Intermittent EPS/MIL Codes

[Big McLargeHuge]

Hi all,

Figured I'd post a new thread here to start a discussion about an issue I've been having that I haven't been able to figure out.

Basically, after about an hour of driving my '90 NSX, some combination of things will happen, sometimes all at once, sometimes just one:
-Speedometer starts jumping around, comes back, sometimes stops working completely
-EPS light comes on, usually codes 31 and 33, or just 33
-MIL comes on, code 17

Once the EPS code 33 comes on, the EPS works until I turn the car off and on again, then I need to reset the codes to get EPS back. The same problem happens almost right after starting up again if I don't wait for an hour or more, perhaps for the car to cool off again. This leads me to believe this is a heat related issue, it only ever happens if I've been driving the car for a long period of time and goes away overnight with a clock fuse reset.

When I had the auto trans, I replaced the NM speed sensor with a new aftermarket one since I thought that would fix both the speedo & EPS problems but still the same issues.

I swapped out the auto trans for a manual trans that came with a "new" (used) VSS1, and transferred over the NC speed sensor so my EPS would still work. I also converted to the S2000 gauge cluster in the meantime. Even with all that, still the exact same issues.

Is it safe to say the NC speed sensor I transferred over from the auto trans to the 5-spd is what's causing my problems? That's all I have left to replace unless there's an intermittent short or something in the wiring harness somewhere that repairs itself overnight.

Can I take off the new NM sensor from the auto trans I replaced and use it in place of the old NC sensor even if their part #'s are different? I don't know if they're interchangeable, it might just be the wire harness lengths are different (NM is longer than NC wiring), but the NC sensor is twice as expensive so if I can re-use the NM sensor that would be great. They should be the same design as far as I know.

Thanks in advance for any suggestions.

 

[Big McLargeHuge]

Self update, I threw more money at it by buying a new NC speed sensor to replace the one I took from the A/T.

Initially I tried using two spacer washers instead of one, and the EPS light came on almost right away, so I figured one spacer is the sweet spot. BTW, the spacers are needed so that the tip of the sensor doesn't hit against the teeth of the larger NSX-R 4.23:1 ring gear.

The new sensor seemed to fix the problem, I went on some longer drives and didn't have any more issues, but the EPS light came on with temporary speedo issues once more last night after a particularly long outing. Still, it seemed to have lessened the problem somewhat but hasn't solved it.

Any other ideas to address the root cause? Maybe some bad contacts on the EPS rack itself? (doubtful with the speedometer symptoms). VSS1 issues? Overheated wiring?

 

[Old Guy]

I haven't followed your build thread in detail so I don't know the details of your modifications. I am also not up on the auto transmission arrangement other than knowing that the AT controller gets a speed signal from the VSS and signals from the NC and NM speed sensors. What confuses me about your comments is that as far as I can tell the NC and NM sensors are dedicated to the automatic transmission and would have been history once you installed the manual transmission. Or are you talking about just using the old NC or NM sensor to replace what you think is a faulty VSS? If the latter, I don't know that they are interchangeable. I thought that the NC and NM are reluctance sensors (2 wire devices) and the VSS is definitely a Hall sensor (3 wire device). Although it could be an optical sensor; but, those are quite uncommon.

Given the MIL and EPS codes, I think you have a VSS / VSS circuit problem. I know zip about the S2000 cluster; but, in order for your original NSX speedo to work you had to be hooked up to a Hall type 3 wire sensor. That VSS device has some electronic circuits in it. These tend to be pretty reliable devices; but, it is possible that heating is causing an intermittent failure in the device (solder bridge opening up?). I would be inclined to replace the VSS with a known good / new VSS and test it out. If you search on Prime you can find links to economical replacements for the VSS.

The problem may not be in the VSS device itself; but, in the circuit. If the connector to the VSS is dodgy you could be getting a temperature related problem in the connection that is causing loss of the signal. As I recall the output of the VSS is an open collector transistor. It needs a pull up voltage from one of the receiving devices - odometer drive circuit or the ECU. If something happens to the pull up voltage, even temporarily, the signal disappears which can cause your error codes. I think that the pull-up signal originates in the ECU so if you are not having ECU problems (other than code 17 ) I think that the pull-up voltage disappearing is unlikely. A temporary short circuit on the VSS wiring would cause loss of the signal and it could be thermally related.

Short answer. I would get a known good / new 3 wire VSS sensor (not 2 wire NC sensor) and see if that makes the problem go away. If the problem persists then I would start checking the wiring and connectors associated with the VSS circuit. If you can't find a wiring problem then its time for the heavy hitting. I would purchase one of those low cost portable digital oscilloscope boards (you can find cheapies for less than $50) and hook it into the VSS circuit to monitor its operation (check for the presence of the speed pulse signal). If the VSS signal disappears just before the error conditions appear, then you know what the source of the problem is. Why the signal disappears will be subject to further testing.

 

[Big McLargeHuge]

Thanks for the reply Old Guy.

In short, the auto trans has the 3-pin VSS1, NC sensor (or VSS2 on M/T), and NM sensor, yes. When I swapped my auto for a 5-spd, I transferred over the old NC sensor from the auto to use it as the VSS2/diff speed sensor on the 5-speed which is required to send signal to the A/T controller and have the EPS still work, as you say.

The 5-speed I got also came with its own used VSS1 which I am now using. So the original VSS1 and NM speed sensor from the auto are off of the car. This is the reason I haven't bought a new VSS1 yet, since I figured the chances of the VSS1 that came with the 5-speed also being bad were low.

I had the same speedometer issues with the original NSX cluster, and the problem carried over when I swapped in the S2000 cluster. Once the speedo starts jumping around, the EPS light is usually not far behind in either case.

Long way of saying I agree that a replacement VSS1 seems like a sensible next step, it just so happens that I chose to replace the NM & NC sensors first which was probably unnecessary. I've been a bit hesitant to do oscilloscope testing partially because the EPS troubleshooting steps are missing from all the online NSX manuals I have, and the conditions under which the speedo starts freaking out are very intermittent and tends to fix itself after a few minutes of driving, so catching it with a scope seemed difficult (plus I don't have one). If I replace VSS1 and still have issues I'll get a scope and learn how to do real diagnosis, lol.

EDIT: I think the cheaper alternative VSS1 is 78410-SV4-003, correct? I remember reading something about it having the male terminal on the sensor body itself instead of on a wire pigtail but can't find the thread now. Not sure if the stock female harness terminal will reach that far or if I need to make a short patch harness myself, either way it saves like $180.

 

[Old Guy]

Here is at least one of the threads with the VSS interchangeability discussion.

http://www.nsxprime.com/forum/showt...rchangable-with-any-other-Honda-transmissions

The following part numbers will get you search hits on non NSX applications that have the correct pigtail

78410-SY0-003
78410-SR7-003
78410-SM4-003

There are likely other numbers. Unfortunately, it looks like a lot of the aftermarket versions are NLA. The thread mentions that you can reposition the harness to 'make' the later non pigtail sensor fit.

Being a little slow on the uptake, I now understand your reference to transferring over the NC countershaft speed sensor to your manual transmission case. I gather that even though you have deleted the AT, you have left the AT controller in place and are letting it do the signal processing on the NC r eluctance sensor so that it can be used by the EPS unit. The 96+ EPS cars with manual transmission use the reluctance sensor in that position but use a module called a pulser to condition the signal for the EPS. I was perhaps confused by the reference in your first post to changing out the NM and NC sensors to address the problem (this problem with the EPS and speedo predates the manual conversion?).

My first observation is that as far as I can tell, problems with the VSS2 reluctance sensor on MT cars or NC on AT cars should not cause speedometer problems or the ECU VSS error code. Speedometer erratic behavior and the ECU VSS error code should all be related to the VSS1 sensor or its wiring. The VSS1 and VSS2 circuits appear to be separate and unrelated (except that they are both used by the EPS); however, that is based upon the Service Manual and the SM does have some errors. I think your speedo and ECU error code problems are purely in the VSS1 sensor / circuit. Replacement of the VSS1 with a known good is the easy (although maybe not the cheap) first step. If that doesn't address the problem then you are in to chasing wiring problems. For that I advise getting an NSX Electrical Trouble Shooting Manual (if it exists for the 1991). The wiring diagrams have much more detail on plug connector numbers and pin assignments and the manual has photos of the locations of all the connectors which make tracing the VSS circuit through the car easier (not easy).

In your second post I noted your reference to the R final drive. I thought the R final drive and EPS was a no-go mix because the reluctance speed sensor could not be made to operate reliably with the R final drive? I remember reading something about Science of Speed coming up with a fix of some kind to allow the use of EPS with the R final drive; but, I don't recall that going anywhere, although I didn't really follow it because I did not have any interest in the R final drive.

 

[Big McLargeHuge]

Thanks for the thread reference, I ordered one 78410-SV4-003 for $10 that says it's for a 93+ NSX because of the lack of wire pigtail. I also got a cheap pigtail kit made for other Honda VSS that will let me chop off the old VSS male connector and make my own crimped patch harness if I can't move the factory harness far enough.

I have left the AT controller installed purely because it's required for EPS signal processing as you say. All the EPS/speedo problems occurred before the manual conversion, yes, which is why I thought changing to M/T with a "new" VSS1 would fix the problem. I sorta knew the VSS1 was the one driving the speedo circuit but figured the possibility of both the old VSS1's I had being bad were so low.

I have the ETM for 1994 which I've used a bit already, but I think (and hope) the new VSS1 will fix everything.

For the R final drive, I did some digging before I swapped transmissions because I was also told by SoS that it would not work with any year's EPS (I was also told by them the swap would be impossible without a wrecked car but that was proven incorrect as well). However, I messaged Blake @ Speedhunters who had his NSX swapped in Japan by Advance with the R final drive and he said his EPS still worked just fine. Because of that and another past Prime post by LarryB also corroborating that experience, I went ahead with the swap and can confirm that the R final drive does work with the 91-94 EPS racks/computers. I do not feel any difference in steering effort from before, but it's possible the computer might be providing different levels of assistance with the different ring gear/sensor signal generation but it's imperceptible to me. I know Honda changed the EPS design in 95 and again (I think) in 97 along with the 95+ differential design which would not work with the R final drive.