My last idea is that there is a vacuum leak that is compensated for by maxing out the idle screw - this is then messing with MAP readings and causing surge.

 

Blu/Wht should also be 5V. This is a single wire from ECU D17 pin to that map connector. You can leave the map unplugged and try to measure voltage at pin D17 on the ECU to a good body ground. If it is 2.5V at the ECU, then you have a bad driver on the ECU that can't supply enough voltage to the MAP. How the detecting works, you can do that jumping trick like the reference. You might want to unplug the ECU and look for a dirty/corroded/loose pin on the connector and ECU.

If you have 5V at the ECU, then you will need to trace the blu/wht wire like you did with the red/wht for damage, etc between the map sensor and ECU.

 

https://www.vadriven.com/forums/tech...wiring-328453/

On this page it says D17 should be ~3v key on engine off

 

I measured that pin, D17 wht/blue with key on engine off and i'm getting the same reading that I did on the wire on the MAP - about 2.5v

 

I measured it by putting a t-pin in the back of the connector to the ECU on that pin.

 

The ECU is the issue. You can pull the ECU and remove the cover. You can try tracing pin D17 to see if there are any cracked solder joints or obviously burnt components to be 100% sure.

The male pin on the ECU could also be loose or the femaile connector not making good contact.

I'd find a u-pull-it junkyard and find an ECU with the exact OEM part number as yours. You can also look at hondapartsnow.com for the OEM number. Sometimes there are updated versions, so if your ECU is 37820-PT3-L52, a 37820-PT3-L53 should work. I'd reference hondapartsnow.com to be sure.

 

Where did you read that the D17 pin should read 5v? Where I read it said it should be about 3v: https://www.vadriven.com/forums/tech...wiring-328453/

 

I was looking at the diagnostic tree on a 90-93 shop manual I kind of trust, and my 95 accord (with the same setup) that a 100% trust.

Wht/blu should read 5V when the map sensor is unplugged earlier in the diagnostic tree. Along a different path in the diagnostic, the map sensor is plugged in and you should get around 3V as the map sensor has a resistance at atmospheric pressure.

The 5V on the red/wht is the constant voltage input to the map sensor. The blu/wht is the ground wire and grounds through the ECU. Wht/blu is the return wire with varying voltage as the intake pressure changes. The 5V coming out of the ECU with the map unplugged is in the shop manual without explanation. The 5V out with MAP unplugged has to deal with the ECU internals and how the ECU measures the voltage inside the ECU. Since you only have 2.5V coming out, the ECU almost has to have an internal problem on that specific pin and/or the trace on the circuit board.

On a newer OBD2 car (from 1996), I'd suggest to use a scanner to see what values the ECU is reporting on the map to "test" this wire. Unfortunately, your obd1 accord doesn't have this option.

I hate calling a bad ECU. Kind of why I suggested opening up the ECU case to see if something was obvously burned on that trace to pin D17.

You can use that link to test for powers and grounds to the ECU. Anything battery or IG1,2,... KOEO (key on engine off) means key in the II position, but engine doesn't have to be running, should have 12V. Any ground should have zero resistance to ground. The sensor grounds won't because the ECU is the actual ground.

 

The reading I gave earlier was with the sensor plugged in and KOEO, I just measured when the sensor is unplugged KOEO, backpinning D17 reads 4.85v in that scenario and I used the battery as the negative ground for that reading

 

Opened up the ECU and nothing appears to be obviously burned up.

However I broke and took down the throttle body and someone has had some fun with gasketmaker on the throttle body mating surface to the intake manifold. I replaced that gasket and will fire it up today to see if it helped.