67 replies to this topic

[ECOBRAP]

Hello all, after playing detective all weekend I have not been able to diagnose and fix a problem that appeared after installing my new motor. Hoping the collective knowledge of the forums can help me out. As some background information, the car sat outside during a rainstorm (covered) when the motor was out, and that will explain some of the unusual things I tried. Also, with the old motor the car would never idle below 1k rpm due to a gunked up idle control valve, which was replaced when I put the new motor in. So this is the first time since I have owned the car that it has been idling at the correct rpm. Car is not throwing any diagnostic codes, and has not been dyno'd with the new motor yet so it is hard to tell how it is on power.

 

Symptoms: A very rough idle that shakes the entire car. Car runs fine once above 1500 rpm and under load, but the car sometimes stutters a bit when taking off from a standstill. You will see in the warm idle video that the car sounds like a big cammed V8 when idling, and as cool as that is, it doesn't seem healthy. The spark timing jumps around due to the rough idle, but I would say it is set at stock 10 degrees. Also, I haven't driven it much to know for sure, but it seems I may be getting poor MPG.

 

Cold Start Video: 

Warm Idle Video: 

 

So let's get to it, things that I HAVE checked or experimented with:

-New Spark Plugs

-Plug Wires

-Ignition Coil

-Throttle Body

-Water in Gas Tank

-Rain damage in the ECU

-Electrical Grounds

-Vacuum Leaks

-Spark Timing (10 degrees)

-Fuel Pressure (38-40psi)

 

 

Things I have NOT checked or experimented with:

-Air Flow Meter

-Air Fuel Ratio

-Timing Belt / Cam Overlap

-Fuel Pump

-Cam Angle Sensor

 

Any suggestions or diagnostic advice is highly appreciated   First race in 3 weeks!

[B(Kuch)Kucera45]

Did someone build it for you or is it a used motor ?

[ECOBRAP]

It was an old Rebello motor (3 seasons on it) and I sent it back to them for a complete freshen up. Motor seems to run great apart from this issue which I think is independent of the build.

[B(Kuch)Kucera45]

I couldn't see Rebello getting the timing off because he has been around for awhile and knows these motors. But checking the timing is easy,just take off the valve cover so you can see the cams. On a 1.6 you have to pay close attention to the timing marks and that they are straight up at 12 o clock. I can't remember witch one it is but one of them will look like it's off when it's in the right position.

I would start there.

Did you say the ECM got wet ?

If any electrical parts got wet you will have to dry them out,but if the ECM got wet your probably going to have to replace it.

[GT40Jim]

It really does sound like retarded ign timing? You cked and set timing I see. Have you checked the harmonic balancer for rotational slippage? I just toss a long Philips into the plug hole and rotate back and forth over the tdc mark.

[Steve Scheifler]

I think it's trying to shake off those red hoses, like a dog rejecting a cute sweater.

But if that doesn't fix it, check cam timing. I've never actually run one of these with a cam a tooth out, but bot sure it would be rough if it runs at all. You need to pull the upper front cover as well as the cam cover to do it correctly.

Disconnect the ECU and plug it back in, working the connectors on and off several times to clean the contacts. Likewise the big connector to the AFM and the one to the CAS. I've seen eack of those cause serious running issues if not fully connected or corroded.

Borrow a few key parts and swap them out. You really souls"d have a known good spare ECU, CAS, igniter, coil pack, AFM, at a minimum anyway, but right now you need known good parts. If there is nobody around you, I can send a set for a quick test.

[Mark]

I'd start with verifying the cam timing. It only takes a few minutes and then you can rule out the cam timing and focus on the other things.

 

Set the engine at TDC with the timing marks (grooves) on the edge of both the cam gears both pointing straight up. Using a piece of string with a weight on the end (a 6mm nut works well), plumb bob the timing marks. With the string held against the cam gear and lined up with the timing mark the string should pass through the middle of the cam gear bolt head.'

 

Mark

[ECOBRAP]

LOL the red hoses look a lot better in person I promise 

 

I will check crank pulley timing mark and cam timing this week, along with double checking the AFM and ECU connections. I cannot seem to figure out the AFM plug, and have already snapped one latch on the side of it trying to get it off... it won't seem to budge.

[ECOBRAP]

Also, something interesting I observed during the cold starts. Watch the video, and you will see I start the car and it starts idling at 900rpm COLD, then 10 seconds later it realizes "oh wait I'm a cold motor, better increase rpm during warmup", and THEN it goes to a normal 1200-1300rpm cold idle speed. It realizes this 23 seconds into the video and you can hear the engine pick up.

 

The thermosensor on the back of the head is attached, and temp registers on the stock dash, so does this mean it is definitely sending information to the ECU? Or can it send information to the dash temp gauge but still not send information to the ECU if the sensor is broken?

[Bench Racer]

The thermo switch in the thermostat housing sends the signal to the OEM temperature gauge.

 

The thermo sensor at the rear of the head sends the signal to the ECU.

 

The engine runs open loop at wide open throttle and also when engine is cold. I don't know the engine speeds. Some signals are not sent to the ECU during open loop.

 

What is idle speed when engine warm?

 

Have you tweeked the idle adjust screw under the rubber cap at the top left hand side of the throttle body.

 

Is the throttle body from your previous engine? 

 

For the latch spring clip on the AFM and other places use what looks like an ice pick with a 90* bend (make one or they come with little pick and poke kits, cheap at local hardware store) at the end to disconnect one end (or both ends, don't remember) of the spring clips. 

[Rob Burgoon]

Does the idle stabilize if you jumper TEN to GND?  What rpm and what does the timing read when you do that?

 

Or do 1.6s not have that?

[davew]

I guess it is this time of year again:

 

Start with the basics, befor jumping to any conclusions. Do a compression and leak down test. Which cylinder(s) are missing? Do not pull plug wires off, spray water on the header and see which one is cooler than the rest.

 

Check vacum with a gauge, what is it and is it steady?

 

I would lean towards a vac leak or a bad injector

 

Davey Dew, you are wrong. The sensor on the thermostat housing is for the cooling fan. The sensor under the intake manifold on the rear of the cylinder head is for the factory gauge. The green square "Bosch" type sensor on the rear of the head is the ECU sensor.

 

Matt, check the wires on the green sensor. many times I have seen the wires pull out of the green plastic housing. This gives the ECU a ice cold reading and it goes pig rich. You have to remove the connector and look inside to see if the wires have pulled out. They look fine from the outside, but are notorious on the inside.

 

Dave

[ECOBRAP]

Does the idle stabilize if you jumper TEN to GND?  What rpm and what does the timing read when you do that?

 

Or do 1.6s not have that?

 

When I do the jump the idle rises to 1000-1200rpm, timing goes up to 12 degrees or so.

 

 

I guess it is this time of year again:

 

Start with the basics, befor jumping to any conclusions. Do a compression and leak down test. Which cylinder(s) are missing? Do not pull plug wires off, spray water on the header and see which one is cooler than the rest.

 

Check vacum with a gauge, what is it and is it steady?

 

I would lean towards a vac leak or a bad injector

 

Davey Dew, you are wrong. The sensor on the thermostat housing is for the cooling fan. The sensor under the intake manifold on the rear of the cylinder head is for the factory gauge. The green square "Bosch" type sensor on the rear of the head is the ECU sensor.

 

Matt, check the wires on the green sensor. many times I have seen the wires pull out of the green plastic housing. This gives the ECU a ice cold reading and it goes pig rich. You have to remove the connector and look inside to see if the wires have pulled out. They look fine from the outside, but are notorious on the inside.

 

Dave

 

Thanks Dave, sorry I am a bit of a noob, we don't have autoshop class anymore so I have to learn on my own!

 

Because of the cold idle irregularity, where it delays in realizing the motor is cold and takes a while to jump up to the correct warmup idle, I believe the issue is electrical/sensor. First thing I will check when I am back is the square sensor that sends temp signals to the ECU.

 

 

The thermo switch in the thermostat housing sends the signal to the OEM temperature gauge.

 

The thermo sensor at the rear of the head sends the signal to the ECU.

 

The engine runs open loop at wide open throttle and also when engine is cold. I don't know the engine speeds. Some signals are not sent to the ECU during open loop.

 

What is idle speed when engine warm?

 

Have you tweeked the idle adjust screw under the rubber cap at the top left hand side of the throttle body.

 

Is the throttle body from your previous engine? 

 

For the latch spring clip on the AFM and other places use what looks like an ice pick with a 90* bend (make one or they come with little pick and poke kits, cheap at local hardware store) at the end to disconnect one end (or both ends, don't remember) of the spring clips. 

 

I have messed with the idle adjustment screw and it works, raising or lowering the rpm accordingly, but it is still choppy whatever I set the idle to. The throttle body is from the old motor, but I cleaned it out. I replaced the idle control unit that sits under the throttle body, so I may have gotten a crap one? I don't think it would cause such a rough idle, only an unusual idle rpm.

 

As for the AFM, yes the plug looks self explanatory, I tried using a flathead to try and push down the tabs to clear the springs clips, but the tabs do not budge whatsoever which is why I wanted to double check I was trying it correctly.

[B(Kuch)Kucera45]

Dave made a good point check for a bad fuel injector.i totally forgot that I had one that was bad on one of my cars and it sounded exactly like your car. The way I found it was by disconnecting one injector wire at a time to see if it changed at all. Do one at a time and if it doesn't change plug it back in and go to the next one.

I think that's your problem.

Oh and be careful with the clips on the injectors,you don't want to break them or you will have to replace them.

[ECOBRAP]

Thermosensor plug is #1 suspect, when I pulled it off I felt play in the wires, they may have come out of the plug altogether. I have a new plug harness arriving in a couple hours, I will provide an update.

[Bench Racer]

Matt, sorry about the miss info, rear of engine 12:00 clock switch sends signal to engine, rear of engine 3:00 clock switch sends signal to dash gauge and thermostat switch sends signal to cooling fan.

[Bench Racer]

I hope you have success with the thermo sensor plug/harness.

 

If thermo sensor plug does not solve the issue, take a jumper wire and jump terminal GND and TEN in the diagnostic box. This will place the ECU in open loop and the ECU will ignore signals from ignition rpm, temperature, throttle position sensor  and sends a fixed signal to the idle speed control valve. If the idle speed issue changes, the issue would seem to be narrowed to the one or more of the items mentioned above.

 

See, Miata.net/garage/isc.html  for info for the idle speed, air valve and the idle speed control valve and the info typed above.

 

I had an issue a year ago where when the accelerator was held constantly at 2,200 rpm the rpm would pulse back and forth between 2,200 rpm and 1,700 rpm. Per the FSM, I checked and rechecked many items more than once. Found the issue, but went a bit further on the whack-o side.

[ECOBRAP]

Replaced both the thermosensor and plug harness, no change.

 

Cleaned ECU plug contacts and AFM plug contacts, no change.

 

Hit AFM with sledgehammer, no change 

 

When I do the ten to ground jump the motor increases idle to around 1100rpm, but still has the stutters.

 

Almost out of ideas, so on Wednesday I will be checking compression, timing belt, fuel injectors, idle control valve replacement, and seeing if the crank pulley timing mark has slipped. I will also spray the exhaust manifold to see if one cylinder is not firing. I may also take the car to a shop and have them attach a fog machine to the intake manifold, that's how they check for sneaky vacuum leaks. Open to any other ideas as well.

[Diller]

For vacuum leaks I have used carb cleaner in very conservative sprays around the intake. If the RPM jumps, I know I have found a leak. Another option is to enjoy a nice cigar and keep blowing the smoke around the engine. If cigars aren't your thing, I came across some "smoke in a bottle" solution from an company that makes aluminum wheels. They use it in the leak testers. I believe this is the same thing from McMaster racing supply...http://www.mcmaster....3880k31/=w1s6tk

 

I used it a lot when I built a turbo system for a Mustang to track down vacuum leaks and it worked great. 

[Steve Scheifler]

If that's a vacuum leak, it's a huge one. Do the plugs look rich or lean? If it's one cylinder, injector, plug, sticking valve, whatever, then the header tubes check is a good start. Every racer willing to open his own hood might benefit from a cheap IR temp gun. That will catch the difference without having to run the car long enough to get it hot (and maybe compound your problems with washed down cylinders etc.).

Nothing wrong with checking the big/easy things first even if they aren't most likely, but when that fails it's time to be methodical as Dave suggests.