Car: 1999 Honda Accord EX (5-speed)
ECU: P28 — all capacitors replaced, J12 removed, and J1 jumped
S300v3 serial: 67277
Mods: Ghost key push button from Jordan Distributions, eBay cold air intake, full speaker system (removed for testing), 3-to-2-wire IACV conversion with Rosko conversion plate

1. Symptoms / History

• Installed S300v3 onto the P28 and initially uploaded my tune — S300 was turning on (not sure if sensors were reading).
• When finishing the installation of the 3-to-2-wire conversion, the S300 showed no live data. At that time D3/D4 (green) were lit but D2 rapidly flashed red.
• I reseated the S300v3 a few times to make sure it was correct; now when the S300 is hooked up it shows no lights at all.
• Plugging my original stock ECU back in makes the car work normally (no issues).
• Sent my ECU and S300v3 combo to be tested on an engine sim, which came back all good. Also, I had the P28 board switched out just in case.

2. What I’ve checked/measured (measurements taken at the OBD1 ECU connector side of conversion harness):

2.a. Baseline (ECU unplugged):

• A25 (IGP2): 11.92 V (key ON)
• B1 (IGP1): 11.92 V (key ON)
• A7 (fuel pump relay): 11.90 V key ON, drops to <0.4 V key OFF
• Grounds: A23, A24, A26, B2 → 0.0 Ω to chassis (good)
• Conversion harness continuity for D19–D22 → OK

2.b. Continuity (ECU unplugged):

• D19 → chassis: OL (open) — no short
• D20 → chassis: OL — no short
• D21 → A23: 0.0 Ω
• D22 → A23: 0.0 Ω

2.c. ECU plugged in (S300 removed), Key ON:

• A25: ~2.17 V (should be ~12V)
• B1: ~2.17 V (should be ~12V)
• D19 (VCC1): ~5 mV (should be ~5V)
• D20 (VCC2): ~5 mV (should be ~5V)
• D21 / D22: ~2 mV (should be ~0V)
• D4 (SCS): ~-0.9 mV (I forgot what this should be but this is only testing if the ECU is initialized, I think)

2.d. ECU plugged in, TPS / MAP / IACV disconnected (Checked after each sensor removed and still no change):

• A25: ~1.90 V ((should be ~12V))
• B1: ~1.88 V (should be ~12V)
• D19/D20 remain near 0 V (should be ~5V)
• D4 measured ~8.9 mV and slowly increasing to >22 mV if probe left in pin (odd behavior)

2.e. Connector isolation:

• Plugging Connector A (only) → A25/B1 collapse to ~1.8–2.0 V and constant CEL.
• Plugging Connector B (only) → A25/B1 collapse to ~1.8–2.0 V and no CEL.
• Plugging Connector D (only) → A25/B1 remain ~11.9 V (no collapse).

With ECU unplugged, A25/B1 read ~11.9 V (so vehicle/harness switched feed is fine until ECU is inserted).

2.f. Fuse & relay checks:

• ECU fuse (driver panel): 12.19 V both sides (battery and load side)
• Fuel pump fuse (driver panel): 12.18 V both sides

IGN1 relay continuity OK; replaced both fuses and the 7.5A for clock backup in passenger side panel — did not fix the issue.

3. Other notes:

• Wiggle-testing the harness/connector at A25 didn’t change the low voltage significantly.
• My multimeter is a cheap one from Amazon so it could be slightly off but from what I can see that isn't the issue.
• I've tried some of these tests with the S300V3 installed and removed but still got the same issues.

4. What this data shows / my current conclusion

• The harness and vehicle switched feeds are good (A25/B1 read ~12 V when ECU is unplugged).
• When the P28 is plugged in, A25/B1 collapse to ~2 V and the ECU 5 V rails (D19/D20) sit near 0 V — the P28 is not powering up its internal 5 V rails in-circuit.
• The collapse happens when either connector A or connector B is plugged into the ECU (connector D alone does not cause the collapse). That localizes the problem to something tied into pins present on A and B (power distribution/components near connector A/B on the board).
• Grounds are good and fuses/relays appear fine, so the load is being introduced by the ECU or something that only connects when the ECU is mated to the harness.
• This strongly points to an internal ECU fault (power regulator, shorted pass device, shorted diode/MOSFET, bad cap, or a connector/pad problem on the P28 board).

5. What I’ve tried already (recap):

• Checked & replaced fuses (ECU, fuel pump) and reseated/swapped relays
• Unplugged TPS, MAP, IACV (no change)
• Removed S300 when testing the ECU (no change)
• Confirmed grounds and harness continuity for VCC / sensor ground rails

6. Planned next steps (DIY repair path)

• Open ECU and inspect PCB around connector A/B and the power regulation area (looking for lifted pads, cold joints, heat damage, bulging caps, solder bridges).
• Do diode/continuity testing on board traces from A25/B1 to power components to find a low-ohm path to ground.
• Use a series lamp (12 V incandescent), test off a battery with an inline fuse (5 A) to safely apply 12 V to the board and identify the hot/shorted component. I have a Sylvania 7443 bulb available for this.
• Replace/repair the faulted SMT/THT part (likely regulator, diode, MOSFET, or cap) and re-test.

7. Questions to the community:

• Has anyone seen a P28 behave like this (A25/B1 collapsing to ~2 V only when connector A or B is mated)?
• Any suggestions for the most likely components to check first around Connector A/B on a P28?
• If you have a common “spot” or part that fails on these boards after S300 installs, please share (photo/part number/repair tip appreciated).

This is the pinout diagram I'm using for all measurements (02D01720-1500 is the board in my ECU):

 

I have no personal experience with this ECM setup/mod, so wait for further information from those who do, BUT,

As someone who manufactures Class 2 and 3 electronics, if the ECM tested good, you have shorts somewhere in the A or B plug harnesses, or something is not connected properly on those one or two harnesses. That's the common denominator.

Process of elimination using the information you have provided. That's where I would look while you await input from others with specific experience in this case.

 

You are exactly right, I found a short in one inside the car. Took me a few hours of testing but the wire goes from the main relay to connector B. I currently have my dash out and I'm undoing my wiring harness to figure out what wire I need to replace.

 

Good stuff.

Since the main relay essentially is connected to both the A and B harnesses, chances are fixing the short in the B harness will be enough, but it wouldn't hurt to check through the A harness while you're in there.