Honda VFR800 FI system acting badly and I'm lost

This is a repost from another forum that I'm on that is model-specific to the Honda VFR. I'm posting here for ideas from the GRM Hive because there's a lot of smart folks here and maybe someone has a suggestion. If nothing really comes out of this post, well, that sucks but I get it. Here goes;

I've got a 2004 VFR800, non-ABS, with about 54k miles on the clock. Nearly a full year ago I took my 6th gen off the street and started to very slowly do some overdue maintenance as well as electrical improvements. Basic things; new thermostatic switch for the radiator fan, manual over-ride switch for the fan in case I have a problem with the switch, a bus bar for accessory grounds. Nothing really earth shattering there. New plugs, new cooling hoses almost everywhere, a new fuel filter.

A couple days ago I had the bike in a state where I could try starting it again. I have the airbox off the bike, the velocity stacks installed, and hoped I would finally do a starter valve sync in the next couple days. All sensors are installed, the PAIR valve is attached electrically, and I've made sure to the best of my abilities that I have all of the vacuum lines attached properly. Gallon of fresh 91 in the tank, hit the starter switch, and it started pretty easily.

It does not respond to throttle well at all. If I really slowly roll in some throttle opening the bike will rev but it's not happy. Quick opening of the throttle usually kills the bike. I let the temp get up to about 135F on the dash display and called it a night, knowing I had more troubleshooting ahead of me. The feeling I had was TPS problems of some sort.

Last night I go out, put the side stand down, and start the bike. Sure enough, 8 blinks of the FI light, which the manual says is a TPS fault. I grab my old digital voltage/ohm meter and some T-pins and start going through the manual's walk-thru for troubleshooting. With the ECU connected I have voltages on the yellow/red and gray (I think? I'm at work, doing this from memory) wires at the TPS with the key & run switch ON. These voltages were nicely in the ranges described by the service manual.

With everything turned off and back probing pins at the TPS I have a change in resistance as the throttle is twisted; I believe the range was about 500 ohms closed, 1.5k ohms wide open, and the change in resistance seems pretty linear and smooth. I have continuity through the wiring harness; with the TPS cable side plug pulled, each conductor rings through from the pin in the connector at the TPS to pins at the ECU connector. So unless a gremlin has done a depin/repin job on the connectors to the ECU, I believe there's good paths from the ECU to the TPS, and I don't see any pushed back pins or damaged terminals anywhere.

My brain says that I have a damaged ECU somehow. My gut says this is nonsense because how would that have happened? I'm not sure how to proceed. I see used ECU's are about $300 on ebay which is an awfully expensive part to purchase not knowing why I'm having to replace it. I have not spoken to my local Honda (and other makes) motorcycle shop about this at all so I'm not sure if they can test the ECU somehow.

What do I do next?

Have you tried disconnecting the manual over-ride switch for the fan from the wiring harness to see if the problem goes away?

If it is raising or lowering the voltage on a pin for the ECU when the engine is running it could cause the the ECU to interpret data incorrectly causing issues.

At least then you can rule out your modification is contributing to the problem.

In reply to jstand:

ECU doesn't control the fan. Fan control is done by a thermostatic switch in the radiator. Think early-90's control tech.

EDIT: You know, I may try to find the old switch and just swap them with the fan unhooked. As long as I don't run it for a really long time, the coolant won't get overly hot, and I can see if there's somehow an effect from the change of the switch I added and the over ride vs. stock. Will try that tonight.

If it were me 1st thing I'd do is open all the electrical connections one at a time and spray them with DeoxIT® Gold from CAIG Laboratories stuff works wonders ask anyone with an older English or Italian bike and or the strange guitar player down the road. No I don't work for them, but I use it a lot on my Ducati and other bikes an even at work now too.

I hit the connectors with CRC brand contact cleaner already. Everything looks good.

I pulled the ground side wires from both the fan switch and the over ride. No affect. Then I pulled a couple more sensors from their cabling. Added a MAP sensor error which it should have done so the ECU hasn't completely lost the plot. Put it all back together and I still have a TPS fault.

Wondering what to do next.

Okay, cats and kittens, here's where I'm at. I believe I have determined which of the three wires does what function at the TPS. This took some thinking at my desk this morning and I wish I had done more testing last night. Took a lot of notes last night while going through troubleshooting steps in the factory service manual and here's some of the interesting stuff.

With the connector pulled from the TPS, the key and switch on, I get the follow voltages in reference to chassis ground; Signal wire to Ground - 4.92VDC (This has me concerned) Ground to Ground - 0.04VDC 5V Ref to Ground - 5.05VDC

With the connector attached to the TPS, throttle closed; Signal to Ground - 2.28VDC (This has me concerned) Ground to Ground - 0.03VDC 5V Ref to Ground - 5.04VDC

With the ECU pulled from its connectors there's no path to ground on any of those conductors. With the TPS attached, ECU connected, key on I get the following voltages; Signal wire to what I believe is the 5V Ref wire - 2.69VDC with throttle closed, 4.62VDC with the throttle wide open. Signal wire to what I believe is the TPS ground - 2.31VDC with throttle closed, 0.40VDC wide open.

What's going on here? The TPS seems to be acting badly. OR I have a wiring fault. What should I do to further troubleshoot this thing? The thing that has me wondering is that signal to ground voltage of 4.92V with the cable connector pulled from the sensor. Could be an induced voltage by laying next to the 5V reference wire in the harness assembly. Anyone have any thoughts on this? Honda doesn't sell the TPS seperately, which is apparently an issue with all of their products, althoug there are some ebay sellers of TPSen that should be swapable.

The sensor grounds are likely isolated in the ECU from the chassis ground. The sensor signal voltage is probably floating when unplugged with a pull up resistor and then that voltage is pulled down to ground when its plug in through the potentiometer of the TPS. The TPS then pulls that voltage back up as it turns.

I would be concerned about what the voltage is when the throttle is totally closed. The manual should give you the values for fully closed and WOT. Did the sensor get twisted in the throttle body? Chech all your grounds. Test you voltage against the sensor grounds not the chassis.

Hope this helps.

In reply to badfrogg:

Agreed with your first paragraph. Voltages with the connector attached have me more concerned.

Second paragraph, Honda's manual has the tech working against both chassis and sensor ground. The voltages I saw with the ECU connected and powered are all screwed up to the point that I'm not sure how great I was at identifying the purpose of each wire. It seems like they used different colors for their motorcycle EFI wiring vs. the cars.

Grounds all seem good but I'll try to check again to be sure. I've also looked around and it seems like there are other bikes from that era that use the TPS from DPFI cars of the late 80's to early 90's. I'm going to try and pick up a car TPS locally this evening on my way home from work. Even if the direction of rotation is incorrect and I need to find a different one for a real fix, I can check how it works electrically against my ECU to see if this kills my CEL.

Well you identified which wire is ground. And the wire that is always 5 volts is the supply voltage. The third wire should change in a linear fashion from low to high as you rotate the throttle. That the signal wire. You just need to verify what low should be. High should be close to 5 volts.

In reply to badfrogg:

The manual states that it should be .4V to .6V closed and 4.2V to 4.8V fully open. This is pretty well inline with every other TPS I've read about. The Signal wire, I believe, is the Lt. Green/Yellow wire at the TPS and ECU. The problem is that this is the wire pair that showed 2.31V with the throttle closed and 0.40V wide open. Which is completely wrong. If it was 4.8V with the throttle closed and 0.40V open things would somehow be backwards. This is different.

I'll check both the Light Green/Yellow wire against Green/Orange as well as Red/Yellow against Green/Orange tonight and then Green/Orange against chassis ground to double check my theory on the wire purposes.

I think you are reading the wrong pair. Those readings sound like the 5v supply and the signal. You want the ground and signal pair.

I think your really close to diagnosing the TPS. If it checks out okay then look for vacuum leaks. That can give a false positive for the TPS.

In reply to badfrogg:

The bike wasn't running for any of these tests after the first "verify that you still have a check engine light" test. After that it was jacked up voltages between the wires I expect to be the signal (Lt. Green/Yellow) and sensor ground (Green/Orange). I'll do the above checks that I listed. Everything I've done and re-read tells me that the Lt. Green/Yellow wire is the signal wire telling the ECU what the TPS's position is. And that's the pairing that had 2.31V with the throttle closed and .4 with it open.

pres589 wrote: 2.31V with the throttle closed and .4 with it open.

That would be a very odd failure mode for a TPS but if that is the case then it would definitely be a problem.

Okay, did some more testing this evening. Here we go, with the key on / run switch on, and measuring at the connectors at the ECU;

Lt. Green/Yellow to Red/Yellow wires - 2.72V w/closed throttle, 4.63 w/open throttle. 4780 ohms w/closed throttle, 1340 ohms w/open throttle.

Green/Orange to Red/Yellow wires - 5.01V w/closed and open throttle, 420 ohms w/closed throttle, 687 ohms w/open throttle.

Lt. Green/Yellow to Green/Orange wires - 2.28V w/closed throttle, 0.38V w/open throttle, 4750 ohms w/closed throttle, 750 ohms w/open throttle.

I believe this narrows down the TPS as the culprit. As for the TPS itself, it's marked as a "JT3L 31105". Vehicles using a JT3L seem limited; the Honda Passport from 2003 to 2005 seems to use a similar unit but I don't know if the direction of rotation is the same or not. I'm really not impressed by how Honda makes these units harder than they should be to replace. Anyone replaced a TPS on their VFR800?

pres589 wrote: Lt. Green/Yellow to Red/Yellow wires - 2.72V w/closed throttle, 4.63 w/open throttle.

This looks like the correct test and it looks like the TPS is bad (2.72V w/closed = Fail). from all your test results i conclude that the following is the pinout:

Red/Yellow - Ground

Lt. Green/Yellow to Red - Signal

Green/Orange - 5V

Dont bother with the resistance reading with everything connected. If you want to get accurate resistance readings just test the TPS unplugged from the ECU.

In reply to badfrogg:

Discussing this on a VFR forum and measuring between Grn/Orn and Chassis myself, the Grn/Orn wire is ground, Red/Yel is 5V Ref, and Lt. Green/Yel is Signal. The Ground and +5 input will look the same from a voltage standpoint.

The plan for tomorrow is to get a TPS meant for a 2004 Honda Pilot and see what I can see without taking it out of the store. My hope is that the direction of rotation is the same, otherwise I have to monkey with the switch, which is explained over at the Wrist Twisters forum for a similar swap between a 919 and a DPFI Civic TPS.

Oh hey look who didn't get his bike fixed a year ago and is back at it again.

MAP sensor is outputting .3 volts on the signal wire with a good 5V input from the ECU and the ground connection rings nicely to the chassis. ECU is still not reading the TPS on this thing. I'm going to try and find a source for a new MAP sensor. In the mean time, what the hell is wrong with my bike??? I'm trying to work on this stupid thing every day but I've thrown a lot of new factory wiring at it sourced from a dealership, three TPS sensors, and a pair of ECU's at this same issue.

Have you undone your mods? Sounds like the problems started after you did work including adding some grounds.

I would inspect the areas around where you worked for disconnected or compromised connectors, wires or hoses.

It sounded like a vacuum leak initially as well so double check for that. Spray starting fluid or your aerosol of choice around and see if idle is affected.

I don't think they run worth a berk with the air box removed.

All mods have been undone. The ECU flags a TPS fault before the bike is started; electrically the bike is somehow not right, so running without the air box or vacuum leaks are a step beyond where we're at now. I didn't add grounds, I added a more direct path between the regulator and the battery for 12V+, which was done initially a few thousand miles ago in the middle of a trip in 2014.

My current theory is that the ground side for the TPS isn't there [electrically] but I haven't worked on the bike in a while because I hate working on that bike.

pres589 said:

All mods have been undone. The ECU flags a TPS fault before the bike is started; electrically the bike is somehow not right, so running without the air box or vacuum leaks are a step beyond where we're at now. I didn't add grounds, I added a more direct path between the regulator and the battery for 12V+, which was done initially a few thousand miles ago in the middle of a trip in 2014.

My current theory is that the ground side for the TPS isn't there [electrically] but I haven't worked on the bike in a while because I hate working on that bike.

Did you ever get the bike to run good again? I have aquired one that seems to have the same symptoms but no FI codes, TPS ohm readings and voltages all check out with the Honda service manual as well. I can actually crack the throttle wide open and the bike won't even rev past idle just kind of chugs along at around 1500rpm even though the blades are wide open. If I slowly rev it it runs great even all the way up to 10k just dead for throttle response though. Thanks in advance.

In reply to jwilczek :

No, I never got this to work right again, and I sold the bike to someone else on this forum. I don't think he's done anything with it either.  Mine would not be convinced to rev much past 1000 RPM in neutral. 

They're cool bikes but I didn't enjoy working on mine much at all.  And I still, kind of, want another one.  Best of luck with yours.

I believe I own this bike now.  I am posting here for easy location of previous troubleshooting whenever I get around to it.  

In reply to nocones :

Red 2004 VFR800, has a black weather-resistant toggle switch to the left of the instrument panel?