boardrepair.me › Boards › 820-01588
iMac / Mac mini
Board Number 820-01588
A-Number(s) A1418
Model iMac 21.5" 2017
Year 2017

Board 820-01588 — iMac 21.5" 2017

The 820-01588 is the primary logic board for the 2017 iMac 21.5" (A1418) with Intel Kaby Lake processors and discrete AMD Radeon Pro 555 graphics. This board is notable for its integration of the ISL6259 multi-rail power controller and ITE IT8913E embedded controller, making it a critical point for both DC-DC rail stability and thermal management—two areas that frequently fail in the field, especially following power surges or long-term thermal stress.

Key Test Points & Voltages

Net / Test Point Power State Expected Voltage Notes
PPVRTC_G3H S5 (standby) 3.30 V ±0.1 V Real-time clock and NVRAM rail. Powered by backup battery or standby converter. Critical for detecting SMC/EC issues.
PP3V42_G3H S5 (standby) 3.42 V ±0.05 V Standby 3.3 V with boost for memory retention. Sourced from ISL6259. Always-on rail.
PPBUS_G3H S5 (standby) 19.0–20.5 V Main distributed bus voltage. Generated from adapter input. First point to check in no-power cases.
PP5V_S5 S5 (standby) 5.00 V ±0.1 V Standby 5 V for USB, Ethernet PHY, and PCI devices. ISL6259 output.
PP3V3_S5 S5 (standby) 3.30 V ±0.05 V Standby 3.3 V for logic I/O. Core EC and BMC rail.
PP1V8_S3 S3 (sleep) 1.80 V ±0.05 V Substrate rail for memory controllers. Enabled by EC once PPBUS_G3H is stable.
PP3V3_S3 S3 (sleep) 3.30 V ±0.05 V DDR3 memory rail. Sourced from ISL6259. Must be stable before CPU rails enable.
PPVCORE_S0 S0 (fully on) 0.85–1.10 V (dynamic) CPU core rail. Voltage scales with frequency. Monitor for ripple (<30 mV) and transient droop (<50 mV).
PPVDDIO_GPU S0 (fully on) 0.95 V ±0.05 V Radeon Pro 555 IO supply. Critical for eDP/HDMI signal integrity and GPU stability. Known instability point.
PPVCORE_GPU S0 (fully on) 0.90–1.05 V (dynamic) Radeon Pro 555 core supply. Subject to voltage droop under load. Common cause of GPU artifacts and display flickering.
PP1V2_MEM S3 (sleep) 1.20 V ±0.05 V Memory termination voltage. Minor rail but essential for DDR stability.
PP1V05_AUX S3 (sleep) 1.05 V ±0.05 V Auxiliary substrate rail for EC and power management ICs. Always stable if PP3V3_S3 is present.

Diode Mode Reference

Diode mode measurements are taken with a digital multimeter in diode mode, probes reversed (black to test point, red to ground) to measure forward voltage drop across semiconductor junctions. These values are sensitive to temperature; readings should be taken at room temperature (20–25 °C). Significant deviation (>0.05 V) may indicate a shorted or open component in that power domain.

Test Point / Net Red (−) to GND Expected Range Common Bad Reading Notes
PPVRTC_G3H (source node) Black probe on PPVRTC_G3H, red on GND 0.48–0.52 V 0.0 V (open) or >0.65 V (shorted) Primary indicator of PPVRTC FET health. Shorted reading = dead standby converter.
PP3V42_G3H (source node) Black probe on PP3V42_G3H, red on GND 0.49–0.53 V 0.0 V or >0.60 V ISL6259 boost converter output. Test at ISL6259 pin 1 and pin 2 simultaneously.
PPBUS_G3H (adapter input) Black probe on PPBUS_G3H, red on GND 0.45–0.55 V 0.0 V (open) or >0.70 V (internal short) Main supply bus. Very sensitive test—bad reading almost always = main fuse F9700 blown or connector issue.
PP5V_S5 (ISL6259 output) Black probe on PP5V_S5, red on GND 0.46–0.54 V 0.0 V (dead rail) or >0.60 V (load short) If open, check PPBUS_G3H first, then ISL6259 enable line and load-side FETs.
PP3V3_S5 (EC rail) Black probe on PP3V3_S5, red on GND 0.49–0.53 V 0.0 V or >0.55 V EC and standby logic. If shorted, board will not enumerate in any power state. Measure after confirming PP5V_S5 is good.
PP1V8_S3 (substrate bias) Black probe on PP1V8_S3, red on GND 0.44–0.50 V 0.0 V (converter fault) or >0.60 V (load short) Present only in S3+ states. Test with system awake. If open, CPU will not POST.
PPVCORE_S0 (CPU core)—measure at CPU socket Black probe on PPVCORE_S0, red on GND 0.48–0.58 V (at ~1.0 V nominal output) 0.0 V (converter dead) or >0.75 V (extreme short) Dynamic rail; diode reading is meaningful only if system is powered off. Bad reading = CPU VRM fault or short under CPU.
PPVDDIO_GPU (Radeon Pro 555 IO) Black probe on PPVDDIO_GPU, red on GND 0.46–0.52 V 0.0 V (open) or >0.65 V (shorted GPU IO domain) GPU IO supply. Shorted reading here = GPU package or surrounding cap issue. Often seen after thermal cycling.
PPVCORE_GPU (Radeon Pro 555 core) Black probe on PPVCORE_GPU, red on GND 0.47–0.54 V 0.0 V (converter dead) or >0.60 V (GPU shorted) GPU core supply. Instability here = common cause of GPU artifacts and kernel panics. Test load-side caps for ESR.
PP1V2_MEM (memory termination) Black probe on PP1V2_MEM, red on GND 0.50–0.56 V 0.0 V or >0.60 V Minor rail but essential for memory stability. If open, expect memory errors or kernel panics under load.
PPBUS_G3H FET source (pre-fuse side) Black probe on source side of F9700, red on GND 0.45–0.55 V 0.0 V (open connection) or >0.70 V (internal fuse fault) Diagnostic only if PPBUS_G3H reads open. If this reads good and main rail doesn't, fuse or connection is open.
PP3V3_S3 (memory supply) Black probe on PP3V3_S3, red on GND 0.48–0.52 V 0.0 V (converter fault) or >0.65 V (memory short) DDR3 supply. Shorted reading = bad DRAM module or power plane short. Test after ruling out PP1V8_S3 issue.

Common Faults & Symptoms

No Power — Blown Main Fuse F9700 After Surge

Symptom: Device does not power on; adapter LED does not light; no fans, no chime. PPBUS_G3H measures 0.0 V in diode mode.

Likely Cause: Main fuse F9700 (typically 20–24 A SMD fuse on the power input stage) has blown due to voltage surge, lightning strike, or failed main FET creating back-EMF. This is the single most common fault on 820-01588 boards after power transients.

Diagnostic Step: Visually inspect F9700 (located near J6200 power connector). Look for discoloration or visible crack in the fuse element. Use multimeter in continuity mode across the fuse; it should read ~0 Ω open. Measure PPBUS_G3H diode mode on both sides of the fuse. If pre-fuse side reads ~0.50 V and post-fuse side reads 0.0 V, fuse is definitely open. Check for short on PPBUS_G3H rail by measuring diode mode at several test points downstream; if all read 0.0 V, short is not present and fuse replacement is safe.

Fix: Replace F9700 with equivalent 20 A rated SMD fuse (typical part: Littelfuse 0805L200/24PR or equivalent). Before applying power, measure PP5V_S5, PP3V42_G3H, and PPVRTC_G3H in diode mode. All should read 0.48–0.54 V. If any reads 0.0 V or >0.60 V, short exists; do not proceed with power-on. If clean, apply power and verify PPBUS_G3H measures 19–20.5 V DC.

GPU Artifacts — Radeon Pro 555 VCore Instability

Symptom: Display shows random colored pixels, checkerboard patterns, or flicker during graphics-intensive tasks (gaming, video editing, Final Cut Pro). Kernel panics with GPU-related stack traces. System may appear to hang briefly during graphics load.

Likely Cause: PPVCORE_GPU rail is unstable under load due to capacitor ESR (equivalent series resistance) degradation, or VRM FET failure. The ISL6259 and associated PWM controllers struggle to maintain tight voltage regulation when the Radeon Pro 555 draws transient current spikes. Thermal stress exacerbates this—the GPU and surrounding power stage heat up, increasing ESR of output caps.

Diagnostic Step: Boot into macOS and run GPU Benchmark tool or open Activity Monitor > GPU. Measure PPVCORE_GPU with multimeter under load (probe placed directly on GPU VCore output capacitor). Voltage should remain 0.90–1.05 V; if it drops below 0.85 V or ripples >50 mV peak-to-peak, rail is unstable. Measure ESR of large output capacitors (typically 100 µF ceramic, 16 V, located within 5 mm of GPU package). ESR should be <5 mΩ; readings >10 mΩ indicate failed caps.

Fix: Replace all ceramic output capacitors on PPVCORE_GPU net (typically C5501, C5502, C5503, C5504—check schematic for exact designators). Use exact replacement value and type (e.g., 100 µF / 16 V X7R ceramic). After replacement, test under sustained GPU load for 30+ minutes. If artifacts persist, suspect GPU package itself; perform reflow of GPU BGA as secondary step. Verify PPVDDIO_GPU is also stable (should not exceed 1.0 V or drop below 0.90 V).

No Display — eDP Timing Board Fault

Symptom: System boots (fans spin, chime heard, keyboard backlight on), but display remains black. Using external display via Thunderbolt or HDMI shows video correctly. Internal panel connectors show no eDP signal.

Likely Cause: eDP timing controller IC (part of display support circuitry, typically on 820-01588 or separate eDP board) has failed, or power rails feeding the eDP subsystem (PP1V8_S3, PP3V3_S3) are unstable. The 2017 iMac uses an eDP backlight driver separate from the main logic board; communication between logic board and eDP board occurs via eDP protocol on dedicated header.

Diagnostic Step: Connect external Thunderbolt or HDMI display while powered on. If external display works, internal panel fault is confirmed. Measure PP1V8_S3 and PP3V3_S3 on the logic board; both should be stable at specification. Inspect eDP connector (typically a 30-pin or 40-pin FPC connector, J2000 or similar) for loose contacts, corrosion, or bent pins. Reseat the eDP cable firmly. Check for liquid damage or thermal paste migration around the eDP connector area. Use a magnifying glass to inspect for cracked solder joints on the eDP timing IC (if present on main board) or on the eDP daughter board.

Fix: If eDP cable is loose, firmly reseat it and test. If rails are unstable, repair PP1V8_S3 and PP3V3_S3 converters (check for bad capacitors or FETs, similar to GPU rail repair). If eDP IC has visibly cracked solder or is suspected dead, reflow the IC with a BGA rework station. If eDP daughter board is available as a separate spare part (common on 21.5" 2017 iMacs), consider replacement rather than attempting micro-soldering repair.

Diagnostic Workflow (No Power)

Use this step-by-step procedure when presented with a board that shows no signs of life: no chime, no fans, no lights on adapter.

  1. Verify adapter voltage: Unplug the power cord. Using a multimeter, check the AC mains voltage (120 V or 240 V depending on region). Plug in the 185 W adapter and measure DC voltage at the adapter output jack (barrel connector). Should read 19.5 ± 0.5 V DC. If 0 V, adapter is dead—replace it. If present, proceed.
  2. Check main fuse F9700: With adapter plugged in (system unpowered), use multimeter continuity mode to test fuse F9700. Should show ~0 Ω. If open (∞ Ω), check for short circuit before replacing (see step 3). If good, proceed.
  3. Measure PPBUS_G3H voltage: With system powered off and adapter plugged in, measure DC voltage directly on the PPBUS_G3H net using a DMM probe on a labeled test point or via the voltage rail after the main fuse. Should read 19.0–20.5 V. If 0 V, fuse is open; if present but low (<18 V), check connector J6200 and main sense line PSENSE. If <15 V, adapter regulation has failed—replace adapter. If voltage is good, proceed.
  4. Test standby rail diode mode: With system powered off and unplugged, measure PP3V42_G3H in diode mode (black probe to test point, red to ground). Should read 0.49–0.53 V. If 0.0 V, the standby converter is open circuit—likely failed FET or inductor in the boost stage. If >0.60 V, a short is pulling down the rail; trace to the short and remove or reflow suspect components. If reading is good, proceed.
  5. Measure standby logic rail PP3V3_S5: In diode mode with system unplugged. Should read 0.49–0.53 V. If open (0.0 V), the ISL6259 standby converter or its load (EC, memory, logic) has a short or open. If the reading is shorted (>0.60 V), suspect a bad capacitor bank near the ISL6259 or embedded controller IC (ITE IT8913E). If good, proceed.
  6. Power on and monitor PPVRTC_G3H: Plug in adapter; do not press power button. Immediately measure PPVRTC_G3H DC voltage on the board. Should rise to 3.30 ± 0.1 V within 2 seconds. If it rises slowly or remains <2.5 V, the RTC backup converter is weak; check capacitors and FETs. If it reaches spec, the standby power tree is functional. Proceed to step 7.
  7. Press power button and monitor rail sequence: Observe the following rails with a multimeter or oscilloscope in real-time or via fast logging: PPVCORE_S0 (CPU), PPVDDIO_GPU, and PPVCORE_GPU. The rails should sequence in this order: PP3V3_S5 → PP1V8_S3 → PP3V3_S3 → PPVCORE_S0 → GPU rails. Each rail should reach spec voltage and stabilize. If a rail does not appear or is stuck low, check the corresponding enable signal at the FET gate and the ISL6259 chip select and feedback pins.
  8. Check EC communication: If power sequencing stalls or system hangs partway through boot, use a logic analyzer to monitor the I2C bus between the main board and EC (typically at SDA and SCL test points). Look for repeated NACK responses or a stuck clock/data line. If I2C is hung, suspect EC firmware corruption or EC IC malfunction (ITE IT8913E).
  9. Verify CPU VCore under no-load condition: Once system reaches the BIOS screen or begins POST, measure PPVCORE_S0 once more. Should stabilize at ~0.95–1.05 V (varies by CPU frequency and stepping). If below 0.90 V, VRM is weak; if above 1.10 V, VRM is overvolting. Either case warrants VRM component inspection and replacement.
  10. Record all measurements and document: Photograph any visibly damaged components, record all test point voltages, and note the sequence of events. If board does not power on after step 9, issue is likely CPU, memory, or EC related; escalate to advanced micro-soldering diagnosis or board replacement.

Related Boards

The 820-01588 shares significant schematic and power delivery topology with the 820-01892, which is used in the 2017 iMac 27" (A1419) with similar Kaby Lake CPUs and Radeon Pro 555 graphics. The main differences are: the 27" variant requires higher current delivery on PPVCORE_S0 and PPVCORE_GPU due to higher-clocked CPU/GPU options; the 27" board has additional VRM phases and larger output capacitor banks. The eDP timing controller and display support circuitry differ between models due to panel size and backlight architecture. Both boards share the ISL6259 and ITE IT8913E chip placement, making many standby and sequencing faults directly transferable between the two models. Schematics and board layouts for the 21.5" variant (820-01588) and 27" variant (820-01892) should be cross-referenced when available.

See Also

← All board references · A-number lookup · Print