Hi all. Longtime listener. First-Time caller.

My question to to community, in short: If you have a rock-solid 280z with the stock EFI, what happens if you "Tap Control Unit" while running? Anything? What if you knock it "a bit harder"?

I have a '78 280z that will very random-ish-ly very occasional-ish-ly simply turn off as though I had removed the key. After 10-20 mins of sitting, it fires right up and runs like a train...until it randomly turns off again, leaving me in some scary circumstances. Sometimes it fires up immediately like nothing happened. I have been trying to diagnose the issue for quite a while. After the last episode, I've decided she's never leaving the garage again until this is sorted.

While testing in the garage I randomly closed the car door and it immediately shut off. I thought it was a coincidence. Later, when I opened the door, it shut off...but then started immediately. I found a couple posts online of similar examples:

https://old.reddit.com/r/Datsun/comments/1f38p25/280z_shut_the_door_car_shuts_off_bad_ecu/

https://old.reddit.com/r/Datsun/comments/1iu2huj/280z_random_stalling_issue/

I have since checked every ground, pulled the drop resistors, cleaned/de-oxed all the connections and made sure they were tight. Just for good measure I swapped-in my spare ECU. I thought that had fixed it...I've been running it for an hour-per-lunch for the past 4 days, opening and slamming the door every 5mins, without issue. I was beginning to think I was good, until today. After about 20 mins of running I opened/closed the door and...stumble stumble stumble...running on 3 cyls...stumble. I gave the ECU a kick, and Poof. Back to normal. Kicked it again...sumble stumble...kick again...stumble...OFF. Won't start. Wait 10 mins. Starts immediately, runs like a train.

In the reference book "ELECTRONIC Fuel Injection - 280Z Electronic Fuel Injection Theory Troubleshooting" on page 39 it also has a small reference under "Engine Misfires - HC Reading Too High" it says "5. Tap Control Unit While Driving To See If This Aggravates Or Alleviates The Problem. If So, Try Another ECU."

So...have both my ECUs aged-out and need rebuilt? Or is "kicking the ECU" always going to result in some stumbling while 'something' re-sorts itself, even on a "known-good" ECU?

I found a thread of a similar-but-different issue where a fix was successful:
https://www.zcar.com/threads/efi-harness-continuity-voltage-problem.311536/page-2

And a quick-comment-on-the side by @zedhead that "In my case, I replaced the transistors and that fixed it. " (shot him a DM)

My follow-along questions would be about rebuilding and where to get parts...Trying to find a rebuilt unit seems like a non-starter. I spoke to David at MSA today about timelines, because "The rebuilder is locating a qualified Bosch-experienced technician"...for about 3 years now.

I'd be entirely happy to rebuild it myself and document the process...but I have no idea where to find components. Arrow and Mouser came up empty.

Someone mentioned I should ping @Captain Obvious, which is what brought me here.

Would love to hear everyone's thoughts.

Edited 20 hours ago20 hr by f1d094

I took a bunch of photos of the ECU for inspection and parts hunting

Edited 20 hours ago20 hr by f1d094
Oops. Forgot to convert to .png

PNG versions

These units are primarily “pth” assembly. They have aged with time IMHO have passed there designed life span. There isn’t as far as my knowledge a test bench that will run it through post diagnostics. My suggestion would be to search this forum and find someone that doesn’t have any use for the spares they have accumulated over the years. Mine are gone to other folks in the same predicament.

Most of the components in the ECU would be available generically. Caps, resistors, discrete transistors... No problem. The integrated circuits, however... Nada. You should be able to get pretty much anything but the ICs at Arrow and Mouser.

But before you condemn the ECU, I would suggest try to isolate the problem. Even if it's an electro-mechanical issue, it would be important to know if the problem is in the wiring or connector to the ECU, or inside the ECU itself.

So, the first thing I would try, would be to slowly and gently wiggle the wiring harness to the ECU and see what happens.

I'm just learning what's what. I have near-zero electrictronics background but I am handy with a soldering iron and recently bought an oscilloscope to learn on (Tektronix TDS 1012)

I can see that the WMIC labled parts are all metalized polymer capacitors and the resistors I recognize are pretty common...but the ICs? Are those the silver 14(?) pin guys with the heat sinks? And what the heck are those copper-coil towers with blue things next to the plug? I need a magnifying glass. I can't read 'em. Nothing looks like a discrete transistor that I recognize, but the round caps on the side say NEC...so I guess that would be those?

When I was running the car previously I've grabbed the harness connector and given it a good hearty wiggle with no change but today when I did the second "tap" it was on the plug and not the unit itself, and it went off like a switch. Mostly though, when I tap it with my foot it is on the lower metal bracket.

Tomorrow I will be sure to try and move/wiggle each bit of the harness from the connector up on into the dash and see if that makes any changes.

The large cans bolted to the outside of the chassis.... Labeled NEC D411A... Those are discrete transistors, but are a unique pair in the design. Those are the final output transistors that drive the injectors. There are only two of them, and they are known as "Darlington Transistors" because of the internal construction. The discrete transistors I was talking about are the much smaller black plastic bodied things with three legs.

I whipped this up labeling some of the components:

And you had asked about the copper tower and blue things near the connector... The copper towers are coils (aka inductors or chokes) and the blue Chicklet things are capacitors. Those two components in conjunction act as a filter on the incoming signals to the ECU in an attempt to eliminate high frequency noise. They originally designed the ECU with provisions to put those filters on many of the signals going to/from the main connector, but it appears they decided somewhere along the way that they are not all necessary. So they only populated a couple of them by the time they got to the end of the run. Most of them are just shorted across with a piece of wire.

Hope that helps.

16 hours ago, f1d094 said:

And a quick-comment-on-the side by @zedhead that "In my case, I replaced the transistors and that fixed it. " (shot him a DM)

Sometimes I take a best guess, no matter what the depth of my knowledge on a subject, and just try something. Then hope my brain will take the results and come up with something better. In retrospect, I doubt that the transistor(s) was/were the cause of the problem. I imagine that almost any open circuit or short inside the ECU could cause it to malfunction.

Your situation is unusual in that it seems so repeatable that it's almost controllable and is the same for two different ECU's. Maybe you're just losing a connection at one of the pins.

Probably one of the multi-vibrator functions! 🥴

3 hours ago, Zed Head said:

Sometimes I take a best guess, no matter what the depth of my knowledge on a subject, and just try something

Haha. Indeed. This is me, all day every day.

No matter which way this goes, I will be using at least one of my ECUs as a learning/traning tool to do a refurb from the ground-up and truly learn every last function, and to really play with tuning. This is why I already had: A spare ECU, a spare AFM, the FSM, the FI Bible, Bosch Fuel Injection & Engine Management, an Ignition 1-2-3 distributor, and an oscilloscope...as well as a host of articles and posts from around the Internet. But I had planned to do this on my own schedule. It seems my Z has other ideas...

In related news, last night I discovered that rebuilt ECUs are actually a dime a dozen; I just failed to Internet properly. I've fallen into some sort of pavlovian response mechanism where I assume everything is unobtainium and start scouring ebay and specialty z-parts shops. But if you look at AutoZone, O'Reilly, RockAuto, NapaAuto, etc. a rebuilt unit can be had for as little as $200, or $175 if you are willing to return a core. I've now re-added the standard parts-jobbers for future needs. It looks like they still have some bits and bobs. If it is useless I can return it no problem, so I figured: why not? When it gets here I'll do some compare-and-contrast.

More news at six. Film at eleven.

5 hours ago, Captain Obvious said:

The copper towers are coils (aka inductors or chokes) and the blue Chicklet things are capacitors.

Just a gut feeling, but based on the how-and-where-and-how-hard I tap on the ECU/connector, it seems like there might be a break in one of these coils...I think this only because they can rattle around a tiny bit in their respective tubes and would likely still "settle" with some sort of contact again because of the confines and 45+ years of work-hardening through vibration might do them in. I'll try and have a peek down the tubes with some sort of magification and see what I can.

Alternatively, based on your comment about the darlington transistors, is this any sort of failure mode that makes sense for them? Where rattling or knocking with make them fail temporarily?

The very temporary nature of the problem is what is most baffling to me.