93 5-speed manual barely running

[mskrotzki]

My car, '93 w/ 205,000 miles barely runs. TCS light is on, but no codes retrievable. Previously got 1 - 3 (or 3-1, don't remember) but now I cannot retrieve any code. CEL flashes intermittently, rarely stays on. Engine pulls and coughs, catches again, coughs, etc, allowing movement -- a miserable 9 mile drive home. Tach drops perceptively at cough, then engine re-catches. Idles at 1400, slowly drops, then perceptively drops to ~100 a few times, finally stalling Fuel pump relay clicks every time the engine starts to die (and tach drops at the same time.) Main relay is also clicking -- sometimes. It is a brand new relay, just replaced today.


It behaves as though ignition is being cut, since it is instantaneous. In the engine compartment I can hear fuel pump running and fuel coursing through the fuel rail when ignition is on. The flow is directly connected to the fuel pump relay clicking (obviously.)


To get home I disconnected top TCS plug, to no avail. I replaced it upon return with no difference.


Without codes I know this is difficult. I have obtained codes numerous times so know the paper clip MUST be firmly inserted against the metal contacts.


I suspect it is something in the control modules affecting the fuel pump/igniter, but any help or thoughts will be appreciated.

 

[solidol]

Start small, inspect ignition and fuel delivery systems. Do compression test and go from there. These cars are so simple.

 

[Old Guy]

So many possible causes.

Did this problem emerge rapidly or has it been a gradual deterioration in performance? If it showed up quickly, I would be inclined to look at electrical problems. If you are suffering from momentary loss of power to the ECU, the ECU will cut power to the fuel pump portion of the main relay causing the fuel pump to stop. However, if the engine dies from other causes, as the RPM drops below the critical threshold that is defined as 'the engine is running' the ECU will stop the pump from running even though the ECU is still powered up. That is a safety feature in the event of a vehicle crash. So, the pump cutting in and out is not a guarantee that the ECU is losing power. However, if the CEL is coming on intermittently, that could be a sign that the ECU is 'restarting' following a loss of power. It is normal for the CEL to come on when you first turn the ignition key to run. If the ECU loses power it will cut the fuel pump. As you coast down in speed, if the engine is still turning at an RPM that is above the 'engine is running' threshold and power is restored to the ECU, the engine will start back up resulting in the cough, catch, cough behaviour.

Electrical supply problems will generally not trigger an error code in the ECU. Look for other evidence such as low voltmeter reading and the dash lights flashing. Common sources of electrical problems on the NSX are the battery post clamps which after a while can become stretched through over tightening and are no longer tight on the battery post and a worn ignition switch.

 

[Larry Bastanza]

Agreed 100% about the battery post clamps, make sure you cannot twist them by hand. Also I would get a fuel pressure gauge on it. That will tell alot. Also check all ground connections especially on the engine block. I expect since it is a 1993 the ground lead is located on the front valve cover close to the coolant tank. Make sure it is there and tight.

HTH,
LarryB

 

[mskrotzki]

Thanks all! Tomorrow I'll check battery clamps, check ground connections, ignition switch. If nothing there, then a fuel pressure reading.

An additional point: With the engine stopped and the ignition on, I can hear the fuel running along the runner. The relay click and the flow stops. After a few seconds when pressure is reached, the fuel flow stops. I've never heard this before in the 12 years I've owned this car. Any significance?

The problem has been intermittent, happened perhaps half-dozen times previously but "cleared itself" after perhaps 15 -60 seconds. It is almost certainly electrical, since on my trip out this afternoon it ran strong, and on the way back it reverted to a previous, rare behavior.

 

[Old Guy]

The 'normal' start up is if you turn the key to on (not start), the fuel pump will run for 2 -3 seconds to prime the fuel system. Its purely time based (not pressure based) and is controlled by the ECU. If your ears are good or your main FI relay is noisy you might hear the relay click. I can hear my pump whirr; but, I can't say that I have ever noticed the relay click. When you turn the key to start the ECU will detect the engine turning over and the pump will again start running and should then run continuously if the engine starts successfully.

A lot of people do not do the turn key to on, wait and then turn to start. They just put the key in and turn to start and the engine usually fires up fine. This works because there are check valves in the fuel system that are supposed to retain fuel pressure in the fuel rail on shut-down. As a result, they have never heard that prime pulse that the ECU gives the pump.

The only time you should hear the pump running if the engine is not running is in that short period after you first turn the key on; but, not to start. If you are hearing the pump getting a prime pulse just after the engine quits on you and you have not power cycled the ECU (done the key off and back to on thing) that would seem to be a sign that the ECU momentarily lost power.

 

[mskrotzki]

BINGO! The negative battery teminial twisted easily by hand. Re-secured it, checked the positive terminal -- just fine. Turned on the key, no strange noises, started it and the TCS light went out, took it out for a drive and all A-OK.


I've always said I can fix anything if I have the right tools. I think I'll modify that to "I can fix anything if I have the right tools AND the right advice."


Thanks to all the respondents, especailly Larry who nailed it. Now I am ready for the next 205,000 miles! This one was so easy I didn't even change clothes -- talk about luck.


I think I'll put the original main relay back in. Anybody need a new main relay?


Thanks again.


Mark

 

[Honcho]

Keep the relay as a spare, Mark. Also, you might want to invest in Ben's negative battery cable. The factory negative cable clamp deforms if tightened too much and will be difficult to get snug afterward. Ben's cable solves that issue. It's on my shopping list for my NSX-R project.

 

[mskrotzki]

Unfortunately I have to revisit this thread. While it ran fine for several hundred miles it started all over with the same symptoms. I previously replaced the sketchy negative terminal with a new, formidable unit from auto zone and believe me, it is on tight. The positive is also AOK.


As suggested by Larry, I next checked the engine ground and found someone had previously stripped the threads in the valve cover! How has it run for the 85,000 miles I have had it? I gather from reading several articles that you don't just re-tap threads in magnesium (or other low ductility materials.) It also seems you can get into trouble with electric couples between differing materials. With used valve covers running $500.00, I don't want to screw this up!


I'm wondering if I can get some aftermarket ground cable and make another ground from the body to the engine block. There appear to be no shortage of potential locations to do this. I would appreciate some thoughts.


Thanks in advance.


Mark

P.S. Good advice, Paul, I will follow it and keep the spare main relay.

 

[mskrotzki]

This is what I have done. Beautiful? NO Does it work? At least for the time being. An AutoZone 19"Starter Cable is doing the trick for now. I am taking the car up into the mountains -- over the Continental Divide -- on Saturday. That should be a pretty good proof of concept.

I always feel good when I can identify a problem, fix the problem and then the item works. Did this the first time, but it turned out it was the same problem, but at a different location. I'll keep everyone posted for future inquiries.

 

[Old Guy]

The NSX (at least my 2000) has two engine grounds. The ground cable you replaced is called ground cable B. The other is ground cable A (imagine that). If you were having operational problems caused by engine grounding, chances are that both A and B were compromised. Replacing B may have resolved the problem; but, only temporarily. It would be best to check the condition of cable A.


Cable A is connected to the transmission housing. From the service manual, I know roughly where it is supposed to be; but, I have never been able to spot it with all the other stuff that is down there. Anybody who has pulled an engine will have had to remove the ground so they may be able to give you a better description of the location and advise as to whether you have to remove stuff to gain access. Perhaps access is from the bottom of the car?

I don't know what you grounded the B cable to on the engine; but, it probably can't be worse than the valve cover if it is a reasonable sized bolt.

 

[asylum]

The other ground cable is connected at the top of the gearbox, under the air filter box.

 

[mskrotzki]

Thanks asylum. I'll check it out.

The situation is getting really weird. I just returned from taking it out onto I-70 and running it at ~85 for ~15 miles. All was AOK for the first 14.5, when the TCS light came on and it started hesitating, with the tack needle going crazy. The speedo needle is fine. I pulled of the interstate and got onto a parallel road where, while not running correctly, was definitely driveable. Later I went through some small town streets and it degraded, then recovered slightly. When it got really bad, I looked for a tree to park under and take a look. (Finding a tree in eastern Colorado is no mean feat.) I opened the engine lid and took a 1/2 hour walk to let it cool off to see if I could find any temperature correspondence. Upon return I turned on the ignition and the relays were clicking and fuel was coursing through the fuel rail for about a minute before it stopped. Starting it, it still ran crazily. I decided to nurse it home (~25 miles) and go for it on the service road. FWIW, the A/C was not working during this time.

About 15 minutes later it started running fine, the A/C came back on and all was back to normal -- for about another 15 minutes. By the time I got home it was as bad as I have ever seen it.

I'm thinking fuel pump, fuel filter, TCS computer. I replaced the filter less than 3 years ago and less than 30,000 miles ago. The fuel pump was a prime candidate when it started running bad at a steady 3750 rpm, but then it started going bad while driving through neighborhoods, so I am not sure. At 205,000 miles I wouldn't be surprised. I'll pull the computer and examine for swollen capacitors later.

Any thoughts?

Mark

 

[Old Guy]

. Upon return I turned on the ignition and the relays were clicking and fuel was coursing through the fuel rail for about a minute before it stopped.

Mark

Time the prime pulse duration. If the pump really is priming for 60 seconds before start-up, it might be time to start shopping for an new ECU. You should not have a bunch of relay clicking. The fuel pump control portion of the main relay should click once when the pump starts to prime and click off when the prime is finished. It should not click again until you turn the key to the start position and the engine starts to crank. Other than that, no clicking.

 

[mskrotzki]

I have a lunch meeting tomorrow morning in the mountains. I am going to pull out both the ECU and the TCS computer in the afternoon and look for the dreaded failing capacitors. The grounding issue, while it was sketchy, now seems to be covered. This is going to be interesting. And perhaps expensive.

 

[mskrotzki]

Upon a closer look at the bottom of the ECU, I found this:

​

I have the RSX injector set installed, and this is the socket where the original factory chip was installed. To me it looks awful, and I'm hoping this is the source of my problem. However, I'm not good with electronics, so I really don't know. On top of that, it has be running fine for at least 7 years, if not longer, so why now?

Looking forward to your suggestions.

Thanks again,

Mark

 

[crxguy52]

I've been talking to Mark, and to me it sounds like cracking solder joints in the ECU. It's the common denominator when you consider the AC not working, tach (and not speedo) dropping, lack of codes, and odd relay behavior. The car I bought had the ECU replaced by the second owner, citing "degraded performance", and it resolved the issues they were having. It would also not surprise me, given that other control modules (main relay, fan control) mounted in the same location suffer from similar issues. The lack of solder in some of those joints probably isn't helping, either. I'd be interested to hear what others thought.

 

[solidol]

^ given observed behavior and newly emerged evidence I'll second this.

 

[mskrotzki]

I am declaring this problem solved.


Unfortunately it is such an odd ball the solution may not be useful to future NSX owners.


To review: Car become undrivable. Engine would sputter, backfire, have 0 power, stall. The relays on the firewall were clicking like a 1950s telephone exchange. Miserable situation.


First thought and action, reflow the solder joints on the main relay. “Success” ran for about 50 miles
Second thought. Buy new main relay. Same “success.” Same ~50 miles
Third thought, check and as a result, change battery negative terminal (It was twistable by hand) Same “success.”
Fourth thought, Check engine ground “B.” Discovered stripped valve cover bolt receiver. Re-ran ground with a starter cable to the intake manifold, right next to the igniter ground. Same “success.”


The same symptoms returned after each one of these actions, each for around 50 +/- miles. While it ran for a while and then crapped out again I cannot explain. Any thoughts might be a help to the community
Time to get serious. I pulled both the ECU and TCU computers and examined each with a watch loupe.


Many years ago I started fooling around with the prospeed chips, even before the RDX injector swap was created. To keep the car running and not have it down for a week, I swapped out my ECU with a unit from Dan (yoojeankim.) I have no idea who installed the chip socket but I know it was NOT Brian. He never saw nor laid hand upon it. I've had a least 4 different chips in the car.
Under magnification the capacitors all appeared AOK and none of the solder joints showed any discoloration or signs of fracture. The chip socket installation, on the other hand, was a disaster (see picture above.) How it ran all these years is a mystery to me, and it was a painful final failure mode. Anyway, I reflowed the 28 joints, replaced both computers, started up the can and ran it in ever increasing distances and stress situations.


After 250 miles driven between 4500’ and 11,500’ above sea level, a few redlines in 3rd gear, multiple start-ups, temperatures ranging from 40 degrees to 95 degrees, I declare it fixed.


Takeaways: (1) I am satisfied with the trouble shooting path. Everything done was proper and logical, and I have probably short circuited a future problem or two. (2) Nothing in the original design of the NSX was to blame. (3) Any modifications to electrical circuits should only be done by people who know that the f*** they are doing. (4) Although the NSX is becoming a bit of a fossil, there are advantages in that things like computers can actually be repaired rather than simply thrown away.


Thanks to all the people who helped me in reaching this solution. NSX Prime is truly a community of interested and interesting people.


Be safe.


Mark

 

[furinax]

Glad you didn't give up on it. Given the ECU was busted, would you have considered going the standalone route?

 

[mskrotzki]

I never considered giving up. I may or may not be a reasonably bright guy, but I am a stubborn SOB and would have (will) stayed with it until I fixed it.

If this wasn't the fix, I'm not sure where I would have turned. Likely further investigation of the ECU, or picking up a used unit. Off the top of my head, a standalone unit, with the requisite dyno testing and programming sounds about as $$$ as a new replacement unit. But who knows? It is an interesting thought.