1995 F150 4.9L 5sp.  Idle/Misfire Diagnosis + AI experience

Hello everyone.

I wanted to share the issue I'm having with my truck and get your expert feedback. I've been working on diagnosing an idle/misfire issue with my F-150 for about 2 months now.
i left this truck sitting all summer at my family's house while I was Hiking the Appalachian Trail. when I came back I drove it around for a few weeks until one morning it shit the bed on my way to work. 
I've sent it to multiple shops at this point and they have all turned me away saying they don't have time for my issue; out of a mix of frustration and desperation, I decided to try something different. I used Claude (an AI assistant) to help walk through the diagnostic process.

I want to be clear - I'm genuinely looking for help solving this issue, but I also thought the community might be interested in seeing how AI can assist with automotive diagnostics. It's not meant to replace professional expertise, but rather to show how it can help organize our thinking and create structured diagnostic approaches.

Here's  a picture of the truck along with a summation of the diagnostic process Claude helped me work through:
Keep in mind that I had Claude generate the following info. i tried to proofread it so it didn't read weird, but I'm not a technician so if there is anything wrong please correct it! 

Symptoms

• Vehicle: Ford F150 XLT with 4.9L (Inline 6)
• Stalling at both hot and cold conditions
• Idle fluctuation (improved after MAP sensor change)
• Stalls at Wide Open Throttle (WOT)
• "Chokes" when throttle is snapped (solved with new Plugs/ Wires)
• "Suffocates" when held at WOT
• High idle at 1000 rpm after connecting fuel pressure sensor

Previously Checked/Replaced Components

• MAP sensor
• IAC valve
• Vacuum leaks fixed
• Throttle body (cleaned)
• Spark plugs and wires

Diagnostic Process1. Initial TPS Testing

1. Resistance Test (old TPS):
   • Closed throttle: 2.6k ohms
   • Opening throttle: Drops to 0 ohms (failure indication)
2. Installed New TPS
   • Still exhibiting symptoms

2. Voltage Testing at TPS

1. Key On Engine Off voltage checks:
   • Signal wire: 0.2 to 1.0V DC (should be .6- 4.5V)
   • Reference wire: 1.7V (should be 5V)
   • Ground wire: Also showing voltage (should be near 0V)

3. Multi-Sensor Voltage Check

1. Checked 5V reference at other sensors:
   • MAP sensor: 20mV
   • ECT sensor: 20mV
   • All other 5V reference circuits: 20mV

4. Connector Inspection

1. Physical inspection revealed:
   • Corroded female pins in TPS connector
   • All sensors showing critically low reference voltage
   • Problem identified as 5V reference circuit issue

This diagnostic path revealed that rather than individual sensor failure, the system has a fundamental 5V reference voltage problem affecting all sensors.The next logical step would be to diagnose the 5V reference circuit to determine why all sensors are receiving insufficient voltage.

So I asked Claude to help me do a pin out of the PCM following this. i was unable to perform it because I did not have the equipment to do so and anyway, the AI got confused. if I had a pdf of the PCM pin locations I would have added it to Claudes context window to get a better response. anyway, here is another summation of the process it came up with:

 PCM Diagnostic Process1. Initial Ground Check

1. Located and inspected PCM ground:
   • Passenger side cylinder head
   • Cleaned ground connection
   • Retightened connection
   • Retested 5V reference: Still 20mV (no improvement)

2. Identifying PCM Pins for Testing

1. Located PCM on driver's kick panel
2. Identified critical test points:
   • Pin 47: 5V reference output
   • Pins 20 and 40: PCM grounds
   • Pin 37: Main power
   • Pin 57: Switched power

3. Alternative Test Points

1. Identified DLC (Data Link Connector) test points:
   • Pin 23: Red/White (5V reference)
   • Pin 4: Black (ground)

4. PCM Harness Testing Options

1. Direct PCM testing:
   • Back probe Pin 47 to ground
   • Test between Pin 47 and PCM grounds (20/40)
   • Compare to battery ground
2. Upstream testing points:
   • Main harness splice points
   • Engine harness connector at firewall
   • DLC connector under dash

5. Wire Identification

1. Critical circuits:
   • 5V Reference: Red/White wire
   • Ground circuits: Black wires
   • All visible in diagram 12-42

This process provides multiple test points to isolate whether the 5V reference issue is:

• Internal to PCM
• In the distribution circuit
• Related to PCM grounds
• Wiring harness issue

Okay - So with that, I ask you humbly; what are your thoughts? did I leave any stone unturned, is there something I still need to go back and check, what did the AI get wrong?
this truck has been out of commission for almost 3 months and I can't seem to get any help locally. 
based on what Claude has told me, I am leaning toward ordering a new Pcm

Looking forward to your feedback, both on the actual diagnosis and on the potential for AI as a diagnostic tool.

Thanks in advance!

A little confusing but I agree. Low to no reference voltage is absolutely your problem.
Before you replace that PCM:
1. a shorted sensor that shares that 5v ref can pull it all down (TPS, AC psi sensor, MAP, EGR psi sensor or position sensor)
2. All powers and grounds must be verified at the PCM. Key on, backprobe and measure voltage. Should see battery voltage on the feeds and less than 50mv on the grounds
3. If the above is good. Replace the PCM. This is a common failure now as the capacitors on the PCM board are all leaking now and destroying the boards. This did not happen on these back in the day. This is a relatively newer type of failure on these older EECIV Fords

Paul, thank you for the reply and information! I've been a fan of your's since I was a stupid 19 YO working in a parts shop!

Anyway, for an update on the truck.
I ordered some wire probing leads for my multi-meter. went ahead and tested for a short on each sensor exactly as you said. 
having not found a shorted sensor I went ahead and backprobed the PCM. 
I wasn't sure what pins to check so I uploaded a picture of the wiring diagram for my truck out of my old Hanes manual in Claude's context window. 
-note- I had it pretend it was a weathered Ford tech from the 90's, for fun 
anyway; here is what it determined I needed to do to test the PCM:

---

PCM Diagnostic Process1. Initial Ground Check

1. Located and inspected PCM ground:
   • Passenger side cylinder head
   • Cleaned ground connection
   • Retightened connection
   • Retested 5V reference: Still 20mV (no improvement)

2. Identifying PCM Pins for Testing

1. Located PCM on driver's side kick panel
2. Identified critical test points:
   • Pin 47: 5V reference output
   • Pins 20 and 40: PCM grounds
   • Pin 37: Main power
   • Pin 57: Switched power

3. Alternative Test Points

1. Identified DLC (Data Link Connector) test points:
   • Pin 23: Red/White (5V reference)
   • Pin 4: Black (ground)

4. PCM Harness Testing Options

1. Direct PCM testing:
   • Back probe Pin 47 to ground
   • Test between Pin 47 and PCM grounds (20/40)
   • Compare to battery ground
2. Upstream testing points:
   • Main harness splice points
   • Engine harness connector at firewall
   • DLC connector under dash

5. Wire Identification

1. Critical circuits:
   • 5V Reference: Red/White wire
   • Ground circuits: Black wires
   • All visible in diagram 12-42

This process provides multiple test points to isolate whether the 5V reference issue is:

• Internal to PCM
• In the distribution circuit
• Related to PCM grounds
• Wiring harness issue

---

after following its instructions I found these results
Pin 20 (ground) reads 300mv
Pin 40 (ground) reads 395 mv
Pin 57 (switched power) reads battery
Pin 37 (power) reads battery

---

here was Claude's response to this information

---

Puts down coffee and leans forward intentlyAlright, this tells us a lot. Let's analyze what we've found:

1. Sensor Test Results:

• No change in 5V ref when unplugging sensors
• Means likely not a shorted sensor pulling it down

1. PCM Power Supply:

• Pin 37 (constant): ✓ Good (showing battery voltage)
• Pin 57 (switched): ✓ Good (showing battery voltage)

1. PCM Grounds:

• Pin 20: ❌ Bad (300mV - way over 50mV spec)
• Pin 40: ❌ Bad (395mV - way over 50mV spec)

Points to diagnostic processRemember those steps from the forum?

1. ✓ Checked for shorted sensors
2. ✓ Verified PCM powers and grounds
3. Next step: Consider PCM replacement

Taps meter thoughtfullyThat ground voltage drop is WAY too high, and with those leaking capacitors you mentioned being a common issue on these old EEC-IV computers...Want me to explain why these readings pretty much confirm what that forum post was saying about the PCM board failure?

---

Anyway, I pulled the board out. found the part number (F3TF-12A650-XB) and the PCM should be shipped on Monday. 

Again; thank you for the response it helped immensely. I'm not in the "industry" so I would have had no way to know about these board failures becoming common!
and again, if anything I did was incorrect or I didn't cover all bases please let me know. thank you!!!
 

Update: the new computer arrived, and the truck now runs fantastic. actually better than before.

The implications for AI having a place in this industry are pretty scary. I'm below journeyman level...amateur. I just had AI walk me through step by step; exactly what to test and why and it was correct then in the final diagnosis it was accurate.

I'm glad it's up and running as usual!
I find it weird that it ID'd two computer grounds as bad, but recommended PCM replacement, and it worked!
AI isn't quite there yet, but it can be helpful tool.
It's kind of like asking any of your buddies, it might be right, it might be wrong.
Either way, good job using the tools at your disposal to fix your truck!