Trouble with starting and running - loping and vvt1 err
 

Hi everyone, after running somewhat normal for a month or so, yesterday I was driving home and the miata starting running rough. This morning, cold start is very difficult and to get it to happen, I have to hold the throttle. While running, it sounds awful, like a loping v8 that's dying. I've changing coil packs and spark plugs and hoping someone else has some advice. I ran the coil pack tester in ms3 and can hear clicking on coila and coilb. I ran an inline spark tester on 4 and 2 and they worked. I started the car and held throttle (required) to 2000 rpm and pulled the wire for 4 and then 2 and for each the rpm dropped 500.

I've posted my log and tune.


EDIT: I have a mild 6.5psi turbo (efr 6258) on an otherwise stock internals engine that last compression test was 170 to 175 across the board.

bump for any help. Attached a log from today with me holding down the throttle to keep it running. Just for a running list of things I've done:

1. New spark plugs ngk4644
2. new coil packs
3. tested spark with inline tester
4. tested coil packs with ms3 tester
5. checked ecu ground
6. checked harness ground on manifold
7. checked timing bumps between gears at 19
8. checked timing marks lining up
9. checked 900cc injectors and resistance
10. swapped in 640cc injectors and updated tune

thanks again for any help

Here's a video of the start up and a couple seconds of running. It's quick but the car will just continue to run that way.

You've checked spark, but have you checked fuel? Make sure all injectors are firing. Painful to do on an NB, but you could use a stethoscope to listen for clicks while using the MS3 test mode or while someone tries to keep the engine alive. Careful using the test mode, as you can flood the motor if done incorrectly. To be safe, you could disconnect the fuel feed line while performing the test.

I’m with Ricardo, would check to make sure the injectors all have power and pulse next. Generally done with a test light, but Ricardo’s technique might be easier on your setup.

What test light did you use? Actually easier for me to do it with that b/c I'm doing the troubleshooting without a friend to keep the engine running while I try to listen.

Sorry for the double post. Will they actually click with the fuel relay disconnected?

I tried to use an Autozone fuel pressure test kit and the fittings don't exactly work fo the quick fittings on the nb2. That being said, I know there's at least enough pressure from the pump to blow that fuel tester's shitty connection off.

Fuel Pump Relay? Yes, they’ll still click with it disconnected. If you mean the MFI relay, then no as the ECU is off then.

and a noid light is the usual test method for injectors, but I’m not sure if you can squeeze one in there with the manifold on, hence why I mentioned a stethoscope to listen for clicks instead.

Fair point. I've seen videos of people using a really long screw driver as well. I have much faster access to that vs. a stethoscope.

Ha! Yeah, a long screw driver will work too!

Thanks again guys! I'll give that a whirl. Have you seen injector cabling go bad? I've actually swapped the injector's themselves out from flowforce 900s to flowforce 640s. Is there some other kind of fuse or something that controls them separately?

Oh weird question - does the fuel relay also get rid of spark? When I was testing spark with an inline tester the light would not light up with the fuel relay out but when I put it in, it lit up like a xmas tree.

Looks like it was something with VVT. I had noticed and asked my tuner about a vvt1 error on the gauges when starting up and he said it probably was nothing. Today because nothing else was working, I did the VVT cam test and set duty cycle to 0% and it started up normally. Still not fixed b/c now I have the wrong 640cc injectors in there, but clearly something is either wrong with some part of the VVT system. Hell, I don't really even feel like fixing it. I may just run the system in 0% duty cycle as normal.

That's very interesting, and I'm glad you found the cause. Looking at the logs, it does seem the VVT was doing things. My guess is that your Minimum and Maximum cam angle settings need work and maybe the minimum vvt duty too. I just found this write up: https://bartuneengineering.wordpress...l-setup-guide/. Give it a try and see how it goes.

That's exactly what I found after a ton of googling for the VVT1 err blink. I started down that process and the second i set the the duty cycle to 0% the car returned to normal. It exactly explains the loping sound b/c it was literally 'cammed' having the VVT angle at max. I added the datalog i was running during the eureka moment. You can see my relative cam angle output at like 40+ degrees and then it go to zero when i figured out the duty cycle. Then it goes to like 279 b/c i zeroed out the min/max cam angle stuff. Interesting as all heck. I had to stop messing about b/c now that I have the FF640s installed the idle AFR was like 18+ after WUE stopped. Weirdly no matter how much higher I raised VE i couldn't get that AFR to come down even after re-setting up deadtimes to the FF640s. Hopefully my tuner can look at it tonight and get me out making boost again!

Truly glad your got that figured out. Did you update your REQ_FUEL for the new injectors? Your idle VE should be between 50-65, otherwise some setting is off

Yep, I updated my req fuel to 640cc from the 900cc it was on before. My idle VE has always been in like the 70s to 80s even before. My tuner said at one point he thought I had a vacuum leak, but we could never find it. I did find that flow force's website says their naming convention is based on 3Bar fuel pressure and the miata is 4bar. So I guess technically it makes 640cc injectors 725cc and the original 900s 1057cc. My req time for the 900s was based on them actually being 900s which may have accounted for the VE stuff, but I dont know it well enough to actually commit to that.

https://goflowforce.com/instructions

Correct, you need to scale your injector sizing to the NB 4bar fuel system, so 640cc@3bar -> 739@4bar, and 900cc@3bar -> 1039cc@4bar. Furthermore, the fuel pressure also affects deadtimes so your tuner needs to verify that the deadtimes are correct for your differential fuel pressure (which varies in an NB returnless system....) but that's a different topic altogether. Use the deadtimes given by flowforce for 4bar systems and set the fuel system type to Fixed Pressure in the Fuel Settings > Fuel Pump and Pressure Control. Also set the pressure setting to 58psi/400kPa. MS3 will then do a fuel compensation in the background based on the MAP.

Setting the dead times and req fuel based on 725cc then reloading the tune helped massively. Actually idled rich rather than lead after the change and I had to get the VEs down to the low 70s which is the lowest they've been. My tuner seems to know what he's doing so the 900cc thing on the 900cc injectors is probably just a brain fart or the fact that he was doing everything remotely. That being said, thanks everyone for all the help. I ran a few live tunes and still need to redo some more of the VE table, but everything seems to be working as long as I disable VVT.

Actually that brings up a point. Now sometimes with VVT enabled (test mode off) the duty cycle will be pegged and the vvt angle will start to max out (issues return). Sometimes the same thing will happen for duty cycle (it'll max out) but the VVT wont move during idle and as a result, the car idles normally. I thought maybe it was oil pressure, but the moment I put it into test mode, everything seems to work as expected with about 50 to 60% duty cycle advancing the vvt to maximum degrees.

I had a few boost/vac leaks that were hard to find. Turns out there are at least two thicknesses of injector gaskets. I had both and the thinner ones were leaking. I was able to pressurize the system and listen for leaks with the kit I bought online here https://www.siliconeintakes.com/inta...essure-tester/

Thought I would update for future generations. I got a replacement VVT solenoid in and at least so far that seems to have resolved the issue. Upon removing the current VVT solenoid (for the 3rd time to inspect), I found small 3mm x 3mm piece of plastic in one of the gaps. I can't tell that anything on the solenoid is plastic so I'm not entirely sure how that got in there. The original solenoid may be just fine and that was getting in there locking the cam forward, I just don't know enough about how to works to say for sure.
https://cimg8.ibsrv.net/gimg/www.mia...8a68dab8d6.jpg
https://cimg9.ibsrv.net/gimg/www.mia...dd63a96f01.jpg