Given that the overall troubleshooting approach has been a shotgun parts replacement,
your next step could be to just replace the engine, right?

http://forums.pelicanparts.com/support/smileys/wat3.gifhttp://forums.pelicanparts.com/support/smileys/wat6.gifhttp://forums.pelicanparts.com/suppo...leys/loki5.gif

Wow Loren - still sore ;)

I wish that was the case - it sounds differently, but pretty much runs the same...

-Wayne

Hmm, that's not really an accurate assessment. To date, we have replaced the following:

Cap, rotor, ignition wires, spark plugs, fuel filter, O2 sensor, coil, and we had the six injectors professionally cleaned and tested.

Other than the coil (which we swapped out with one to eliminate 100% of the potential for ignition issues), I would have put everything else on the "should be replaced" for a car that hasn't really been run in the past 12 years, and was stolen, and left out in the rain for a while.

Still open to suggestions.

Bottomline, at the end of the day, I have a feeling this will be traced back to an issue with the conversion setup - either through a faulty / weak wire harness or a fuel supply problem.

-Wayne

Wayne..can we go to the beginning-shortly-? If you don`t mind?
Did you build the 914/Carrera ..or somebody else put the engine in it? when you said it was always running badly since the beginning?
When it was stolen(i did read the story-a while ago-) was it already like this..all the system in the car?...because when you checked the camshaft timing it was not the same on both sides..why, you did not do the job ,right?
I still think it is something simple......
Ivan

Here's the (brief history) of this interesting car:

- Bought the car in 1995 after one week of living in SoCal (age 25). I didn't even have a place to live - I parked it in the hotel parking lot. I made sure that I bought a 2.0L, after all, I had read that the 2.0L was the one to get. I took it for a pre-purchase inspection at Vasek Polak, and asked them to make sure it was a 2.0L. The Internet was a bit new back then, and there was no Google images or much info on what a 2.0L looked like. Anyways, the car turned out to be a 1.8 - the previous owner didn't know, Vasek's tech's screwed up, etc. So, from that day on, I said I was going to convert it into a six.

- Bought another 914-6 conversion whose chassis was a basket case. It had a 2.7L engine in it. I put the two cars side-by-side and swapped all the bits. The car ran great! Until one day the fan belt fell off and cooked the motor. I had removed the shims from the inside of the pulley (I was 26, didn't know any better, and again, there were no books really on that except for the factory manuals, which are not clear and I didn't have). Anyway, this blown-up motor eventually became the "donor" motor for the Engine Rebuild book. That motor was eventually sold off in pieces, as I figured I would just go with a 3.2 for this car. I bought an engine in 2003 from Robert at Dutch Treat. It was in a wrecked 1984 Carrera, and I got to look at the engine, see it, and Robert ran compression and leakdown tests on it and it all checked out okay. This is the best way to buy an engine.

- So, the car sat for about six years. Then I had someone put it together for me, adding the flares, etc. The car came back about a year or so later (2008) and unfortunately it never quite ran right, and had a few other issues that I wasn't happy with. So, I parked it an concentrated on other things.

- Then, the car was stolen from where it was parked, around 2016. Bottomline, I tracked the guy down (he's in jail), and eventually got the car back (had to buy it at the insurance auction). We changed the fuel, the filter, and started the car, where it ran exactly as it did in 2008 (ran okay, but not great - mostly had an idle hunting problem). Then, the car sat for another 6-8 months as I tried to figure out who was going to paint it. Then we got it painted, and it came back in November of 2019. I just finished pretty much putting it back together, and I'm not tackling the engine issues that I knew would be an issue (for the past 12 years).

There you go - there are lots of missing details - like the 150 hours of me driving around LA looking for the guy who stole the car, confidential informants, working with the police, helicopter chases (yes, actually), etc. Some other day I'll write it all down or do a video... :)

-Wayne

Ok...this "Then I had someone put it together for me"..............this person would have known if the car was running ok or not...what kind of accident was it? The camshaft timing still is not ok by me ..as i do not know any mechanic who would leave it like this....Not dangerous, but not really right.....1,16 plus 1,37.............when i do it, i do 1,3 .....in CA the original specs were from factory on the lower scale..
Ivan

interesting, that makes me think how did i parked my 911 on the Santa Monica Beach (by the pier) parking lot for 14 years ..beats me now....with liability insurance only;-)

http://img.pccreation.net/photos/202005132336534953.JPG

The exhaust thing jarred my memory. Years ago I had a car come into the shop from the district service manager that had been to a few other dealerships. Complaint was, at times the engine had NO power. Lemon law, buy back, the whole thing. 70 friggen miles on it. One last chance. Me and another tech test drove it and sure enough, the problem happened on us and we both said, "plugged exhaust." Pulled the exhaust down and found what appeared to be a remnant of a stamping process during assembly that had somehow got left in the exhaust just before the catalytic converter. At times it would pivot 90 degrees and act as a throttle in the exhaust.

Since you've checked everything else, does the #6 exhaust pipe have a spot where the temperature changes drastically?

I can loan the below to you if you want. They came off a lower mile motor many years ago. None of this has been used run or used for years since it was removed, so YMMV.

http://forums.pelicanparts.com/uploa...1589407766.JPG

Was watching your cold start video when the car goes lean approximately 90s into the running. The WBO2 responds much faster than the NB stock 02 sensor as expected.

Do you have a scope trace showing the AFM output during startup and through this leanout condition? Is the AFM voltage also dropping? If the airflow is constant the AFM signal should be relatively constant.

The O2 sensor input signal to me is your #1 clue to this condition. It is the systematic change that makes the engine run smooth or poor. Because a change in ECU does not change running I don't think the PCB or circuit have issues.

Is the supply voltage to the O2 sensor constant? Does the 02 sensor signal voltage at the ECU match the voltage at the sensor? There was a question about wiring of the O2 sensor.

I would trace the O2 sensor wiring all the way through the harness and make sure wires are correct. Since this was a custom harness and likely these wires were extended for 914-6 installation this would be my next focus.

Today’s progress (NO CHANGE, ALTHOUGH SOMETHING INTERESTING – SEE LAST POINT):

- Another video of the cold start where it starts and runs semi-okay and then falls apart:

<iframe width="560" height="315" src="https://www.youtube.com/embed/JAyihnlTo2g" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>



- Checked the timing on cylinder number six. Looks steady, looks good.


- Checked the cylinders with the FLIR camera. Number six looks to be running at proper temp now, number three is a little light (it always has been). No change in the running condition (still missing)


- Checked the sound of the fuel pump today -> sounds perfectly normal:

<iframe width="560" height="315" src="https://www.youtube.com/embed/XqMHe2eKgCQ" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

- I also checked the sounds coming from the fuel rail. It was much louder on one side than on the other. I don’t know if this is normal or not, particularly in this custom configuration (914-6):

<iframe width="560" height="315" src="https://www.youtube.com/embed/tDR3NKppNNc" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>



- Checked the pulses on each injector. They seemed to look fairly even to me (some of pulses are upside down, as I might have swapped the leads on the plug when measuring it on the scope):

http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg

Then, I took some water bottles, removed the fuel injector rail from each side and placed each injector into each bottle. Then I jumped the fuel pump on the DME relay. On both sides of the car, there was no leakage from any of the injectors while fuel was running through the fuel rail. I then turned over the car with the starter (it didn’t start of course due to large vacuum leak caused by the injector holes in the manifold). It released small amounts of fuel when trying to start up, which looked even across all three injectors. I did this for injectors 4-5-6 and then repeated for the other side (had to cut down the bottles to make them fit under the injectors on that side. Photos:

http://forums.pelicanparts.com/uploa...1589425839.jpg
http://forums.pelicanparts.com/uploa...1589425305.jpg
http://forums.pelicanparts.com/uploa...1589425839.jpg
http://forums.pelicanparts.com/uploa...1589425839.jpg
http://forums.pelicanparts.com/uploa...1589425839.jpg
http://forums.pelicanparts.com/uploa...1589425839.jpg

<iframe width="560" height="315" src="https://www.youtube.com/embed/DIzj3N4cfxA" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Other side of the car:
http://forums.pelicanparts.com/uploa...1589426205.JPG
http://forums.pelicanparts.com/uploa...1589426205.jpg
http://forums.pelicanparts.com/uploa...1589426205.jpg
http://forums.pelicanparts.com/uploa...1589426205.jpg



Okay, so the injectors are not leaking when off, they seem to be flowing an equal amount from all of them, and they seem to be volumetrically consistent.

So, the next test I did was the “running on three cylinders test”, but instead of disconnecting the fuel injectors, I disconnected the spark plugs *and* the injectors. I had to ground the plugs in order to prevent the spark from jumping around the distributor (figured that out when it wouldn’t run). Here is the interesting part. The car runs the same when running on three cylinders with the injectors unplugged and the spark plug wires removed from the plugs and when it was running with *just* the injectors removed. However, when I go to plug the injectors back in (so that we have six injectors dispensing fuel, but no spark firing on half of them), the car starts to run rough and misses. This is a big clue, although I’m not sure what the conclusion is.

Allow me to ruminate. As we saw yesterday, the car seems to run okay on three cylinders on one side (unplugging three injectors on one side). Plugging in the other three (so all six are firing) makes the car start missing. Moving to the opposite side and unplugging the three injectors makes it run on three cylinders again, seemingly okay with no missing (on those three cylinders of course). However, I was looking to isolate the three injectors without having the cylinders fire, which we did today by unplugging the spark plugs. Plugging in all six injectors (even though only three are firing) caused the three that were firing to flub and misfire. This makes no sense. Dispensing fuel into the cylinders on the opposite side that have no spark should have minimal or no impact on the other side of the engine (unless I’m missing something obvious).

Yes, it’s not good to wash your cylinders with fuel, but for a short time for testing purposes, it should be okay.

Why would plugging in three more injectors on the other side of the non-functioning side of the motor affect the performance? I can think of only two things. Either the fuel flow is being affected by the addition of three more injectors (keep in mind, this is idle fuel flow, not a tremendous amount), or there is some type of electrical load or short circuit in the harness that is causing problems when the three (non-sparked) injectors are plugged in. I’m just thinking (especially after observing the injectors in the videos above), that the fuel flow of the injectors is just not enough to make any difference on the system. It’s a miniscule amount of fuel. The car revs as well – it just keeps missing faster when it’s revved (which is symptomatic of an ignition problem), but I haven’t been able to find any trace of any ignition issues and we replaced the cap, rotor, wires, plugs, and coil already (as well as an ECU swap). I guess it could be a distributor problem, but I checked the timing yesterday and today on cylinders, 1,2 and six for about 3 minutes at a time, and they were absolutely rock solid. Usually, you can spot ignition problems very easily with the timing light. Nothing showing up here. And even as the engine stumbles, the timing never stumbles – it’s always at the same spot, rock solid - which would indicate the TDC and speed sensors are good (which we already tested).

So what would cause the engine to start to stumble when the non-spark driven injectors are plugged in? The only thing I can think of would be an electrical problem either with the injectors (not likely since I have swapped them around a bit), and/or a wiring harness issue. A few days ago, I found that the resistance in the injector harnesses were higher on #5 and #6. After cleaning some contacts and remeasuring, they appear to be back to normal. Someone here mentioned that they saw one of their harnesses had worn away and had started to ground out. I haven’t seen that show up on the electrical tests yet, but then again, I haven’t necessarily been checking the resistance to ground per se.

I think tomorrow I need to clean up, put all the tools away, button stuff back up that I’ve taken apart, and recheck the harness. Seeing how this is a conversion car with a “custom” harness, this would seem to be the most likely culprit (even though the most obvious tests haven’t revealed that). Rob at DC Auto is sending me a pressure regulator and a fuel pressure damper to try out – they should be here next week. That will be another thing to eliminate – I’m not sure of any other way to actually test these units other than applying vacuum to them.

-Wayne




Wayne R. Dempsey, Manager, RDW Development
wayne@rdwdevelopment.com - 310-261-3604

My tests were only approximate too, not super precise due to the engine being in the 914 backwards and nearly impossible to get visibility on. A tiny turn of the engine here and there can affect the readings. I was looking to see if they were seriously off, which they were not. I spent about four hours or so checking this, I wasn't going to spend another four just to see if it was "spot on".

So, I think it's okay...

-Wayne

It's slightly off now on #3. That cylinder has always been running a bit cooler from day one, although it's hotter / normal at the top near the cylinder. I'm not sure if I'm nitpicking it, but I haven't been able to find anything else wrong with #3.

As for the "banana in the exhaust" theory, that's a good one. But, the car will still rev okay, so I think that eliminates that particular issue for now. You could be on to something, but I don't have a good clue to point to that one. Number six is now running at good temps but the engine still has a problem.

-Wayne

Wow, yes, I will take you up on that offer! There's a lot I can do (swap the regulator and/or the damper), and unplug my injectors and plug those in to place an electrical load on the harness to see if it's an electrical problem.

I'll text you!

thx,

Wayne

Ok Wayne - I think I have one idea that would explain the issue and your observations:

The injectors are supplied with +12V from what is typically referred to at Terminal 15. Depending on how the conversion is done this either comes straight out of the ignition switch or from the first stage of the DME relay.

If that supply line has too much internal resistance (read skimpy wiring) your current through the injectors is limited by that supply. This would explain why hooking up the three injectors who’s barrels don’t get spark brings down the other barrels when running on 3 cylinders.

To test figure out what side of the injector harness is +12V and run a beefy 10 gauge wire from the battery positive to that side. I bet this will make a difference.

To measure hook your o-scope between the +12V side of an injector and the positive terminal of the battery. If you see a signal that’s the non-wanted voltage drop

My money is on this guy! I think he's on to something. You might want to take some contact cleaner to the injector harness connector pins or rough them up with some scotch brite.
I'm about to install a 3.2 in my '69 Coupe and this thread gives me the shakes...

Bill

Okay, cool deal. I'm with you on this (although I had to read your post four times before it got through to the grey matter).

Okay, the injectors are pins 14/15 (the same, as you mentioned they are joined inside the DME), and the other side is pin 87 of the DME relay. I will completely agree with you, as the DME relay harness is custom-made and disappears into the abyss of the rear trunk, surrounded by electrical tape. I have indeed already looked at the DME relay and wondered out loud "where does the current come from for this".

Okay, the Bentley manual (which is not always correct as everyone here knows), says that the power for the injectors is provided through pin 87 of the DME relay and pins 14/15, which is shown below:

http://forums.pelicanparts.com/uploa...1589431979.jpg

The power is provided from the battery via a wire at the top of the diagram (4.0 RD) and applied to pin 30 of the relay. So, to summarize what you've said, you would suggest running a super-beefy 10-gauge wire to terminal 30 of the DME relay, just in case there isn't enough current flowing there from the battery. I can also check the resistance in this wire (from pin 30 to the battery), although that wouldn't necessarily reveal a current / capacity problem. I think I can fairly easily tap in a direct wire from the battery positive terminal to pin 30 of the DME relay.

But it's a little late to run the car in the neighborhood, so I will have to wait until tomorrow.

Also present on this car is the typical 911 "exciter" alternator issue (where you have to rev the engine to excite the alternator and get it started charging). My 1972 911 has the same exact issue. That issue, I believe, is related to grounds through the car and/or the contacts related to the bulbs in the gauge (perhaps the wrong bulb is installed in the gauge - haven't checked that out yet). Just to eliminate any issue with the battery or voltage drops, I've been running the car with a low amp battery charger hooked up, just in case (didn't make any difference).

Steve W. is going to lend me those injectors / rail tomorrow. I will then be able to unplug the injectors from one side, and plug in the spare injectors to see if applying an electrical load (but not fuel load) will reveal any more information.

My gut from day one seemed to hint at the fact that this issue may be "conversion related". Not really based upon anything in particular, but mostly based upon the fact that I have a *lot* of one-off type custom cars, and they all have little weird things like this that pop up from time-to-time.

-Wayne

Wayne,

The voltage drop can happen a anywhere from B+ (battery terminal) to the powered side of the injectors. In a stock setup that could be the DME relay socket, a spliced wire to the DME relay, a dirty fuse holder and the list goes on.

Fact is if you have 6 low-impedance injectors with a DC resistance of 0.5 Ohm it puts serious demands on stable current supply.

And don’t assume whoever did the conversion followed the stock wiring diagram. Maybe they ran power to the coil and/or injectors directly through the ignition switch. You never know.

Wait a sec. So, that's only one part of the equation. So, you're saying that the current feeding the injectors might not be beefy enough, and that comes from pin 30 (connecting to the battery), that might indeed be the case, but then we'd see a voltage drop (maybe) at the breakout box for the DME, and perhaps voltage drop problems with the ignition system (which we may be seeing). Indeed, I will beef up that wire, and run a separate one to the battery.

But, the injectors are driven off of pin 87 (lines / junction A & B in the diagram above). So, I think you're actually referring to the two wires that say 2.5 RD and head to points A/B. If these are not beefy enough, then this will show up as a problem. This would be the wires from 87 to the six-prong connector (which only has four prongs populated) on the injector harness. I understand now.

So, you want to make sure that the pin in the four pronged connector (which has 12V on it from pin 87 of the DME relay) has enough current to drive the injectors. So, I can run a new wire to pin 30 for the DME relay, and that would assure good current to the ECU, etc. but I should also run a wire from the battery to the two terminals on the four pronged connector (two of which will have 12V when the DME relay is energized.

I get it now.

http://forums.pelicanparts.com/uploa...1589433436.jpg
http://forums.pelicanparts.com/uploa...1589433436.jpg

This connector:

http://forums.pelicanparts.com/uploa...1589433386.jpg

Thx,

Wayne

Yup, that's a good idea, and I already did that a few days ago. It improved the resistance of the harness...

-Wayne

Indeed, there are these mystery fuses located here that I have not messed with yet. Now is probably the time!

http://forums.pelicanparts.com/uploa...1589434230.jpg



I've never quite liked this particular connection either:

http://forums.pelicanparts.com/uploa...1589434173.jpg


I found this info on the web too that explains it a bit better:

http://performancefuelsystems.com/InjectorCompattibilitywithECUs-TechCorner.htm

This electrical issue would also possibly explain why the missing part doesn't seem to be concentrated on any one cylinder in particular, and also why the problem doesn't appear to change or be affected when the rpm increases. If it were a fuel pressure problem, I would have thought that there would be an issue as the rpm increases...

Hmm...

-Wayne




-Wayne

You’re getting the concept: think of all six injectors being a 0.5 Ohm load. Even though they are fed through separate harness branches the DME acts like a giant switch to GND via the two output pins. Internally there is one channel that turns on to thr negative pole of the battery. The other side of that load is supplied by T15 (1st stage of the DME relay in a stock setup) going back to the battery positive.

At 12V and 0.5 Ohm you’re running ~24 Ampere through this load. That’s what the wiring needs to support. Any poor connection, e.g. poor GND to the DME or poor +12V to the other side of the injectors will cause issues.

Assuming here that this bank of injectors is NOT monitored in any way by the O2 sensor? Otherwise these fuel vapours would certainly indicate 'full rich' and the ECU will intervene

For the purpose of our testing here, we have the Oxygen sensor unplugged, which puts the Motronic system into "open loop" mode.

Also, specifically to answer your question, the injectors were plugged in on 4-5-6, without spark, and the O2 sensor is on 1-2-3. But - I believe that the O2 sensor will not "sense" fuel vapors - only ignition / burned exhaust. The CO / HC gas analyzer would detect that (but not the O2 sensor). That is at least my understanding of it.

-Wayne

Yes - understand completely. My initial hunch was something in this one-off wire harness was not completely correct, but I was focused on internal resistance and work hardening and I didn't consider the current requirements of the low impedance injectors! But the test today where I electrically "loaded" the six injectors on the harness seemed to point to this theory.

I also ordered more connectors - I can wire each 12V line to each injector directly to the battery if need be to perform the "ultimate" test of current flow.

-Wayne

hmm ..(maybe i overlooked it )..did not know it is home made el.harness.looks like you are in good hands with isSmithz...

Ivan

the forum needs a "like" button.
what a knowledge base!

The thread is so long I can't find if you might have checked the grounds but that's a very good point here as well. All the 3.2 harness grounds are in the engine compartment bolted to the left side of the intake manifold on a 911, but you've got a custom setup there so might want to see where all those ground connections wind up and clean the contacts there as well. The DME is bolted to the chassis but I believe that ground is just for the outer case of the DME unit to shield it from EMI and prevent it from injecting noise into your radio.

I always though it was rather poor design to have such a high current ground going all the way back to the engine compartment, would have put it right next to the DME on a very short wire. The sharp edge on the signal will couple into everything else as they run along the inside of the harness as well.

Bill

24 amps thru the transistor switching to negative :eek:
Big time epiphany

Suggest a nearby mounted terminal board with big ass wire feeding it for all the auxiliary power sources so as to have the same mechanical connection integrity and resistance

It’s not that simple: peak current only happens for less that a millisecond during the opening pulse and 24A is a purely theoretical value assuming zero loss in wiring, the transistor and all connections. It’s to illustrate that even small imperfections will have a big impact on the ability to properly control the injectors.

Transconductance of older BJT Darlington transistors are pretty poor by today’s standards and that’s why Bosch had to employ the P&H topology.

Remember, the inductive load of the connected injectors affects the injector pulse width, given the design of the
911 3.2 fuel injection system versus the later 964/993 design. As a result, the effective fuel mixture gets changed.

You're joking, right? Where did that get pulled from?

Well, I ran the 10 gauge wire last night and just started the car up. No change. BUT - the 12V lead is indeed only one side of the equation. I did check all of my grounds recently for resistance, but not necessarily thinking about current draw. In order to have the large current flow you need to have a supply *and* a return. So, the supply of the 12V to the injectors is only one element - the ground connections that are switched by the DME also need to be well grounded. I can take a few leads from my breakout box and wire them directly to the negative terminal of the battery to ensure that the unit is well grounded. If that doesn't work, then it's on to the next thought...

-Wayne

While you're at it also run a heavy gauge wire to the coil's +12V side. And good point on the GND. That is shared between the 6 injectors and the coil as the biggest complex loads. GND comes to the DME via the DME harness on multiple wires. There is lots of internal splicing going on and one big bundle ends up in the engine compartment on a stock 3.2.

Internally all GND pins of the DME are tied together and they are kept seperate from the the DME housing (shield).

Ingo

Most who have worked on engines that use an AFM as an input for load, understand that it's likely that the AFM has been re-adjusted
over time from the factory setting, e.g. for more power. This usually results in an over-rich running condition when warm. Knowledgeable techs
when having rough or rich running conditions, typically check the spring tension of the wiper by moving it at the problematic RPM,
and noticing the effect. The typical AFM wiper spring tension setting is set at 2K RPM for a 1.0 - 1.5 CO without the O2 sensor,
with a warm engine.

Many who have reverse engineered both the Porsche & BMW Motronic DMEs of the '80s, have learned that Bosch designed Motronic DMEs
for both low & high impedance injectors. The standard BMWs, e.g. 535s, used high impedance injectors. The M3/M5/M6 engines used low
impedance injectors like the 911 3.2, and as a result had basically the same DME ECU design but a different EPROM.

what's your point Loren?

Why do YOU think Bosch employed a complex P&H driver design based on a custom IC and many discrete components instead of just picking a modern IGBT, MOSFet or smart lowside switch :rolleyes:

And if you'd done your homework you'd know there are simpler Bosch designs that only work for high-impedance injectors from the same time period. I wonder why....