The problem car is my '94 Si hatchback with JDM GSR engine and USDM P72 ECU. Both are stock.

Earlier this year, I pulled my trans to rebuild it. Like a boob, I forgot to unplug the VSS and managed to rip the wires out of the harness plug. I was able to repair it and everything seemed fine. Except for my tach.

My tach has been reading wrong ever since. Shortly after my mishap, the tach would stop between 7000 to 8300 rpms at the rev limiter. For a while, I thought the actual limiter was moving and preventing the engine from revving higher. It wasn't a big enough deal for me to pursue so I let it go until I was done for the season. Since then, redline is consistently showing 7500 rpms (sometimes 7200).

I went to Gingerman Raceway and ran four 20 minute sessions. On the way home, I noticed that my rpms were really low for the seed i was driving. I shifted down and ran to the rev limit and it stopped at 6000 rpms. The engine was definitely revving higher than that.

Here's what I've done:

Researching the wiring diagrams I found that other than ignition power, there is no correlation between the tach signal and the VSS. I unraveled my harness and I found no broken, cut or shorted wires.

I used a GPS app to make sure my speed was reading right and it was.

I used a Fluke meter to check RPMS and it showed that I was getting 8300 rpms while the tach showed 7500 rpms.

You would think it was the tach. No. I bought a used OEM tach and it showed the same rpms.

After I tried that, I tried running a wire directly from the ICM tach output on the distributor to the tach and ran a separate power feed directly from the battery (easy to do with the spare tach). Same problem, even read fewer rpms.

The next part in line is the ICM. I went out and bought a new ICM and I still have the exact same problem.

Great thing about buying a part that I didn't need is that I had a spare ICM that I needed to revitalize an old dead distributor.

I threw my now good spare distributor in and I still have the same problem.

One other thing I need to note, this is a bone stock P72 ECU. VTEC changeover is 4400 rpms and IAB changeover is 5750 rpms. Right now, the changeover for vtec is indicated at 4100 rpms and the IAB changeover at about 5300 rpms. I really believe the ECU is happy with it's distributor crank sensors and it's doing what it's supposed to do.

Oh, I also swapped in a spare VSS with no changes. I am also getting good voltage from my alternator.

I don't think I missed anything so that's all it in a nutshell. The only other parts I can think to change out are the engine harness or an ECU of which I do not have spares. I need to try and find a timing gun with tach to verify engine speed.

I've never been so stumped by a problem. Any help would be greatly appreciated.

Thanks,
Alex

 

What wire did you tap the Fluke into to measure engine rpm, which you believe was the correct reading? Do you still read what you believe is the correct rpm using the Fluke?

 

I used the test lead (blue wire) on the driver's side of the firewall as well as the blue wire on the back of the cluster. The Fluke was reading 8300 rpms while the tach showed 7500. The error starts to show itself shortly after 2000 rpms. At 2500 on the tach, the fluke showed 2650 (approximately).

When I ran separate wires to a spare tach, I disconnected the engine harness lead to the distributor (blue wire) and ran a wire directly to my spare tach. Had the same problem.

 

Did the Fluke read the "correct" rpm from the separate wire that you ran?

 

I didn't try reading it from the separate wire that I ran. Seems irrelevant since I was basically getting the same reading on the tach.

 

The relevance of testing the separate wire with the Fluke is simply to verify that the wire is providing a good rpm signal.

You haven't addressed the tachometer ground wire. The tach requires power, ground, and signal. Your tachs may have a ground problem. What are you using for ground? Have you tested it?

Verifying rpm by an independent source like a timing gun is definitely a good idea.

 

I see your point by testing the test wire.

Are you referring to the ground wire on the cluster or when I set up the test? The extra tach was powered by the battery. I ran 14 ga wire directly from the batter to the tach and then ran a lead from the distributor to the signal terminal.

Another point that I didn't make is that everything else, the speedo, other gauges, and the power output of the engine are all normal. If it was a power issue to the ignition coil, then I would think there would be issues with power output especially at higher rpms.

 

The tach in the cluster shares common power and ground with many of the other gauges; however, there's a short segment of power and ground on the circuit board that is specific for the tach. It's possible that the cluster circuit board is bad.

I'm confused here. If the extra tach is powered by a battery, then why would you need to run a 14 gauge wire to the battery?

Have you tested system voltage when the tach is malfunctioning?

 

I would agree that the cluster board could be bad if I didn't have the exact same problem with my test setup. As for the wires, I was just talking about how I wired my test setup. Picture wires coming from the battery and the blue wire off of the distributor going around the hood and door and sitting on my passenger seat. My test setup completely bypassed any wiring in the car.

I never tested the voltage while it was malfunctioning. I tested while it was idling. I did have 14.5 ish volts at idle.

If the problem were involving a ground, I would think that my test setup would have bypassed any issues unless the problem involves my engine ground. I added a couple of extra ground leads a few years ago from the block to the battery.

 

When you say test setup here^, do you mean this:

If so, what exactly did you connect the battery wire and blue tach wire to? And does this test tach setup also use a battery? Can you post a picture?

It might be a good idea to monitor system voltage when the problem is happening - just a thought.

 

They always say, a picture is worth a thousand words:

http://importnut.net/thehatch/testtach/

I hope you can see everything in the pic. Like I said, I ran the power right off of the battery, and the tach signal right off of the distributor.

I will have to check the voltage during use another day.

What are the chances there is something related to the ECU? It's a huge long shot. There is one lead that come from the ICM to the ECU according to the helm manual.

 

Okay, you are wiring both power and ground directly from the battery to the test tach. That tach seen on your cowl cover is different from the tach in your cluster, right?

The Fluke, on the other hand, has its own power and ground source separate from the car battery when its measuring rpm, right? But the Fluke uses the same blue wire from the distributor, right? If so, this suggests that the problem is with power or ground to the tachs as this is the only difference from the Fluke.

If the ECU or Yel/Grn wire were malfunctioning, then the tachs and fluke would yield identical rpm readings.

 

That is the spare tach. Originally, when I did this, I used my original tach and had the spare tach in the cluster. Same results either way. [/QUOTE]

The Fluke did have it's own power source. One interesting point I should ask about is the Fluke was giving me a tach read-out without having to attach the ground lead to the car. Meaning, I could get a tach signal by only using the positive lead on the meter. If I attached the ground, I had the same readings.

I did use the blue wire with the Fluke. I read it off of the test port under the hood then I tried it again where the blue wire goes into the back of the cluster.

How can I have a problem with power and ground when I'm using the battery for power? For my test, should I run a ground lead from the negative terminal to somewhere around the distributor?

Aha! Okay, that's good.

Thanks so far!

 

Can you post the section of your Fluke manual that gives directions on how to take tach readings?

If your car's charging system is not providing a steady voltage, then this may interfere with the tach readings.

Do you have a spare car battery that is fully charged? If so, connect the power and ground wires of the test tach to your spare battery instead of the battery installed in your car. Then connect with the blue wire to the test tach and check whether the tach reading is now correct.

 

I borrowed the fluke. I'll have to see what I can find.

I'll have to try the other test tomorrow. It's getting late.

I'll let you know what I find out. Thanks!

 

Ugh. I ran my test tach off of another battery and I was actually reading lower rpms than the tach in the dash. I used the test lead near the firewall for the signal.

This is driving me crazy. This week I'm going to try and get a different digital meter tach as well as an aftermarket tach for more experimenting. I need to try and find a digital timing gun as well (without having to buy one).

Any other suggestions?

 

I'm dusting off this old thread to try and get some more input to the problem. Read the initial post to get an idea of what's going on. Here's what I've done since then.

Last fall, I came to a conclusion that I had somehow damaged the engine harness. I picked up a used harness this spring and installed it.

No change in the problem.

I bought a cheap sunpro tach to see if it would read differently and that was basically inconclusive. It read 7800 rpms at redline where the stock tach was showing me 7200.

At this point, I needed to know, without question, what my engine was doing. I took the plunge and bought a digital timing gun. With this, I am reading rpms off of the spark plug wire which eliminates any erroneous from any of the cars wiring. This is where it gets interesting. Right off the bat, I was reading higher than the tach. IAB and VTEC kicks in right where they are supposed to according to the digital timing gun however, I am only getting 8050 rpms. The fuel shut-off is at 8300.

At this point, I wanted to start eliminating the wiring as an issue. I made up a few wires with the correct pins to run directly from the ECU to the distributor wiring. I started with the ICM (ignition control module). No change. I then did the same for each sensor, CYP, CKP, TDC. No change. I didn't have enough wire to bypass each sensor at the same time but I have a feeling that I will not see any changes.

To recap, I've tried different distributors, ICM's, tachs, bypassed tach wiring, bypassed ICM and all three distributor sensors and I'm still reading the wrong RPM as well as not reaching the correct rpm.

The only thing left is the ECU. I'm trying to get my hands on one. Has anyone heard of Honda ECU's having problems (when used in a basically stock application)? OBDI Computers are pushing 20 years old at this point!

Any help would be greatly appreciated.

 

Not sure what your logic is to make you believe that the ECU may be responsible for the problem.

How do rpm readings compare from the timing gun, Sunpro tach, and stock tach in the cluster.

At this point, I think you should try swapping in a different known-good cluster.

 

I say ECU because it's the last thing I haven't swapped. It doesn't make sense so I'm not 100% convinced. As for the cluster, I've tried two separate honda tachs with similar results. I've bypassed all of the wiring in the car and ran the tachs off of the tach output on the distributor (the blue wire) There are links in this thread to the pictures I took of that setup.

The stock tach will read different numbers depending on what mood it is in or the alignment of the starts (since I've run out of rational explanations...), between 6000 rpms and 7500 rpms. It's been hitting 7200 recently. Also recently, while the stock tach reads 7200, the sunpro will read 7800 (ish) and the timing gun will read 8050.

Since I've bypassed all of the wiring for each sensor (by running a separate wire directly from the ECU to the pins on the distributor), with the same results, and include the fact that I'm not getting to the 8300 rpm fuel cut, I go back to the ecu.

-Alex

 

The stock tach circuit passes through the cluster. I think the cluster is a candidate because your Fluke and timing gun appear to be reading the correct engine rpms.

The stock tach/cluster is problematic, making me return to the cluster as the potential source of the problem. The Sunpro is probably just inaccurate and the timing gun (and Fluke) is probably giving the real tach reading.

The fuel cut discrepancy is likely a separate issue from the tachometer problem.

 

You refer to a circuit in the cluster. I'm a little confused since the only thing I see in the cluster are the three separate "lines" that feed power to the tach and the actual tach signal. Since I've run the tach (both of the Honda units) independent of the actual cluster, I don't see the impact of the cluster itself.

I trust your experience here so I don't doubt that I could possibly have two bad tachs. I don't doubt the sunpro is inaccurate. It only cost me $30 new.

Any thoughts? ECU? Issue with the distributor?

A friend of mine is sending me a spare ECU, so I'll be able to swap that out for no cost. I'm reluctant to buy another cluster simply because I've already spent more than I want to try and tackle this persistent issue.

I definitely appreciate your input!

-Alex

 

The tach power, ground, and signal wires pass through the plastic cluster printed circuit sheet.

I agree. I forgot that you bypassed the cluster and got the same result with the stock cluster.

Given your overall findings, however, might the explanation simply be that the stock and Sunpro tachs are just less accurate than the tachs built into the Fluke meter and timing gun?

The CKP sensor inside the distributor sends its rpm signal to the ECU, so keep in mind that the ECU sees a different rpm signal than the tach, which gets its signal from the igniter unit (see below). One of those two components or one of the CKP wires might be "bad". Alternatively, it's also possibly just an inherent inaccuracy of the CKP and/or ECU? You can Ohm test the CKP sensor and it's wires, but you'd need to swap the ECU to test it.

CKP --> ECU --> igniter unit --> tach


.

 

I would completely agree that the Fluke Meter and timing gun are more accurate. My only issue is that my tach was accurate, showing roughly 8300 rpms, (and 9200 rpms when I still had my hotrod engine several years ago) until last spring.

I thought my mishap with ripping the VSS wires out started the issue since the problem started almost immediately after. Swapping the engine harness out didn't change a thing so now I'm totally stumped as to what initially caused it. Early on in the spring, the problem would intermittently go away and it would work fine. After a month, it has been persistent. Again, this was last spring.

I have become very familiar with the ignition system. I've been pouring over my Helm manual trying to find anything that could lead me to a solution. As I mentioned, last night I made temporary wires to run from the ECU to the distributor with no change. I could make all 7 to try them all at once, but I don't know if I'll see anything different or end up with issues since my temp wires won't be shielded.

I need to wait for the spare ECU to show up and then go from there.

Of the three sensors, is it just the CKP that specifically sends a signal to the ECU and back to the ICM? I was wondering about that last night...

Thanks

 

Yep

 

I'm dusting this one off for another round. <sigh>

The odd-ball donor ECU that I borrowed (the JDM Auto P72 ecu) has a different pin pattern than every other ECU you can toss in. Once I sorted the three wires out, I was able to try the ECU to see if my tach behaved any differently.

Well, first, I wasn't sure. It seemed like VTEC was kicking in right about where it should and then the IAB's as well. Then, after driving for a bit, it went back to the same problem.

This brings me back full circle. I'm missing something...this is ludicrous.

Ron, anyone, any insight would be great. Since I'm basically going to have to start from point A...

Thanks,
Alex