So Ive asked about this before but I just did a test of the problem where I suspected it. My cars been hit or miss with performance and behavior. So id fix something (like the ignition coil) and the car would do great for a while and then it goes back to hit or miss behavior.

I suspect an issue with the map sensor. I just did a map sensor test based on the FSM. The test said to measure the signal wire to ground and the reference wire to ground. This is with the ignition on and the connector disconnected from the map. What I read was 5 volts at the reference and ground wires like I should but at the signal and ground wire I read 4.66 volts. Its not a lot but its a difference considering .5 volts is about 5 in/hg. FSM said to check at the ECU. So I found the same color wire and it read the exact same voltage at the ECU harness.

This would indicate a bad component in the ECU right?

So Im thinking I should replace the ECU. Any ideas or ways I can verify this issue?

Thanks

 

Is your meter that accurate? That's only a 7% meter error if it's reading 4.67 on a 5v line. Maybe check the meter accuracy on a USB output? Computer power supplies are spec'ed for ±5%, although in my experience they're much tighter under light load.

 

The reference and ground wires measured 4.99 volts. I double checked with the reference wire and battery terminal and that read 5 volts. (.01 volt drop through the wires to the chassis ground) So the multimeter is pretty accurate. Ive compared it with other measurements too. This is a digital one. The voltage from the ECU MAP signal wire is definitely off by .34 volts.

 

OK. The lack of decimal places threw me.

So there's 4.99v between VCC1 and SG1. What's the voltage between MAP output and SG1 with the sensor connected?

 

Ill have to check. It was 3.23 or something. I remember because I had read the directions wrong and was confused as to why I got that reading. Then I saw to disconnect the plug. I will check it again though and update this.

Bare with me please cuz I have a theory...
It may be relevant

if there is a vacuum leak, would the map sensor/ecu adjust to maintain the normal vacuum? I induce a vacuum leak and both my vacuum gauge and map sensor readings show a normal vacuum of 23 in/hg. A vacuum gauge should read low if there is a leak. Or is that just on MAF engines?

Which is another thing that makes me suspect an issue, Ive brought this up before but I still dont know why my engine reads about 23 in/hg. High vacuum is healthy but the timing isnt that advanced and normal engines should be between 15-22 in/hg.

Is that a normal reaction for the engine or does that add to my theory? The ECU map signal wire is reading funny and adjusting the engine to what it sees.

Im gonna look for a junkyard ECU and swap them out to test at least.

 

Here are some photos. The ground to reference, ground to signal and then ground to signal plugged into the map sensor all with the ignition on/engine off. I compared these readings to the battery ground and the difference is about -.01 volts. The ECU reads the same.
I also have a video of just showing the engine revving, inducing a vacuum leak, etc, with my vacuum gauge and multimeter hooked up. The behavior is roughly normal but Im still not sure. You may be able to notice something from it. My biggest concern is the voltage difference at the ECU. I tried looking for the L02 ECU and it is almost impossible to find at junkyard price.

Video- sorry I turned it sideways, try to listen to the engine too, something funny is going on
VIDEO0005 - YouTube

ground to reference- 3p connector disconnected, ignition on/engine off


ground to signal- 3p connector disconnected, ignition on/engine off (reads the same at the ecu)


ground to signal with 3p connected, ignition on/engine off

 

The FSM says if ~3v between MAP and SG1 then it's a bad ECU. I think 2.85v qualifies.

A P0H-L01 should work too, from a '96.

 

Yeah, I was looking around for the 96 or 97. My car is really tough to find. Either its obd1 or its a b2 engine or its an lx. I went to two junkyards and couldnt find a 5 speed f22b1.
I found this thread though. Can this be done if I get an f22b1 auto L52?
The official auto to manual ECU conversion thread - Honda-Tech

Also, from the looks of it, does it seem like my ecu is messed up?

 

The last post in the thread indicates that it will work on an OBD2 Accord:

P0J is the non-VTEC '96-'97 ECU but I'm guessing the P0H has the same board layout. Worth a shot. P0H-L51 or L52 is the automatic ECU and they're cheap and plentiful in boneyards.

 

I did find a lot of p0h-L5* 's but didnt wanna waste my money if I was gonna have an issue with it. It may come down to an L5* ECU but Im hoping to find exactly what my car should have first. Theyre $27+core charge. Im gonna call around, there are about 4 or 5 junkyards within an hour of here.