I've got an 09 Civic Si coupe here at my shop that had a turbo kit installed, customer blew the motor (turned boost above tuned limit) so upon initial inspection I dropped the subframe down enough to just remove the oil pan. To get it low enough the steering connection disconnected from the steering wheel arm. However, now the motor is back in after being rebuilt and I reinstalled the steering arm into the steering wheel arm, and I made sure that the alignment arrows were lined up. Also, I don't know if this matters but the axles are not in yet and both abs sensors are disconnected because they are bad and new ones are on the way.

Now, here's the problem, once I start the car the steering immediately jerks all the way to the left. If I move it to the other direction it jerks it right back all the way left. It's almost like the wheel centering is off or something. Why??????????

Thanks

 

Is the EPS mil on? All connectors ok? Grounds ok? There might be possible torque sensor damage Or steering angle sensor issues. Matter of fact it could be numerous things in the eps circuit including the control unit. As fas as the column shaft u joint and the steering rack itself, that has nothing to do with anything described in your problem. The only thing that can be 'screwed up' is the toe/relation of the steering wheel. Hm. let me know if you need dtc descriptions or troubleshooting info (connector views, component views, terminals, values, ect.)

 

No EPS light that I can tell. I hooked our Modis up to the car and the only code I can find is a U0122, which from what I've read is related to the Hondata flashpro. All connectors seem fine. ABS light is on because the lack of front sensors, would that cause this?

I miss the pre-01 hondas...

 

abs/tcs/vsa fcan code is what that dtc is. I have Flashpro and do not have any body electrical dtcs relating to fcan (see note belwo). That is either a true problem and is causing your ab/vsa mil to stay on or, you say, you have no front wheel speed sensors on this car? Why? And what would cause what?

As far as the eps issue goes, you should be able to pull up a data list once you can communicate with the eps control module with your modis. Check the current fluctuation and torque sensor valus for discrepency. The modis should give you basic operating parameters assuming system is OK. Obviously, further diag is reqd. Again, let me know if you need adv info (please be specific)

(note; the vsa mil may be on due to flashpro on different year si's. Although 'all the same', pcm contruction slightly varies. The oem Etm will not show this, however. So, what this means; that U code (fcan/body dtc) may be infact due to flashpro. But the speed sensors missing, well....)

 

Ok, so I got the new abs sensors and axles installed. ABS light went away of course but the steering is still doing the exact same damn thing.

Kingteg, please send me any info you can, sensor locations and connector views for eps system. I don't see any section on the Modis in regards to EPS either, our Modis is kind of old so that might be why. Anyway, my email is undilutedcarguy@yahoo.com

Greatly appreciate your help.

 

On the rack itself there is only the torque sensor (3p connector) and the motor angle sensor (6p). As far as sensors go. Which connector views would you like? @ the component or control unit?

If nothing else was damaged or touched, then my money is on either a bad input (torque sensor) or faulty control unit. The circuit is more simple than youd think. Also, there is several failsafe modes that will not kill assist.

Be aware, when you replace the eps ecu, you will need to do the relearn procedure. If your modis cant do that, you will have to take it to us (Honda) for relearn.

And just to clarify, the car is assembled, wheels on ect, on the ground, you turn the key to ON (ll).....what happens? Nothing should happen and you should have NO assist. The eps ecu requires signal in pulse width from the pcm of engine rpm. You turn the motor on and the rack/steering wheel/both wheels turns itslef all the way to the left/lock without you touching it? and if you give it force to the right it will jerk back to lock? How quickly? How much force?

 

How do I test the torque sensor? I would just think that I'd be getting an eps light, but I'm not, which I don't understand.

Clarification: Everything installed, car on the ground, the key is just on, "ON" then nothing happens. When I turn the motor on it turns itself all the way left/lock without being touched. Yes, when I turn it the other way it immediately jerks back to the left, with enough force as if the steering was recentering itself after a turn.

 

Is it possible that I need to do this procedure? (copy and pasted from alldata)

Memorizing The Torque Sensor Neutral Position

The torque sensor neutral position must be memorized whenever the steering gearbox, the torque sensor, the EPS motor, or the EPS control unit is replaced. Note that the torque sensor neutral position is not affected when erasing the DTC.

NOTE: The torque sensor is temperature sensitive. This procedure should be performed within the range of 68 °F ±18 °F (20 °C ±10 °C).



With the ignition switch in LOCK (0), connect the HDS to the data link connector (DLC) (A) located under the driver's side of the dashboard.
Turn the ignition switch to ON (II).
Make sure the HDS communicates with the vehicle and the EPS control unit. If it doesn't, troubleshoot the DLC circuit: Except Si model, See: Powertrain Management\Computers and Control Systems\Data Link Connector\Testing and Inspection Si model. See: Powertrain Management\Computers and Control Systems\Data Link Connector\Testing and Inspection
From the EPS MENU, select MISCELLANEOUS TEST, then TORQUE SENSOR LEARN, and follow the screen prompts on the HDS. NOTE: See the HDS Help menu for specific instructions.
Turn the ignition switch to LOCK (0).

 

Honestly I doupt it, but Ive seen stranger things. Regardless, the HDS is what we have here at the dealer so thats gona be an issue.

The blu/red wire is your ref volts. 5ish v constant supply. The other 2 wires are the outputs to the eps ecu. Ill have to pull up the pcb schematic for the eps unit to see if these inputs will affect judgment on moving the motor (on its own to that affect). I dont think so tho. -This will take some time. I believe either the ecu is no good, or the motor is faulty. Magnet sticking or something wierd but more likely the ecu is giving a wack output to the motor commanding full assist to left lock.

In the meantime, since you have access to alldata, look over the circuit and check for loose or poor connectors, ground distribution, ect.

 

So I took it to our local honda dealer and he said that the torque sensor is bad according to their HDS. However, he told me on the '09 si's that the sensor isn't servicable, you have to replace the entire rack. Does that sound right??

 

I had figured it was a command via bad input.

Yes this is true, you must replace the rack as a unit. Honda will only sell it that way. The sensor does come off however. But we won't sell you just a sensor as it is not separate in the parts catalog.

This sensor seems to be a common issue in most of the eps problems on the Si's that I've seen thus far.

 

So here's my next question then, could we pick up a good condition used rack and just swap the sensor?

Also, he said the code that came up is "22-01: Engine speed signal", what does that have to do with the torque sensor??

 

Yes you could provided that sensor is from a known good unit.

Ok now that code leads me back to my original concern of the control unit malfunctioning. These are two seperate codes not dealing with one another. So that was the only code? If there is no code for the torque sensor, yet the datalist proves otherwise, and your test passes for 22-01, then replace the original suspected eps ecu. Here is the test;

Disconnect the 28p connector from the eps ecu. Start the engine.
measure voltage from pin 3 (BLU) of the connector clip and good body ground. Is there voltage? Yes- check for poor connections. If none are found, replace the eps control unit. If there is NO voltage, there is an open or short to ground between this wire @ the eps and the pcm.

If the technician working on the vehicle does believe the torque sensor is not reading right yet no dtc's regarding that are found, and there is no input to the eps ecu from the pcm for engine rpm, either the eps unit is bad, there is a short to ground or open between the eps ecu and pcm, or your pcm is bad. Or, if your running hondata, be sure that is ok too.

(if you have the vehicle, cover all basis in which was moddified or touched.)

 

Ok, so I think the EPS control unit is bad. With the 28pin connector plugged into the control unit, I get positive continuity between all 3 torque sensor wires at both connections at the firewall and at the sensor. But, if you disconnect the 28pin connector from the controller, the continuity goes away. Also, the volt ref wire is testing -0.56 volts and the 2 signal wires are testing 5 volts.

 

I got the pm. Give till tomorrow to edit this post. I need to look @ something; then results

 

From my travels with the EPS system in an 8thGen Si, My steering died from a fire [LONG STORY] Burnt the motor harness and the engine room harness. Had to replace both of those things. After all that hell, I still had my EPS light on. The EPS codes gave stuff about the Torque Angle sensor and stuff being way off. I replaced the whole EPS ECU and it fixed everything. Just my two cents and story. Make sure to test both the Rack and the EPS unit thoroughly. the Units are expensive.

 

Im not sure where exactly your measuring continuity at. You should have cont. at the disconnected connector @ the ecu and the sensor conector (each wire respectively) regardless of being plugged in or not. If you do have continuity, that means that harness and connector cavities /terminals are good. If you have no cont, you have an open or you have high resistence somewhere.

The refvolts should be around 5v however should fluctuate 2.5v per sensor depending on angle.

Again, the ecu is giving a wack output to the motor. Without being @ the car myself, from what weve discussed, Im sticking with the eps ecu being faulty and giving wack output. I dont think we have a bad input here causing the output.

Im not able to get the eps unit solidstate schematic @ this time so; Pull connector D (28p) from the eps ecu and measure resistence between 8 & 9. You should get 12-15 ohms. If you DO; meaasure cont. from p9 to body gnd. If you have NO cont, double and tripple check for poor or loose connections. If you find no poor connections, go ahead and replace the eps ecu and call this a day. *If you found more or less ohms between 8 & 9, let me know.

 

What I meant with the continuity was that with the 28pin plug connected to the module, all wires were testing pos continuity with each other. Went ahead and replaced the eps control module and its still doing the same thing this is driving me nuts.

 

Ok, I understand. And Im guessing you found more or less resistence as you went ahead and replaced the ecu.
Like I said, without being at the car, its somewhat of a crapshoot, however the sympotms youve described and the testing and work youve performed seemed like a bad control unit-via the forum chat here.

Before you buy a new rack and/or motor, tripple check connections everywhere in the entire eps circuit including ground(s). Again, if you are not comfortable (esspecialy now), then go with what the tech @ Honda was saying (if you trust what they say based on their diag-being right at the car. Which you shouldnt have reason to not trust them Id hope.). Id hate to see you waste money unintentioanly and the customers time-including your own time.

 

Ok, so the whole ballgame changed. All along, I've been noticing that wire colors weren't the same as what you've been describing and by alldata but I just assumed that the pin numbers were the same. The connector itself has pin number indicators on the corners of it that I've been going by which I've now figured out why I've been so confused. The connector indicator numbers are WRONG apparently because when I go by the reverse of that the connector says, it all starts making sense and the wire colors match up like they should. My stupidiy for assuming the connector was correct in its label.

So, I've gone back and re-performed all the tests you've mention above and here are my results, test was done with car running and 28pin D plug disconnected from EPS module:

Blu/Red ref voltage: 0 - No voltage
Pin 3 Blue wire voltage: 6.98 volts
Reistance between Pin 8 and 9 - 13.8 ohms
Continuity of P9 to body ground - No continuity

Do you have a diagram that shows where the grounds are for the EPS system specifically? All connections are solid and intact.

Hopefully with these new results we can figure out the issue.

 

Gnd distribution;

-torque sensor; coupe; at the bottom of the eps ecu. 4dr; top left of the eps ecu/ to right of the cabin filter tray.

-eps ecu- to the right of cabin filter tray (regardless of body)

Check the wires from the motor itself up to its connector onm the subframe, and up to the ecu and make sure nothing is shorting to body ground or another wire.

***from here on, diag is assuming all connectors, terminals, and wires are OK as you have stated***

Some of our readings here are not making sense. Like, you cant have this here, and not this here, and this work. Perhaps its due to the incorrect colors. The ref volt wire should have cont to gnd with D unplugged, this makes the trq sensor now prime suspect. Do this, unplug the torque sensor clip @ the sensor, and turn the engine on. Do the wheels jerk left? No-troubleshooting complete, replace the rack (or grab a known good sensor) Yes-unplug the motor angle sensor. Do the wheels jerk? No-replace. Yes- impossible, as there is now no input to the ecu to give command to the motor to assist full left.

(Advance torque sensor diag-with everything plugged in, back probe the torque sensor clip @ the sensor and measure voltage at either output wires, not the ref volt wire thats red/blue. If you find nothing, or one is 0 and the other Xvolts, then there is an isse with the sensor. However, since the wheels jerk left, the reading will read that value. Have someone hold the steering and take measurements)

(it seems we are going to have to start from scratch to clarify the now known mixup, and youve started already by visualy checking connectors and wiring but check again as you go along now.)

 

Hi..I was wondering if anyone can help me out.i have 06 si just replaced the rack and eps ecu...now the problem im having is that the steering completely locks to where I cant move it in no direction when I turn the car on.now when I turn the car off I can move my steering normal.i dont have no eps light on..can someone please help me out.thank you!

 

So I'm having a similar issue. The lower slide shaft came out instead of disconnecting at the joint on the shaft. Now I aligned everything back up and the wheels are straight and so is the steering wheel. Only issue is the eps is driving the wheels to the left.

Is there a way to do the neutral position recalibration on the torque sensor without the hds like the s2k platform? The just short the dlc to chassis ground.