A mechanic installed an F23A4 motor in my 96 Accord LX with automatic and did not tell me he was going to leave the old ECU in. I see now that the VTEC solenoid is unplugged. The engine runs, but roughly at idle and doesn't seem to get great gas mileage. With the spool valve unhooked is it running in low rpm or high rpm optimization? Also, what would it take to get the motor to run as a true VTEC? I've heard about 'boost buttons' inside the car etc. but that those don't factor in all the ECU does for tuning. Would a rewiring and new 32 bit ECU be required? And what part number if you know would be most helpful.

 

the mechanic you took it to should really be the one to set this right. He installed an engine that does not work correctly. i would assume you gave him a fair amount of money to do this for you?

here is some unorganized information that may help:

"boost buttons" only do half of the job. they actuate the mechanical part of VTEC but not the electronic part (ECU "tuning" or fuel and ignition parameters are different when VTEC is active)

a VTEC engine was available for the 96 model year in the EX (F22B1). I'm not certain, but the ECU may be plug and play with the exception of running the wires into the engine bay for VTEC specifics.

The F23 VTEC is a bit different from most. at low RPM it essentially operates only one intake valve to get more swirl in the cylinder. VTEC activation is "available" between 2500 and 4000 rpm depending on things like coolant temp, and throttle position. it's always on above 4000 rpm. so as you have it now, you're only operating "correctly" between idle and 2500 at wide open throttle and up to 4000 at partial throttle.

was you're original distributor used? i don't think the F23 distributor it's compatible with the stock ECU.

the idea i think is to get the engine set up so that the ECU from a 96 EX thinks it's connected to a 96 EX VTEC engine (F22B1) by replicating it's sensor arrangement. that will get you close to 100% OK. VTEC engagement will still be a little off for the cam in your motor... but you may be able to live with that.

depending on the emissions requirements in your area a chipped ECU may be an option, but will also require some wiring modification and at least street tuning session(s)

 

Thanks for the reply, hondamark35. Firstoff let me caution anyone who's thinking of buying a JDM engine off ebay or craigslist or other to ask lots of questions. If it's going to be shipped to you verify that the motor being sent is the one in the picture or get a picture of the one being sent, ergo you might receive one that's been under seawater for months or otherwise not in as good condition. Make sure they're selling you the entire motor and not just the block and head, ask for a serial number so you can check the year and origin and get a compression reading for all cylinders. If you decide to have a local shop install the motor make sure they agree to hook the thing up properly in its entirety, not just VTEC, but recharging the AC and such, upgrade the ECU and wiring and alternator if necessary, and if you have qualms about going from a 2 port to single port exhaust manifold work out the arrangements in advance (e.g. aftermarket 4-2-1 header). Ask every question you can think of no matter how inane they seem, like is a core exchange required, and if other conditions arise within the warranty period like coolant leaks or battery charging problems or even windshield wipers or AC quits working, and rattles and noises, are those covered.

I inquired about a motor for purchase and asked about installation and was quoted a price to install not an F22B2 like original but an 'upgrade' to an F23A4 VTEC with slightly more horsepower and overall a better engine. I thought great, the car could use a few extra HP. I asked if they kept the old engines and was told that they scrap them and since my F22B2 had been well cared for but had lower compression on cylinders 3 and 4 I asked to have it back when they were done and they said sure. I have another nearly identical 96 accord in my driveway with the same F22B2 motor but needing a headgasket job so it was in my interest to keep the old motor. I was not explicit about getting back a complete motor and got only the block and head, as they reused the intake manifold, saying it was better suited for this application (I read that it was a less restrictive intake) and reused the exhaust manifold as the 96 accord LX has a single headpipe and manifold port. As I requested they installed a new timing belt kit I provided. They also reused the power steering pump, alternator, AC compressor and distributor. Had I just ordered the F23A4 it would have shipped as a complete engine. Please understand time was of the essence as the car was needed for daily commuting to work and when picking it up a short test drive confirmed that it was firing on all four and when I asked for a compression test they did one on the spot and each cylinder read 180 psi. I thought fine, well enough. Then a few days ago realized the old ECU was still installed the VTEC solenoid was connected to nothing.

I had the same thought as you that perhaps the 96 LX ECU is already pinned but not wired for VTEC, however I'll have to yank the carpet away from the module to read its part number and then research that. I was thinking about just installing a rocker switch as a low cost solution for engaging VTEC during highway driving but that could be disasterous since maybe the 'warp drive' should not be engaged at idle, and only with sufficient oil pressure and rpms. Some accord motors VTEC solenoid are routed thru an oil pressure sensor as well and I did not find that on the F23A4. And if I did engage it how would the car function at lower RPMs. My F22B2 was quite happy with all 4 valves opening wide at all RPMs but it was a simpler motor and familiar to the ECU. I also understand perhaps the need for swapping out the alternator, maybe, as the 96 VTEC motor uses a higher amperage one although the wiring to the ECU seems identical.

One one side of the coin I'm not satisfied with the end results and want that engine performing as intended and the rough idle remedied. On the flip side my F22B2 with 122,000 miles was well due for a timing belt and head job which would've cost about the same (a grand) for the entire motor and swap. Most interesting here in northern VA that most mechanics I talked to would much rather replace than repair a honda motor unless it's only a timing belt kit, and that's around $500.

Sometimes it's a long lonely trek but if you want something done right, do it yourself. Attend Google U if necessary, but arm yourself with the necessary tools and information.

 

Happy new year y'all. I feel optimistic about 2018 and hope it's a good prosperous one for all. It's been a few weeks since I last reported about this motor swap and wanted to let the fine people of this valuable forum know what I found out. Firstoff, big cudos to ROADSTER for his fine and excellently documented, almost identical motor swap he already performed on his accord wagon at this other forum:

https://www.hondaaccordforum.com/for...gen-5-a-47757/

Using this and lots of independent research I reviewed as much info on honda ECUs as I could find and yikes, there's a lot. I did not find a single concise list of all of them, only subsets. I searched "honda ECU' on ebay and came up with over 12,500 hits. "Honda accord ECU" still produced thousands. Not all identical of course but speed reading thru the condensed list showed me what was popular and available. Based on my 96 accord LX with auto and formerly an F22B2 upgraded to frankenmotor F23A4 I opted for ECU model 37820-P0H-L52. Decoding this number is as follows and correct me if I'm wrong;

37820 are all honda ECUs
P "zero" H means it's a 96-97 VTEC ECU with VTEC support (signal pin A8 for VTEC solenoid is active) and is ODB2 (the ODB1 version is the P "zero" A or P0A).
L is California spec (most common for sale, A is america, J japan etc.)
5 means automatic transmission (0 "zero" means manual transmission)
2 is revision number. The POH comes in an L51, L52 and L53 model for automatics from what I can find. The only difference I gleaned was a mention that the later revision L52 smoothed out some transmission anomalies. If they're anything like windows updates, you just install them and hope whatever beasties lurking inside are now quarantined or fixed.

BTW top contenders for a VTEC motor swap based on popularity were the P0A and P28, but I stuck with this P0H cause it'd save me the cost of a OBD2 to OBD1 wiring harness adaptor. I want this to be a simple good swap but read that the ODB1 ECUs are easier to chip or reprogram. I was under the impression that these older ECUs just reset themselves if battery power is removed for a few moments, then recalibrate F/A ratio as you drive. The newer ones circa 2000 and up I see even have bluetooth interfaces on them for using phone smart apps for reprogramming.

Finding a suitable schematic and ECU pinout of a 96 or 97 accord, VTEC or not was impossible. On the internet I found partials and fuzzy unreadable wiring diagrams. And lots of 'me too' schematics for ford chevy, toyota and whatever slipped thru the browser filter. I even have three after market service manuals here (the popular ones I won't mention by name) and their wiring diagrams were fraught with errors as if someone did wire tracing during a teardown and took notes. E.g. ECU pin C6 used in more than one place, or completely different ECU pinouts and color codes between models that don't seem to match anything. This unraveling I augmented when my ebay ECU arrived and was able to look at the sheared off connector wires. First question was if this is an ODB2a ECU, then why does it have all A, B, C and D connectors in use? The label with VIN on it says it came out of a 97 accord 2 door coupe with automatic and 2.2l VTEC (obviously an F22B1). I have yet to pull up the carpet on my car and compare wiring but that's the next step. I did find one mention that the B connector was used on certain Integras and civics, but what a dilemma, as the fuel injectors should be A1 A2 A3 and A4 but on certain models / years / versions would be on the B connector. The 90s apparently were a renaissance for ECU and ODB development as every year seemed to incorporate more changes. When I went to trace wires and look for VTEC pins, I noticed C15 for the oil pressure switch was unpinned. (???) Then last night I found some thread that said around 2000 honda did away with monitoring that VTEC pressure sender switch (and apparently as early as 1997), as they had found the same unpinned C15. I will verify the VTEC solenoid does engage with proper RPMs, speed and engine temp on a road test before buttoning everything back up.

If you look at ODB connector images it's obvious that ODB1 are two row pin connectors, but ODB2a and ODB2b are supposed to use A, C and D or A, B and D respectively, correct? Seeing all for in use is still a mystery to me but I've read enough about a plain swapout with the added VTEC wire being a success then I'll table my concerns. On the same day as my ECU arrived I also received a pair of connectors with 6" pigtails from Ebay. One for the VTEC solenoid, one for the pressure switch, the pair cost me a mere $6.50 and got here in about a week. The next challenge was how to get wires from them thru the firewall to the ECU under the passenger's feet up front. It turns out that the F23A motor also did away with the second temperature switch near the bleeder valve on the left rear of the head. This 2-wire switch I learned from previous experience was for 'hail mary' overtemp recovery whereas if the coolant temp exceeds its spec, then this second switch does a forced turn on of both radiator and AC fans in parallel, and will continue to run even with the ignition off until the overtemp is reduced. I was tempted to add the cooling jacket fitting to reinstall this sensor but figured if honda is satisfied with using ECT sensor or such then so am I, plus those two unused temp switch wires do terminate on a connector just below the glove box where a cigarette pack size fan control unit resides. Snipping one of those wires off and running it on to the ECU eliminates any need for fishing more wires thru the firewall and up to the VTEC valve. I also figured that since there is no need for a oil pressure signal to the ECU then why not route the solenoid thru the oil pressure switch first and then back to pin A8. After all, the idea was for strong oil pressure being present before the hydraulic actuators can be engaged. Makes sense to me, with the only drawback being if the oil pressure switch fails, VTEC high lift stops as well.

The mechanic who installed my 'frankenmotor' reused the F22B2 intake manifold and eliminated the need for any IACV adaptor or interference fitting. He also reused the F22 distributor and exhaust manifold and my newly replaced F22 alternator. I hope this is not a problem as at the autoparts store, the VTEC alternator is different and higher amperage, different connector too but hopefully the same pulley. The F22 alternator has been running fine for a month on the F23 motor so the only nit may be some incongruity with the VTEC ECU. I'm of the understanding that voltage regulation is not done on the alternator or on a firewall mounted regulator but in the ECU and one pin called ELU does monitor the voltages. I'm hoping the old F22B2 intake manifold is not too restrictive compared to the GEN6 F23A1 that some upgrade to, and figure it's probably better than the ULEV F23A4 manifold that came off. Either way we're talking just a few HP and methinks once VTEC is operational the car will be quite satisfactory for around town and highway.

So wish me luck. Next week when the deep freeze ends I swap in the new P0H and add a wire for VTEC. The old F22 motor with leaky valves on cyl 3 and 4 was still getting 31 mpg and ran fine off idle (static compression test wet or dry showed 175, 175, 140, 140 PSI). The F23A so far has a rougher idle than normal, gulps gas at about 20 MPG and loses 'oomph' around 4000 rpms, and well it should since intake valves do not open at a full 10mm but 7mm and 1.7mm. BTW I read both sides of the concern that it's OK to run a VTEC motor in non-VTEC mode, and others say it's not something you want to do for very long. I could see drawbacks of carbon buildup or too much gas washing oil off piston rings if the non-VTEC ECU profile is too rich or such. After this puppy is up and running fine, I am anxious to remove the head from my old F22B2 and find out exactly why cyl 3 and 4 had lossy compression. Hoo boy did I try every type of fuel system and oil treatment, readjusted the valves several times, replaced the carboned up fuel injectors with new borg-warner ones and even stuck a flexible USB camera with illumination into each cylinder thru sparkplug hole and intake ports above the valves looking for clues or cracks and saw only some carbon buildup. Might be a crack or headgasket blew a via between 3 and 4, dunno yet. I even tried the compressed air into the sparkplug hole and listened for hiss coming from exhaust, intake, oil filler hole and bubbles in radiator. All I could conclude from that was all cylinders hissed and none came thru the oil filler or radiator. (and make durn sure the piston is at BDC before you apply air pressure! :-)

OK, coffee break's over, back on me 'ed.

 

Addendum:

Installed the P0H ECU today and zoom, started right up and idled beautifully, and no CEL. I added an LED thru a 1k ohm resistor to pin A8 so I could monitor VTEC solenoid operation. Interesting that around 2500 rpm or if the gas pedal is mashed it'll kick in, but on the highway the car can cruise steadily at about 40 mph before VTEC kicks in. The car has some serious giddy-up now. The ECU wiring using all 4 OBD connectors is because for 96-97 the accord with auto trans uses an integrated ECU and PSU transmission control. Documentation I'd see previous showed them as two separate boxes but once I peeled the carpet back it looked identical to my replacement ECU. Also, there is no oil pressure switch on the F23A VTEC valve, so just a single wire is needed. I topped up the gas tank and reset the trip meter to see if mileage improves and think it will, now that ECU matches the valve train. I also found out that the P0H ECU for automatics has at least L51-L54 versions for what that's worth. And I might add that VTEC operation at moderate acceleration is seamless, no noticible noise increase or jump, just that the top end acceleration is much better.

 

This is an amazing thread and I appreciate the info . I have some issues on a 99 accord ulev model and I been curious about converting back to obd1 due to some stupid mistakes made during the in frame I completed a week ago .