One more question, concerning the computer:

How did the old engine run, even with low compression on a cylinder? If continued low power output is detected from a cylinder, the computer responds by shutting off the respective injector and coil. It doesn’t want to pump unburned fuel into the exhaust.

The check engine light is also illuminated to get your attention. Believe it or not, the loss of a cylinder isn’t always obvious to the driver. I can vouch for this, both with my Jag (bad coil) and another car (low compression due to a burnt valve)

So this leads me to ask how the old engine was running before replacement. Was it running roughly? Or was the check engine light the only apparent symptom? Even if the engine ran poorly, that isn’t necessarily the computer at fault, so further research is needed. But If running reasonably well, that would point to the computer operating properly.

 

The old engine runs exactly like the new engine.
idles rough, start right up, but after up to temp, you have to crank and crank to start it. You can smell gas all the time. At highway speeds you can’t tell there is a missing cylinder.
it used to get 34-36 mpg, now 17-19.

 

They're just DTCs like any others but the MIL is not on.

Any OBD tool will see them, even the famous elm327.

You've lost your usual touch on this, somewhat.

Unless/until cleared, DTCs are kept so they can be read even with the MIL off, until 40 or so warm ups of consistently not occurring when they are finally auto-deleted. They are kept in case a car is taken in to a tech somewhat tardily.

When the MIL is off many people don't look for codes (quite sensible!) so don't realise they're kept as above.

 

When the newer engine was put in they should have erased all codes. If the new engine has good compression but the similar code(s) then it's maybe coil / wiring / ...

If they maybe didn't clear the codes you can do it and see what then happens.

One snag is that doing so will also clear the OBD monitors and each cleared (unset) one means a bunch of related codes cannot flag (as the PCM is unsure what it can rely on).

 

I just ordered an OBD2, I will have codes tomorrow.

 

Blasphemy!

 

I'm a little confused here. Any time I've disconnected the battery on any OBD II vehicle, that has "cleared" the codes as far as my various paltry scanners can see. It also resets the monitors. In addition, it basically puts the learning process back to square one with factory defaults. Fuel trims, for example, start out at zero but quickly change as there appears to be no accessible history for averaging. Not trying to argue, but that's what I've seen over the years on various vehicles with several inexpensive scanners. I've used this behavior to my advantage for troubleshooting.

As far as the battery, I'd certainly hope it was disconnected for the engine change. Considering the car ran reasonably well previously, even with a dead cylinder, that doesn't point towards a bad computer.

Is anything known of the history of the donor engine? If I had to gamble your hard-earned money, I bet it has one or more marginal/failed coils. Wouldn't it be something if the fault with the donor engine was as simple as a bad coil on #2? That would give the same code as the original engine with low compression. I'd hate to think a shop would fall into that trap, but who knows these days.

Or maybe its something else simple, such as the fuel pressure sensor is inaccurate. In response, the computer is diligently adjusting pump output to match a bad input.

Whatever is going on, I think we could safely put the bad computer idea on a back, back burner for now. Retrieve any existing codes, clear them, and see what returns after several drive cycles. Armed with any active codes, we can plot a plan of attack.

When you report the codes, please give us the 5 character code, such as P0302. Don't rely on the generic definition your code reader may provide. Those definitions aren't always accurate.

 

I will get codes tomorrow, the car has been active (battery not disconnected) for 3 weeks.
this is the OBD I ordered.

 

Battery off should clear learned values such as fuel trims (also accel pedal max/min and so on) and OBD monitors.

It should not, and on 2002.5MY-on does not, clear codes. Maybe the earlier cars which have a very different PCM do not obey that.

The idea is that owners not have an easy way to circumvent OBD codes when presenting a car for an emissions or other test. (Not an issue over here because OBD still isn't used for that. Here they use an actual exhaust gas analyser. Also fail if MIL is on or if self-test MIL, airbag, etc don't work.)

 

The only code was. P1000

 

 

Codes
P1000 system check not completed since last memory clear

B2426 passenger solar radiation sensor circuit open

C1424 Front left damper circuit open.

 

I'm thoroughly confused. If the engine is running roughly, it should be setting codes right away. Does your scanner have a separate function to view pending codes?

The P1000 is normal after the battery has been disconnected or codes cleared. It just means not enough drive cycles have been accomplished for various self tests (monitors) to run automatically. Once that happens, the code will change to P1111. But in the meantime, the computer should still be logging any pending codes.

The other two codes are not related to the engine at all. Don’t worry about them.

 

That is the only code. I away till the 28th. So I can’t check anything until then

 

That's also true with the early models. If you disconnect the battery (or clear codes with a scanner), all monitors will reset and need to run again before you can successfully pass the emissions test. As you said, that's to prevent owners from circumventing active faults.

Several years ago, my '02 logged a new code on the way to the test facility. What horrible timing. It cleared no problem but I had to log a few drive cycles before the monitors were satisfied.

I'm surprised your authorities don't test via the OBD system. That's how its done in Oregon and Washington, and probably most other of the Colonies. Super easy, and no need for the exhaust sniffer.

 

It's harder to fake exhaust gas. Much easier to fake OBD...

Dunno if that's why, though they were probably testing exhausts before OBD II (well, EOBD) came in.

 

You should cancel your trip so we aren't left in suspense.

How much have you driven the car since replacing the engin?. Still baffled why it's not setting codes right away. Perhaps the current rough idle is just below the threshold to set a code? I wonder if more driving would help the learning process so the engine runs better. It might also provide more opportunities to set codes, which will aid greatly in troubleshooting.

Another thought: How does the engine idle with the transmission in neutral or park? Kind of a long shot, but maybe a transmission problem is causing the same symptoms as engine trouble?

 

Most obvious possible reason for not setting codes would be unset OBD monitors. Ah well, 28th will be here soon enough.

 

Hmm, this is interesting. Must be a big difference from the early models.

A while back I was chasing a P0301, eventually confirmed to have been caused by a bad coil (and followed by a bad replacement!). When the codes were cleared, the fault code returned in minutes. This was long before the monitors could complete their cycles.

 

Each unset monitors means a bunch of codes can't flag.

These are for example documented for the various iterations of the Denso PCM used by Jaguar on these cars (2002.5MY-on). The original cars used a cruder PCM and I'm not sure the equivalent doc is available.

But some codes don't rely on a specific (or any) monitor being set. That's what you must have experienced.

To a fair extent you can ask: can the PCM be sure about a fault? If yes, it'll likely flag a code. But if it's drowning in data it's not sure it can trust (e.g. unset monitors and/or contradictory readings) then either it'll guess and flag multiple codes or possibly flag none.

The Denso PCMs have more memory, more I/O and so on, so can do a bit better. And there's doc!!