3500 Swap Finally Underway, a few questions

[CNCguy]

Either your MAP sensor is bad or the wiring was not right. We use the new style MAP with OBD-I all the time without issues. Terminals A-B-C serve the same function on both sensors.
A = Ground
B = Signal
C = 5vt

Did you get a chance to make sure the knock sensor was properly grounded? Have any banging that could be detected as knock like exhaust pipe, metal to metal contact, etc? That 3500 tune should run but has rich/lean spots, knock under WOT and decel/idle issues. Our customer paid a fairly well known company (I won't name names) for that tune, used it for over a year and even drove the car to Vegas from Dayton a couple of times without killing anything.

[jmgtp]

I thought I could use the new map. The thing is that the new map didn't have the ABC on it. Instead it had a sticker that says 5v Vout and Ground and I took the sticker at face value, meaning that I corresponded the order of the 3 wires on the sticker to the pins on the sensor. Is this correct? That is what I did, I double and triple checked it. Things did not run right until I installed the old style sensor.

I ground the KS by clipping an alligator clip to it and grounded the other end of the clip directly to the negative cable. I don't know if that was a good test but I still have the same problems. I ran a new wire all the way from the ECM pin to the sensor so it isn't the wire. The only thing left would be to pull the sensor and chase the threads in the block. To be honest, the reason I haven't done it is because I didn't want to drain the coolant.

The engine is making some unfriendly mechanical noises that may be the cause of the knock counts. BUT, the noises sound like they are the effect rather than the cause. Meaning, it just sounds like an engine that is running poorly. It is loping a bit at idle. The exhaust manifolds seem to be leaking where they mate to the heads. I bought new gaskets today but haven't installed yet.


Datalog: idle with 3500 bin.zip
This is the engine idling with the 3500 bin that Marc spoke about above. I can't say if it runs better worse than the dohc code. But I can say for fact I wouldn't think of taking the car on the street running like this! The idle is crap. It struggles a bit under light throttle, but seems to rev powerfully under heavy throttle - in fact that is the only time the engine sounds healthy. I'm sure the customer that paid for this from whatever company had a better running car. My code 43 must be haunting me. The engine doesn't have the dogbones in securing it to the radiator support and it bounces back and forth a good inch or so as the engine lopes along.

Would anyone like to long distance tune for me? This would be the type of thing where I send you a tunerpro log, you tweak a bin and send to me, I burn the bin to the chip and datalog again and back and forth we go. I WILL PAY YOU. Yes, I know I need to get the knock sensor issue worked out before tuning. I also have to hook up the o2 sensor and get an intake on the motor.

[CNCguy]

John,
The terminal letters are on the mating plug. If you hold the plugs next to each other with the latch to the top, the terminals on both plugs are in the same positions. Looking at the harness plug from the end with the wires and the lock facing upwards the terminals from right to left are A B C.

If the ground was a problem it seems that grounding the KS body would have cleared up the problem. Must be elsewhere. Maybe it is just false knock from the engine bouncing around and something hitting. You should be getting the bracket and mounts pretty soon. Have you tried just unplugging the KS?

The tune sucks but not that bad. It would sit and idle fine. The problem was when transitioning back into the idle table.

Just a thought, but I wonder if having the O2 unhooked would cause the ECM to go into limp mode, thereby bypassing the "tune".

[jmgtp]

I hadn't really considered that not having the o2 sensor connected would impact it and possibly force the ecm into a limp home mode. I assumed it would just run in open loop and not cause any problems. It is next on the list of things to do along with replacing the exhaust manifold gaskets - I know for fact that the front is leaking, not sure on the rear.

Once that is done all the car requires to be mechanically complete is the dogbone mounts and an intake setup. I have some 3" aluminum tubes/bends so I really just need to pickup a filter. The 3.1 airbox I have won't work since it is for a genII motor it has the hose connection in the wrong spot compared to a 3x00 airbox.

I'll take a look at the OBD II MAP connector to see if the ABC lines up with how I wired it. It must not have otherwise I wouldn't have had the MAP troubles. If it turns out that I did it wrong I'll rewire it back in since it is a much cleaner install than how I rigged the old style MAP.

Next time I get to work on it is Saturday.

edit: forgot to mention on that last datalog that I did, the KS was completely disconnected. I believe it did not register any knock counts other than the glitches in the data and I was using a good bit of spark advance. Code 43 was present, as would be expected with a disconnected KS. I am running the 3.4 dohc injectors, is it possible that they are too big? Off hand I don't know what the 3.4dohc injectors are rated. I'm using them from a 97 and I want to say 24lbs sounds right. Marc, do you know what size injector the 3500 bin is using? Perhaps I do have mechanical issue causing absurd amounts of knock. I've mentioned before that the engine lopes a bit and doesn't sound like a healthy running motor. And, it has a clunky sound to it which I think I'm hearing the leaking exhaust manifolds since it doesn't sound like a rocker arm to me (its not bottom end either), but could it be possible that the valve lash is too loose and the KS picking it up? The last time I set valve lash on a motor was on my LT1 and on that motor you have to set preload. There are a few different methods that people seem to use to do it, but the only one that got my motor running right was doing it by feel/sound with the motor running without valve covers. On the LX9 the rocker arms are already mounted to the pedestal in a fixed position and you just torque the pedestal down (I don't remember the torque spec off hand but I know I used my torque wrench). So, could the valve lash be off somehow and causing my issues? Or, since I torqued them down to spec, can I cross valves lash of the list of suspects?

Last thought, I replaced the LIM gaskets. I'm 99.99% sure I got all the pushrods in the right places. I know if you screw that up you bend valves - it should be making a hell of a valvetrain racket if that was the cause right? What is weird about this is that it seems like it is just at idle and small throttle angles that it runs poorly. If you goose it, it revs beautifully, doesn't make any clunking sounds and smooths right out. I still think it is an ecm issue but we will see what happens once the o2 sensor is working, the exhaust leak corrected, and the dogbones in place.

[CNCguy]

In my notes I have 24# @ 55 psi for the LX9 and 21# @ 43.5 psi for a '97 3.4 DOHC. If those are right, it is probably on the lean side. If you had an AFPR, you could try bumping the pressure up to see if it would help. Getting the O2 in there might make it smooth out once it goes into closed loop.

The lash isn't adjustable and one would assume it is right if all the parts are right and installed correctly. The USPS site says the mounts and bracket should be there today.

[pocket-rocket]

I wonder if the lope is causing the transmission to make some clatter giving false knock readings. I know my 92 with the 13* timing has a bit of lope to it and causes the 284 to make a slight bit of clatter that would make someone that didn't know what was going on turn their head. Maybe start the car, push the clutch in and see if the clatter goes away/knock count goes down. I bet if it is, the knock sensor is in too far since they are supposed to be torqued to a certain spec so they aren't sticking out too far (not sensitive enough) or too far in (too sensitive).

[jmgtp]

If anything I feel like it is running a bit rich, going by how it is running and how the exhaust smells.
The FPR is not adjustable. At idle it is running about 35psi. With the vacuum line pulled off the regulator it was around 44 psi if I remember correctly.

Thanks for the update on the mounts, they will surely help since the motor is bouncing around a lot without them.

I had the exhaust cams retarded on the dohc motor and I had heard due to that the o2 sensor could cool to the point of not reading right during prolonged idle. I ditched that setup and went back to 1 wire sensor for the LX9. But, the 1 wire sensor I have has no connector on since I clipped it off to use on the 3 wire. Of course now I can't find the connector that was formerly on the 3 wire sensor. I'm looking for a suitable weatherproof connector to use. I'll prolly pull one off the LX9 harness but o2 sensors seem to use some kind of special wire that I want to say isn't copper. Okay to use copper wire?

[pocket-rocket]

I've read that you almost can not solder to O2 sensor wire because of it being some weird type of wire. I'm not sure what it is, but luckily I haven't cut any of mine. I tend to stray away from harness cutting if I can. I don't see any reason why you couldn't use copper wire, but I've no idea behind the strange wire or what it is. If I had to think of a reason, I would almost say it's some type that resists heat better than copper.

[jmgtp]

I wonder if the lope is causing the transmission to make some clatter giving false knock readings. I know my 92 with the 13* timing has a bit of lope to it and causes the 284 to make a slight bit of clatter that would make someone that didn't know what was going on turn their head. Maybe start the car, push the clutch in and see if the clatter goes away/knock count goes down. I bet if it is, the knock sensor is in too far since they are supposed to be torqued to a certain spec so they aren't sticking out too far (not sensitive enough) or too far in (too sensitive).

The new trans clatters much less than old trans that was bolted up to the dohc, in fact I don't hear any clatter at all with this one, and the LX9 lope is mild compared to how the -13* cams LQ1 ran. Up until a few days ago I didn't have the axles in the trans, which meant no lube in it either, so when I started it I would always keep the clutch pedal in so as to not spin up the trans dry, and still had the knock issue. I torqued the KS to some spec I found via google for a 3100. I was too lazy to go locate my service manual upstairs at the time. Off the top of my head I don't remember what it was. I suppose I can remove the sensor, check that the threads in the block are clean and torque to the spec in the service manual to see if it makes a difference.

[CNCguy]

John,
If you have the O2 from the 3500 why not use it? At one point, someone had said they were too fast for the OBD-I but we have been using them them in our conversions without a problem. It's nice since the engines always come with two sensors. All O2 wiring from the plug to the ECM is just copper.

A = Ground
B = O2 Signal
C = Ground
D = 12 vt for heater

[jmgtp]

I may just use the LX9 o2 sensor then - the reason I didn't in the first place is because I too had read that my OBD I ecm wouldn't like it. It solves the problem of the connector.

Do I need to bother hooking up the heater portion of the sensor? I know I don't have to but is there a real benefit? On my 3 wire sensor I used the hot wire from the egr plug (I don't run an egr) to power a relay which provided power to the heater on the sensor. Now, I went back to the 1 wire setup since it is more simple and eliminates the extra relay. Since I'm not running an egr can I just use the egr hot wire to go right to the sensor? Does the o2 heater draw that much power that the relay would really be necessary?

[pocket-rocket]

If you still have the relay around, or have another around, better safe than sorry in my book. If you don't use the heater element, depending on where the sensor lies in your exhaust it might not get hot enough to read. O2 sensors have to be a certain temp or hotter to read correctly, hence the reason OEMs added heaters to them (especially down stream post cat O2s).

[jmgtp]

Well, I can counter argue that it is in the stock location which is right where the front and rear bank converge - it definetly gets plenty hot there and the OBD I setup doesn't use a heated sensor there.

What I meant to ask if a sensor is equipped with a heater, does the heater need to be function in order for the sensor give a proper signal? Could a sensor designed to use a heated element not acheive its operating temperature on exhaust heat alone even if it is extremely close to the exhaust ports, as it is?

I really don't want to bother hooking this relay back up! haha

[robertisaar]

you could disable the heater and have the O2 work correctly, it will just work as if it's a non-heated O2...

kinda self-defeating IMO.

[CNCguy]

The O2 heater doesn't require much power and the EGR feed should work fine. Once you start tuning, the time to go into closed loop can be reduced with the heated O2. To me, that's makes it worthwhile by itself.