximum efficiency achievability with minimal hardware
Proves implementability with only N+1 MOSFET + 2N-3 diodes. Supports the academic validity of SUNGUIDE achieving PE optimizer-level power recovery with only relay/MOSFET switches.

Energies 2025, Section 2 ↗

✓

Basis ④: Confirmed automatic recovery (1,500W) after shading removal
Simulation confirmed that when shading is removed, the system automatically returns to series configuration and recovers maximum power (1,500W). This exactly matches SUNGUIDE's "zero loss under normal conditions" claim.

Energies 2025, Section 4 ↗

Key Question — Why Do the Remaining 28 Panels Come Alive When Only 2 Are Paralleled?

When IPC switches only 2 panels — contaminated panel (PV15) + adjacent normal panel (PV14) — in a 30-panel series string, the remaining 28 panelspanels, whose circuits were not touched at all, increase their output from 50W → 1,000W (20× increase). This explains step by step using circuit theory how this is possible.

BEFORE — Prior to IPC switching (all 30 panels in series)

①

KCL Series Constraint (Kirchhoff's Current Law):
A series string has only one current path. Therefore I_string = min(I_PV1, ..., I_PV30) = 1A (based on contaminated panel PV15).
Even though the remaining 29 panels can generate 20A, physically more than 1A cannot flow.
Each normal panel output: 50V × 1A = 50W (only 5% of rated 1,000W used)

②

MPP Departure (Maximum Power Point Departure):
In a solar cell's P-V curve, MPP is around I_mp ≈ 18~20A. When string current is forced to 1A, all 28 normal panels operate at operating points far from MPP. This is an impedance mismatch state, in complete violation of the Maximum Power Transfer Theorem.

Overall system state: Total output = 30 panels × 50V × 1A = 1,500W (only 5.0% of rated 30,000W recovered)

AFTER — Post IPC switching (PV14 ‖ PV15 parallel, remaining 28 panels remain in series)

③

Norton Equivalent Transform (Norton's Equivalent Theorem):
When IPC connects PV14 (normal, 50V, 20A) and PV15 (contaminated, 50V, 1A) in parallel:
· Common voltage: V_pair = min(50V, 50V) = 50V maintained (KVL)
· Summed current: I_pair = 20A + 1A = 21A (KCL parallel)
→ This pair forms a (50V, 21A) Norton equivalent single-panel equivalent joining the string.

④

KCL Constraint Released — 28 Panels Simultaneously Freed:
New string: [PV1~PV13] + [PV14‖PV15 equivalent] + [PV16~PV30] = 29-panel equivalent series.
New string current: I_string = min(21A, 20A, 20A, ...) = 20A
IPC only touched PV14·PV15, but with the KCL constraint released, all 28 panels simultaneously return to 20A.
This is the physical mechanism of "local switching → global effect."

⑤

MPP Return — Impedance Matching Re-achieved:
28 normal panels return from 1A (5% MPP) → 20A (~100% MPP).
Per the Maximum Power Transfer Theorem, the operating point moves to where load impedance = panel internal impedance, i.e., near MPP.
Each panel's P-V curve operating point converges from LMPP → GMPP. (Verified by Energies 2025 paper experiment Fig.11-12)

Overall system state:
· 28 normal panels: 50V × 20A = 1,000W/panel × 28 = 28,000W
· PV14‖PV15 equivalent: 50V × 20A = 1,000W
· Total output: 29,000W (96.7% of rated capacity recovered)

Before vs. After Contaminated Panel Switching — PV14‖PV15 Parallel Pair Highlight Visualization

Before switching — normal panels (KCL 1A limit) Before switching — IPC parallel pair PV14·PV15 (blue, same 1A limit)

▼ Before switching (30 panels, all 50W — KCL 1A limit)

After switching — 28 normal panels (1,000W, MPP restored) After switching — IPC parallel pair PV14‖PV15 merged 1 panel (1 blue bar, 1,000W)

▼ After switching (29-panel equivalent — PV14‖PV15 parallel → 1 blue bar)

Graph interpretation: Before switching: 30 panels → after switching: merged into 1 blue bar (PV14‖PV15 parallel equivalent), displayed as total 29 panels. By Norton equivalent principle, 2 panels become 1, and the remaining 28 panels all return to 1,000W.

Circuit Theory Summary — Chain Effect of 3 Laws

Step	Applied law	Formula	Result	Scope of effect
Step 1	KVL parallel voltage sharing	V_pair = min(V14, V15) = 50V	Common voltage 50V maintained	PV14, PV15 only
Step 2	KCL parallel current summation	I_pair = 20A + 1A = 21A	Contaminated current absorbed	PV14, PV15 only
Step 3	Norton equivalent theorem	(50V, 21A) → 1-panel equiv.	2 panels → 1 panel substituted	Entire string structure
Step 4	KCL series current determination	I_string = min(21A, 20A×28) = 20A	KCL constraint released	All 28 panels simultaneously
Step 5	Maximum Power Transfer Theorem	P_max when Z_load = Z_source*	28 panels MPP restored	All 28 panels simultaneously
Result	Joule power law	P = 1,450V × 20A	29,000W (96.7%)	Entire string

Key insight: IPC switching physically acts on only PV14·PV15, but through KCL — the fundamental law of series circuits — it simultaneously shifts the operating point of all 28 remaining panels. This is the electrical engineering reason a simple relay switch can change the entire system efficiency from 5% → 96.7%. The remaining 28 panels are "liberated" even though their circuit connections were not changed.

IEEE TPEL 2020 — IPOS ICS/OVS auto-achievement principle ↗ | Energies 2025 — LMPP→GMPP convergence experimental verification ↗

Advanced Theory ① — Shockley Single-Diode 5-Parameter Model

★

Why voltage maintains 50V after parallel switching — semiconductor physics proof
The current-voltage characteristics of a solar cell are precisely described by the following equation:
I = Iph − I₀·[exp((V + I·Rs)/(n·Vt)) − 1] − (V + I·Rs)/Rsh

· Iph (photocurrent): Proportional to irradiance. Under contamination, Iph decreases from 20A → 1A.
· I₀ (reverse saturation current): Depends only on temperature, unaffected by contamination.
· Vt (thermal voltage): Vt = kT/q ≈ 26mV (at 25°C), unchanged.
· Voc formula derivation: Substituting I=0 gives Voc ≈ (n·Vt)·ln(Iph/I₀ + 1)

Voc change calculation for contaminated panel (Iph: 20A → 1A):
ΔVoc = n·Vt·ln(1/20) ≈ 1.5 × 0.026 × (−3.0) ≈ −0.12V

Conclusion: Even though contamination reduces Isc by 95%, Voc decreases by only 0.12V (0.24%). This is the semiconductor physics basis for the common voltage maintaining nearly 50V after parallel switching. The "V_pair = 50V maintained" explanation in this tab is not mere intuition but formula-proven by the Shockley equation.

Advanced Theory ② — Fill Factor Degradation Analysis (FF Degradation)

★

How contamination reduces the P-V curve "area" and how parallel switching restores it
FF = Pmax / (Voc × Isc)

· Normal panel FF: Pmax=1,000W / (52V × 20A) = 0.962
· Contaminated panel FF: Pmax=50W / (51.88V × 1A) = 0.963 (FF itself maintained)

Core paradox: The contaminated panel's FF barely changes because contamination only reduces Isc. The problem is that series connection forces the effective FF of all 28 normal panels to collapse to 0.048. (Each normal panel: P_actual = 50V × 1A = 50W, FF_eff = 50W / (52V × 20A) = 0.048 — 95% of MPP area vanishes)

After IPC parallel switching:
· 28 normal panels effective FF restored: 0.048 → 0.962 (×20 times)
· P-V curve area restored close to original rectangle
· Energies 2025 paper Fig.11-12 confirmed P-V curve area restoration by measurement

Advanced Theory ③ — Mismatch Loss Formula Definition (IEEE definition basis)

★

How many W and why — quantitative loss decomposition
P_mismatch = P_STC_sum − P_string_actual

Pre-switching Mismatch Loss decomposition:
· P_STC_sum (sum of individual STC outputs): 29 panels × 1,000W + 1 panel × 50W = 29,050W
· P_string_actual (actual string output): 30 panels × 50V × 1A = 1,500W
· P_mismatch = 29,050W − 1,500W = 27,550W (94.9% loss)

Loss breakdown by cause:
· Current-Forced Loss: 28 panels × (1,000W − 50W) = 26,600W
· Contamination Loss: 1 panel × (1,000W − 50W) = 950W
· Total: 27,550W

Post-switching Mismatch Loss:
· P_string_actual: 29 × 50V × 20A = 29,000W
· P_mismatch = 29,050W − 29,000W = 50W (only 0.17% remains)
→ 1 IPC unit operation reduces Mismatch Loss from 27,550W → 50W by 99.8%

Advanced Theory ④ — Thevenin Equivalent (Norton's Dual Theory)

★

How the string "seen" by the inverter MPPT changes
Norton equivalent (current source) and Thevenin equivalent (voltage source) express the same circuit from different perspectives:
V_th = I_N × Z_N | Z_th = Z_N

Before switching — Thevenin equivalent seen by inverter MPPT:
· V_th = 30 × 50V = 1,500V (open-circuit voltage)
· I_sc = 1A (short-circuit current = contaminated panel Isc)
· Z_th = V_th / I_sc = 1,500Ω (abnormally high internal impedance)
· MPPT result: Multiple LMPP (Local MPP) appear on P-V curve → algorithm confusion, optimal point tracking fails

After switching — Thevenin equivalent seen by inverter MPPT:
· V_th = 29 × 50V = 1,450V
· I_sc = 20A (string short-circuit current normally restored)
· Z_th = 1,450V / 20A = 72.5Ω (normal internal impedance)
· MPPT can immediately track single GMPP → inverter efficiency maximized

Thevenin impedance change: 1,500Ω → 72.5Ω (reduced to 1/20.7)
This is why the inverter MPPT algorithm immediately finds the optimal operating point after IPC switching. The Maximum Power Transfer Theorem (Z_load = Z_th*) is finally satisfied.

Advanced Theory ⑤ — Superposition Theorem

★

Mathematical proof that 1 contaminated panel dominates the entire string
Superposition theorem: In a linear circuit where multiple current sources act simultaneously, the sum of each source's independent contribution equals the total response.

Applying superposition to 30 current sources:
Treating each panel as an independent current source to analyze their contribution:
· PV1~PV14, PV16~PV30 individual contribution: 20A each (normal)
· PV15 individual contribution: 1A (contaminated)

Superposition result in series circuit — min() dominance law:
A series circuit has a single current loop, so each current source cannot contribute independently; the lowest value dominates the entire circuit:
I_string = min(20A, 20A, ..., 1A, ..., 20A) = 1A

PV15 (1A) alone completely cancels the contribution of the other 29 panels.
Potential current contribution of 29 normal panels: only 1A of 29 × 20A = 580A utilized (0.17%)

Superposition re-analysis after IPC parallel switching:
PV14‖PV15 → substituted with Norton equivalent I_N = 21A.
New min() operation: min(20A, ..., 21A, ..., 20A) = 20A
→ "Dominance effect" of contaminated panel eliminated → all 28 panels return to 20A contribution

The core message of the Superposition Theorem: In a series circuit, the single weakest current source determines the entire system's operating point. IPC reverses the dominance of superposition by reinforcing that "weakest link" in parallel. Mathematically equivalent to removing the dominant element of the min() operation.

Experimental Data Dashboard — SUNGUIDE IPC