I'm pretty sure it's only when actually pulling load, it seems to be just fine sitting on the ground, which is going to make it fun to diagnose since it's raining for the next while here.

Since it only happens under load you could be fuel starved or ignition system is weak. Have you looked at the engine bay at night? You may see ignition arcing around the wires at night.

If he wanted to do that fuel delivery test found in the manual where would he source that fitting? Is it the same as the pressure test port?

It wants you to remove the fuel return hose back by the speed and ref sensor plugs (not forgetting to counterhold;)), hook a hose up (fitting required) and jumper the fuel pump. You should see .85 liters in 30 seconds.

Is that point after the regulator and this is the volume you can supply at 2.5 bar?

You are correct, you want to disconnect the return line at the FP Regulator front of the LHS fuel rail, you simply want all return fuel going into a catch can instead of the fuel tank. This after the regulator, so it's going through the regulator. And you simply start the pump up for 30 seconds and you need to have no less than .85 liters of fuel in the catch can.

You don't need the exact matching fitting you just need a hose that fits very snug over the fitting and won't leak. You can even clamp the hose to the fitting if need be. You won't have any significant fuel pressure at that return port or in the hose during this test.

The fuel pressure test is done at LHS rear of the fuel rail near the Distributor while the delivery rate is done at the return hose at the FP Regulator, they are not the same port.

finally had time to do somethings last night, I don't have a motronic breakout box so I was only able to test some of the 30 pin connector electrically, I did that, things seemed good. We checked the coil 0.7 ohm and 5200 ohms I think that's good, but found in my manual it says the values are in "Table 2" but table 2 is no where to be found.

Also popped in a different computer loaned to me by the fabulous folks at Chris's German and started the car. Then we took it out around the neighborhood and it seems just fine. well damn it! we didn't change anything apart from re-securing every connector and those tests. I'm going to have to wait until the weekend when I'll have time to drive it and try the original computer as well.

Your DME likely has solder joint issues. I see a lot of DMEs with intermittent issues these days, I've been fixing a lot of bad joints lately as well as Land lines that are worn through. PM me if you need help with the DME.

I think you just found the issue, it's the DME.

The coil ohms spec is:
Primary winding - 0.4 - 0.6 ohms
Secondary - 5000 - 7200 ohms

That's right out of the DME Master Test Plan

Your coil seems fine, you are in the general area on the resistance and some meters are not that accurate at the very low ohms reading below 1 ohm. So if the meter is not a fluke or hi-end meter I don't trust the readings below 1 ohm.

Not necessarily:

1. The original fuel injection computer needs to be put back in and re-tested.

2. If the computer is bad, it may not be an intermittent connection problem but
a faulty component in the computer.

"Land lines" - You maybe mean circuit traces.

If it's a bad computer, Pelican Parts provides a rebuilding service for that:

http://www.pelicanparts.com/cgi-bin/ksearch/pel_search_2014.cgi?SUPERCAT_FLAG=Y&make=POR&Conte xt_make=&please_wait=N&LastVisited_input=911M&Prev ious_Section=&forumid=8&threadid=823096&command=DW search&description=DME&I1.x=0&I1.y=0

Dave,

Sure it could be a bad dirty contact or a bad component. But I see a lot of DMEs that are failed at the ignition circuit. It may be worth while to have the DME checked out.

I also am starting to see some DMEs with 'circuit traces' literally rubbed off the board from vibration. This I had not seen in prior years but now I'm starting to see this type of failure as well.

I agree with you that he should put the DME back in for another test but having that DME inspected would be prudent at this time, given that a donor DME worked just fine.

Meter is a Fluke so it's likely quite accurate. I'm going to go for a test drive with the new computer, and then swap the old one in and drive more. I've got it out in front of the seat so I Can change it in seconds.

Last night I took the car out for a slightly longer driver, only 3 miles or so but through the subdivision and out on the main road up to 60 and then back in, and it ran very strongly. Then I my computer back in it and took it out again, and it worked fine again! Last time I didn’t get even as far out of my house as I did both times last night before it was clearly broken. I even thumped on the box (it’s sitting in front of the seat at the moment) and nothing changed. So I’ve got three theories:

1. There was a loose connector somewhere and when I went over the engine compartment and checked everything I fixed that
2. Fuel, either the fuel pump is starting to fail and is now working okay again but will shortly fail again, or when we cracked the fitting at the end of the fuel rail we got an air bubble in that took a few runnings to clean up. My buddy Nate who spent time between high school and college as a Ford Mechanic said he had a Toyota that failed like this once.
3. When I originally forgot to put the CHT Sensor on (it was disconnected when I first started the car last Saturday) the DME learned something wrong, and by disconnecting the power (i.e. pulling the DME out) this got reset. Does this happen?

3) is impossible. The 3.2 DME doesn't learn anything. Every time you shut the car off it's a full reset. It also doesn't adapt during driving other than the O2 sensor driven mixture adjustment.

Ingo

I've fixed a few DMEs behaving exactly like you describe. Working great most of the time but then with no apparent reason they act up. These DMEs had bad solder joints. What happens is that the joints are cracked and sometimes they make contact but then as temps change inside the DME they erratically act up. It's prudent at this point to get the DME inspected, when I find a bad joint or 2 I can even photograph them and show the customer. You decide but I highly recommend getting that DME checked out.

I just fixed one last week with exactly same issue as yours. Customer complained to the shop that car runs like crap from time to time, it runs so bad they need to pull over but every time it goes to the shop they can't reproduce the issue. Shop owner finally sent the DME to me and sure enough it had 3 cracked joints. It's back in car no complaints so far.

Since now it's working fine I say just drive it but if the problem comes up again you need to get the DME inspected.

You generally can't visually find problematic solder joints which then requires that
the complete circuit board be re-soldered. Some joints may look good but be bad and
vice versa. Without fully verifying that the computer is bad, i.e. using another computer,
rebuilding that computer is a waste of money.

Most likely the problem/problems:

Dave,

I respectfully disagree, I locate bad joints all the time via my inspection process, I've fixed many 911 and 944 DMEs by inspection alone.

But if it's running now without issue then just leave as id. However, if the problem just pops up again then the DME needs to be look at by someone who has experience in fixing these.

Obviously this method can't find all the bad solder joints and is a very poor way to
properly rebuild, i.e. if that's what being claimed or advertised, any electronic device that may be intermittent.

Dave,

The boards have common places where cracks develop and if you repair DMEs often you know the locations. The reason the 911 DME is so prone to the problem is simply because it's mounted parallel to the ground plain meaning it's horizontal to the horizon. Very bad mounting, in aviation industry this is a huge NO NO practice. Then the DME is also directly bolted to the floor board and what happens is that with years of rough service the heavy components literally drive them self through the board, the pins keep pushing down into the board till they crack. Knowing this is the issue you can easily inspect the appropriate component areas and find the cracks. Here's a perfect example of a DME I fixed and you can clearly see the cracks in the hi-res picture. This DME was DOA and after just fixing those 3 solder joints the DME worked. My point is you can easily find the problems using good inspection process.
http://forums.pelicanparts.com/porsche-911-technical-forum/652375-dme-bad-solder-joint-repair.html
I see DMEs come in like this often, it's very obvious what's wrong, why would you disagree with visual inspection as a decent way to find issues? I'm not saying it's the only way but to say it does not work seems wrong.

And by the way I never claim I rebuild, I repair DMEs. To claim 'rebuild' implies refreshing components to prevent failure and I'm not sure anyone does this? I simply try to make broken DMEs work again and questionable DMEs are tested and visual inspected.

Good DIY (Do It Yourself) info which should save those with a 911 3.2 computer repair
charges. Based on the image of the circuit board, anyone with a screwdriver and
a soldering iron should be able to fix their computer in about 10-15 minutes if they
have a no-spark/intermittent spark condition.

Maybe not but if they have a soldering iron they have some basic interest in electronics. If you saw an obvious fault you would be pretty dumb to ignore it.

How does the flat rate rebuilt board differ from the board that is just troubleshot and repaired. Do you replace all the electrolytic capacitors? Replace all the low mean time between failure parts?

Rick,

I'm done with Dave, the guy loves to argue.

I repair plenty of these DMEs, most repair shops are robbing folks blind! These boards are failing mostly for cracked joints like the one I showed. Burned circuit lines, also very easy to spot, from shorted components. And once in a while I find a bad transistor. This is not guess work, it's what I see from the 50 or more DMEs I've fixed or converted over the years. I may not have a full blown repair shop but I know what the most common failures are and 50% of the repairs are simple easy to see cracked joints. Almost all issues are on the analog board.

If you feel better about the repair because it costs more so be it.

I'm sure that most shops just fix what's broken, no preventive parts replacing is taking place. It's also silly to re-solder an entire board, you need to simply understand the reason the joints fail and you quickly realize what needs attention. I have yet to find a failed solder joint on the digital board of any repair I've done. Why is that? I already explained this, but guys like Dave don't care to listen. The failures are from stress fatigue in the very heavy components.

I have years of experience in Plant Floor Control system that see very harsh environments, we saw failures like this in months. This problem in our DMEs is simply from 30 years of minor vibrations compared to some industrial equipment.

Also don't trivialize how hard it is to spot a failed solder joint visually, it's not that easy unless you know what you are looking for. Also most folks don't have the proper skills or tools to fix the joints correctly. I've seen real hack jobs where folks tried to solder boards, you really should let a pro fix these, I don't think it's a DIY project for most folks.

I am reading these threads because I am trouble shooting a DME problem and thought I would add my 2 cents worth to this thread...



I appreciate the input that people like scarceller bring to this forum and this thread has been helpful and informative for me. Anyone that is servicing many examples of these (be it the DME or any other part of our cars) can see trends and understand what age & time can do to these parts.

What he says about the durability and failure mode of the DME make sense and certainly the approach of fixing what's broke and not fixing what ain't broke makes sense to me and many other DIYers.

I don't understand the arguments sometime brought into such discourse like that in this thread. The approach that scarcellor uses may not be what everyone in this forum thinks best but if it isn't your cup of tea, it doesn't mean it isn't what others would choose.

So +1 scarceller -1 Dave