B20 Running Poorly
I have a 93 civic with a b20b4 swap, Integra p75 cylinder head, intake manifold, engine harness obd1 and p75 unchipped ecu.
I bought this whole swap from a running and driving Integra, ever since I took it out of the donor car and installed it its been running poorly.
Runs very rich, smoke at startup, idles rough. Seems to idle a lot smoother as soon as it fully warms up and goes into closed loop. No fault codes. Car has full power but stumbles and hesitates in rpms a lot when not fully warmed up.
I've checked timing 3 times it is spot on, I've cleaned the injectors and replaced all the injector seals. The spark plugs are properly gapped at .045. New plugs and plug wires. Everything is connected. It had a top dead center sensor fault, I've replaced the distributor fault is now gone but still runs the same. Injectors and spark plug wiring is orientated correctly. Tested compression all between 120-128 dry on cold engine
What could I be missing??
I bought this whole swap from a running and driving Integra, ever since I took it out of the donor car and installed it its been running poorly.
Runs very rich, smoke at startup, idles rough. Seems to idle a lot smoother as soon as it fully warms up and goes into closed loop. No fault codes. Car has full power but stumbles and hesitates in rpms a lot when not fully warmed up.
I've checked timing 3 times it is spot on, I've cleaned the injectors and replaced all the injector seals. The spark plugs are properly gapped at .045. New plugs and plug wires. Everything is connected. It had a top dead center sensor fault, I've replaced the distributor fault is now gone but still runs the same. Injectors and spark plug wiring is orientated correctly. Tested compression all between 120-128 dry on cold engine
What could I be missing??
Last edited by JDMBart; Apr 9, 2020 at 07:53 PM.
MAP sensor wiring ? Does your MAP sensor wiring run from the MAP sensor all of the way to the ECU through the passenger side shock tower junction ? Did you have an original MAP sensor mounted on the firewall ? Is there still a 3-wire plug there ?
What do you mean by map sensor mounted on the firewall? Yes it Goes through the factory firewall grommet. It is the 3 wire connector i double checked on the Honda wiring schematic and I have the correct colors running to the MAP. No fault codes either.
Some 1992/1993 non-Si/EX civics have a firewall mounted MAP sensor... and the wiring comes through the firewall and goes directly to that sensor. This means that where the MAP sensor wires would connect to the engine harness on the passenger side shock tower, there are three open pins (without wires) in the chassis side plug that do not meet the MAP sensor wires normally found in the EX/Si engine harness that plugs into the MAP sensor found on the top of the throttle body.
Some 1992/1993 non-Si/EX civics have a firewall mounted MAP sensor... and the wiring comes through the firewall and goes directly to that sensor. This means that where the MAP sensor wires would connect to the engine harness on the passenger side shock tower, there are three open pins (without wires) in the chassis side plug that do not meet the MAP sensor wires normally found in the EX/Si engine harness that plugs into the MAP sensor found on the top of the throttle body.
Trending Topics
Do you have a scan tool that can give you live data so that you can monitor IAT and ECT values ? Maybe one or both of these sensors is reading wacky and causing improper fuel trim values... and maybe even a psycho O2 sensor.
Can I do that on OBD1? As far as I knew you have to jump the connector and count how many times the CEL blinks. I am not getting any faults.
I have just recently replaced the fuel pump and 4 injectors with new units. It is still dumping fuel. There is a lot of fuel entering my engine oil and diluting it pretty quickly. My exhaust manifold gets hot fast. I am not sure what causes it to run this rough.
"Seems to idle a lot smoother as soon as it fully warms up"- this makes me think you have a vacuum leak. Good time to learn about Short Term and Long term fuel trims, put a scan tool in live data mode and monitor them SHFT and LTFT and see if they get better at high rpm/speeds and if they do then you have a vacuum leak and confirm either with smoke machine(DIY) or some use water bottle spray to find vacuum leak
My car is a "95 EX with a JDM B20 low compression, P75 cylinder head non-vtec, LS manual transmission. Phearable 2.0L basemap ECU (o2 sensor delete).
My engine puts out 190 psi for each cylinder when I did a dry compression check last year. That's down from around 205 psi per cylinder when I bought it 4 years ago.
Try adjusting the valve's if you haven't within the last 30k miles. Also repeat a dry and add a wet compression the following way:
1) trickle charge battery for 12 hours or more before and during compression test. weak battery equals weak compression during cranking
2) add oil to whip's threads prior to inserting into spark plug holes. this ensures lo air loss and helps prevent whip from getting stuck in spark plug holes.
3) during testing don't use a set number of cranks. instead crank engine to max psi reading and then go a few cranks past the peak reading to ensure the exact peak psi number.
4) after dry test do wet test with 2 teaspoons of oil in each cylinder. write down all numbers both wet and dry..they should not vary from cylinder to cylinder by more than 20% per respective test (dry compared to dry)(wet compared to wet).
Other possibilities may be
1) weak or incorrect o2 sensor
2) ecu issue
When I purchased my car (prior owner did great job with swaps) it came with a P75 ecu. It ran well other than it would idle lower (600-/+ rpms). About a year and half after purchase the engine would intermittently cut off due to low idle at stops. I purchased that Hondash unit (free software download) and with the help of a 20+ year Honda dealer mechanic on a another site I tested everything (always test first before parts, if possible). I've never owned a scanner prior to that point so I would youtube him live data and recorded data of various sensor (ex: tps, air intake, temp sensor, other) readings, fuel trims readings, engine temp, etc.. I checked mechanical, ignition timing, checked valves (all were spot on), adjusted idle screw per service manual, cleaned IACV, cleaned TB, compression check, smoke tested for vacuum leak, replaced o2 sensor.
After confirming everything listed above checked out fine he concluded the ecu must be the issue. I researched a lot about B20/LS swaps and there were more than few people on various Civic forums mentioning idle issues with a B20 swap when using an obd1 that is calibrated for a 1.8 liter engine. Some people in those threads think the .2 liter difference (from 1.8 to 2.0) is enough to confuse readings between the o2 sensor and a 1.8 liter ecu. I talked this over with the owner of Phearable: http://www.phearable.net/ and he said it's plausible but unlikely the .2 liters would make a difference concerning idle. I asked him if he could create a 2.0L basemap ecu and he said yes, no problem. I also asked him to program it for an o2 sensor delete. At the time I lived in Florida where there are no vehicle inspections. I installed the basemap ECU he sent and it has been idling great since (for past 2.5 years) at a constant 750-/+ . Unfortunately, the Hondash doesn't work with the Phearable ecu.
Thanks for the response! Yes I am interested in the scanner
I have previously checked for vacuum leaks with no issues found. I was thinking of doing an o2 sensor however I am not thinking that is the issue because it idles poorly and very rich as soon as the engine is started in open loop. The sensor is not warm enough at this time to read and the computer is using a preset value. I was thinking about trying another test ecu but haven't been able to find another p75 local.
I have previously checked for vacuum leaks with no issues found. I was thinking of doing an o2 sensor however I am not thinking that is the issue because it idles poorly and very rich as soon as the engine is started in open loop. The sensor is not warm enough at this time to read and the computer is using a preset value. I was thinking about trying another test ecu but haven't been able to find another p75 local.
Yes, that is how to read and obd1 Civic. Hondash is an obd1 bluetooth scanner that works with many stock honda obd1 ecu's including a P75 ecu. This link has some info..my past names here have been Jimi Hondrix and Megalodong: https://honda-tech.com/forums/honda-...-tool-3299758/ I still have the scanner (version 2.0) for sale if you are interested? Will sell it for $20 (shipping not included). Here's the Hondash site: https://www.hondash.net/p/e-shop.html and the conversion price on that site is US $32 (not including shipping).
My car is a "95 EX with a JDM B20 low compression, P75 cylinder head non-vtec, LS manual transmission. Phearable 2.0L basemap ECU (o2 sensor delete).
My engine puts out 190 psi for each cylinder when I did a dry compression check last year. That's down from around 205 psi per cylinder when I bought it 4 years ago.
Try adjusting the valve's if you haven't within the last 30k miles. Also repeat a dry and add a wet compression the following way:
1) trickle charge battery for 12 hours or more before and during compression test. weak battery equals weak compression during cranking
2) add oil to whip's threads prior to inserting into spark plug holes. this ensures lo air loss and helps prevent whip from getting stuck in spark plug holes.
3) during testing don't use a set number of cranks. instead crank engine to max psi reading and then go a few cranks past the peak reading to ensure the exact peak psi number.
4) after dry test do wet test with 2 teaspoons of oil in each cylinder. write down all numbers both wet and dry..they should not vary from cylinder to cylinder by more than 20% per respective test (dry compared to dry)(wet compared to wet).
Other possibilities may be
1) weak or incorrect o2 sensor
2) ecu issue
When I purchased my car (prior owner did great job with swaps) it came with a P75 ecu. It ran well other than it would idle lower (600-/+ rpms). About a year and half after purchase the engine would intermittently cut off due to low idle at stops. I purchased that Hondash unit (free software download) and with the help of a 20+ year Honda dealer mechanic on a another site I tested everything (always test first before parts, if possible). I've never owned a scanner prior to that point so I would youtube him live data and recorded data of various sensor (ex: tps, air intake, temp sensor, other) readings, fuel trims readings, engine temp, etc.. I checked mechanical, ignition timing, checked valves (all were spot on), adjusted idle screw per service manual, cleaned IACV, cleaned TB, compression check, smoke tested for vacuum leak, replaced o2 sensor.
After confirming everything listed above checked out fine he concluded the ecu must be the issue. I researched a lot about B20/LS swaps and there were more than few people on various Civic forums mentioning idle issues with a B20 swap when using an obd1 that is calibrated for a 1.8 liter engine. Some people in those threads think the .2 liter difference (from 1.8 to 2.0) is enough to confuse readings between the o2 sensor and a 1.8 liter ecu. I talked this over with the owner of Phearable: http://www.phearable.net/ and he said it's plausible but unlikely the .2 liters would make a difference concerning idle. I asked him if he could create a 2.0L basemap ecu and he said yes, no problem. I also asked him to program it for an o2 sensor delete. At the time I lived in Florida where there are no vehicle inspections. I installed the basemap ECU he sent and it has been idling great since (for past 2.5 years) at a constant 750-/+ . Unfortunately, the Hondash doesn't work with the Phearable ecu.
My car is a "95 EX with a JDM B20 low compression, P75 cylinder head non-vtec, LS manual transmission. Phearable 2.0L basemap ECU (o2 sensor delete).
My engine puts out 190 psi for each cylinder when I did a dry compression check last year. That's down from around 205 psi per cylinder when I bought it 4 years ago.
Try adjusting the valve's if you haven't within the last 30k miles. Also repeat a dry and add a wet compression the following way:
1) trickle charge battery for 12 hours or more before and during compression test. weak battery equals weak compression during cranking
2) add oil to whip's threads prior to inserting into spark plug holes. this ensures lo air loss and helps prevent whip from getting stuck in spark plug holes.
3) during testing don't use a set number of cranks. instead crank engine to max psi reading and then go a few cranks past the peak reading to ensure the exact peak psi number.
4) after dry test do wet test with 2 teaspoons of oil in each cylinder. write down all numbers both wet and dry..they should not vary from cylinder to cylinder by more than 20% per respective test (dry compared to dry)(wet compared to wet).
Other possibilities may be
1) weak or incorrect o2 sensor
2) ecu issue
When I purchased my car (prior owner did great job with swaps) it came with a P75 ecu. It ran well other than it would idle lower (600-/+ rpms). About a year and half after purchase the engine would intermittently cut off due to low idle at stops. I purchased that Hondash unit (free software download) and with the help of a 20+ year Honda dealer mechanic on a another site I tested everything (always test first before parts, if possible). I've never owned a scanner prior to that point so I would youtube him live data and recorded data of various sensor (ex: tps, air intake, temp sensor, other) readings, fuel trims readings, engine temp, etc.. I checked mechanical, ignition timing, checked valves (all were spot on), adjusted idle screw per service manual, cleaned IACV, cleaned TB, compression check, smoke tested for vacuum leak, replaced o2 sensor.
After confirming everything listed above checked out fine he concluded the ecu must be the issue. I researched a lot about B20/LS swaps and there were more than few people on various Civic forums mentioning idle issues with a B20 swap when using an obd1 that is calibrated for a 1.8 liter engine. Some people in those threads think the .2 liter difference (from 1.8 to 2.0) is enough to confuse readings between the o2 sensor and a 1.8 liter ecu. I talked this over with the owner of Phearable: http://www.phearable.net/ and he said it's plausible but unlikely the .2 liters would make a difference concerning idle. I asked him if he could create a 2.0L basemap ecu and he said yes, no problem. I also asked him to program it for an o2 sensor delete. At the time I lived in Florida where there are no vehicle inspections. I installed the basemap ECU he sent and it has been idling great since (for past 2.5 years) at a constant 750-/+ . Unfortunately, the Hondash doesn't work with the Phearable ecu.
Hey JR what you said about the map being firewall mounted holds true. I have a 93 dx and I have a firewall mounted map that is connected. I also have the map on my integra throttle body that is connected with the integra engine harness. I don't think that this is connected properly though because the ecm isn't sending power or ground to the MAP on the throttle body. Any tips on how to wire it in?
Hey JR what you said about the map being firewall mounted holds true. I have a 93 dx and I have a firewall mounted map that is connected. I also have the map on my integra throttle body that is connected with the integra engine harness. I don't think that this is connected properly though because the ecm isn't sending power or ground to the MAP on the throttle body. Any tips on how to wire it in?
Please report your findings.
Thanks for the response! Yes I am interested in the scanner
I have previously checked for vacuum leaks with no issues found. I was thinking of doing an o2 sensor however I am not thinking that is the issue because it idles poorly and very rich as soon as the engine is started in open loop. The sensor is not warm enough at this time to read and the computer is using a preset value. I was thinking about trying another test ecu but haven't been able to find another p75 local.
I have previously checked for vacuum leaks with no issues found. I was thinking of doing an o2 sensor however I am not thinking that is the issue because it idles poorly and very rich as soon as the engine is started in open loop. The sensor is not warm enough at this time to read and the computer is using a preset value. I was thinking about trying another test ecu but haven't been able to find another p75 local.
You shouldn't have to wire anything. First, use a multi-meter or volt meter to check continuity or voltage between the ECU and both the MAP sensor plug on the firewall and the one on the top of the throttle body. My suspicion is that the one that is actually connected to the ECU is the one on the firewall. If so, do you still have the MAP sensor mounted to the firewall ? If so, simply connect a vacuum line between the MAP sensor and a vacuum port on the intake manifold. Crank the car and see how it runs. If you do not have the factory firewall mounted MAP sensor, you will need to connect the firewall mounted MAP sensor plug over to the MAP sensor on the top of the throttle body. If the wiring isn't long enough, then cut one wire at a time and lengthen it by the amount you feel is needed to both work and appeal aesthetically so the wires are routed in a such a way that you are happy. Once all three wires are lengthened, connect the plug to the MAP sensor and crank the car.
Please report your findings.
Please report your findings.
Thread
Thread Starter
Forum
Replies
Last Post
supracrx
Honda CRX / EF Civic (1988 - 1991)
4
Oct 29, 2006 10:00 PM








