JT's M20 Build Thread

[JehTehsus]

have you used LM programmer to verify the LC-2's analog 1/2 output to make sure it's actually outputting what's expected?
I set up an LC-2 a year back and it was fine, but my LC-1 came preset to the wrong multiplier for gasoline.

I have not - but I also don't think that is the issue, because sometimes I can get it to hold stableish AFRs for a a bit.

So screwing around last night at (high) idle of 1300RPMish, I shut off auto tune because according to the AFR gauge it was massively leaning things out (and the VE table looked really low). Took about 2 minutes to get the VE table to reasonable numbers and the idle much more stable just by manually tweaking.

I am starting to suspect a vacuum leak and/or an intermittent injector, and plan on doing some tests there this afternoon. My reasoning here is as follows;

• The fuel pressure regulator often doesn't seem to 'respond' to changes in MAP below ~80 kPa. This results in inconsistent fueling, at least when driving. I have seen pressures (absolute) from ~270-330kPa, they should be bang on at 300 and steady. This is being logged by the bosch sensor on the fuel filter, which is not super close to the rail, but even so it seems odd to often not see the FPR do anything at all (when MAP is 80-100kPa it definitely does - you can see the FP change basically immediately to follow change in MAP). The regulator also seems to be leaving me a bit high on the FP side, in the below log snapshot it is holding about 312kPa, should be 300.
• My plugs seem to be all over the place. I have attached a couple pictures of them - cylinder six is on the left, one on the right, six seems a bit fouled.
• I may be reading too much into this... but the AFR gauge, under my relatively steady idle conditions (@1300rpm) seems to be spiking every 1.8s, which to me indicates a single cylinder is under fueling (based on plugs, probably *not* six... which looks to be overfueling :/). I don't know if this is an issue with the IDs or not - my understanding is they are relatively good at low pulse widths, and I believe I have the low pulse width control setup properly in the MS, based on the data they have online.

If anyone has any other ideas, I am all ears. Hopefully a vacuum leak check doesn't take too long...

I may also route my FPR pressure reference to the manifold directly instead of the vacuum station. Right now the vac station + FPR reference line totals maybe 2.5-3ft of (4-6AN) hose. I could probably route it right into manifold and shorten hose to a foot or less. I would expect this to show up as a delayed response though - not just plain 'ignoring' low MAP levels.

These are the plugs (6 - 5 - 4 - 3 - 2 - 1 left to right)





Here is a log snapshot. This is after quickly dialing in VE around where she was idling to keep things in the AFR gauge range. I think the gauge is working fine and telling me something else is wrong lol. Note that the pulses are almost always ~1.8s apart, with some occasional other odd behavior.

[JehTehsus]

Here is a video of that last minute or so idling as well, maybe someone can hear something I can't:

https://youtu.be/_h2mGPmqaOw

[varg]

Definitely sounds like there's an occasional miss, that'll throw the AFR readings off. Fuel pressure looks pretty stable in that log snapshot, 12kpa high really isn't that much, that's under 2psi and probably within the tolerances of the fuel pressure regulator. Do you get 300kpa with your engine off and the fuel pump running? Pull a variable vacuum on it with a handheld vacuum bleeder pump or something you can check the steady state response of it with no fluctuations due to injector operation. Assuming the inputs are correct and the pulse widths are accurate, your pulse widths are fairly small, I'm seeing 1.429ms in that log capture, but you mentioned even shorter pulse widths before. A quick google search didn't yield the minimum pulse widths for your 1050X injectors but 1.3-1.4ms seems to be close if not in the region of instability on the stuff I did see on the injector dynamics website. IMO you're really pushing it at idle with such large injectors on this little inline six. Have you tried just making it idle really rich to see if it runs nicely at a consistent indicated AFR? You might just have to idle rich to get the pulse widths up in an area of consistent response. I've tuned cars where the owners overdid the fuel injectors and it just had to idle in the 12s to keep the pulse widths where the injectors were happy. Try bumping your VEs up while idling until you see 1.5ms or so and see what it does. You have like 700bhp worth of injector on E85, you planning on using all of that?

[JehTehsus]

Definitely sounds like there's an occasional miss, that'll throw the AFR readings off. Fuel pressure looks pretty stable in that log snapshot, 12kpa high really isn't that much, that's under 2psi and probably within the tolerances of the fuel pressure regulator. Do you get 300kpa with your engine off and the fuel pump running? Pull a variable vacuum on it with a handheld vacuum bleeder pump or something you can check the steady state response of it with no fluctuations due to injector operation. Assuming the inputs are correct and the pulse widths are accurate, your pulse widths are fairly small, I'm seeing 1.429ms in that log capture, but you mentioned even shorter pulse widths before. A quick google search didn't yield the minimum pulse widths for your 1050X injectors but 1.3-1.4ms seems to be close if not in the region of instability on the stuff I did see on the injector dynamics website. IMO you're really pushing it at idle with such large injectors on this little inline six. Have you tried just making it idle really rich to see if it runs nicely at a consistent indicated AFR? You might just have to idle rich to get the pulse widths up in an area of consistent response. I've tuned cars where the owners overdid the fuel injectors and it just had to idle in the 12s to keep the pulse widths where the injectors were happy. Try bumping your VEs up while idling until you see 1.5ms or so and see what it does.

Thanks a ton for taking a look.

The ID1050s are specified with minimum pulse width data, which I have entered for the injectors (I believe correctly. I used that same spreadsheet you linked for my info, including setting up the small pulse width curve). All that said - yes, these are bigass injectors for a small motor, I fully anticipate having to idle rich, and my cam probably doesn't help either. My concern is that it does seem to be missing, which fucks AFRs, which I am pretty sure also fucks the autotune and makes me pull all my hair out. The reason for the large injectors is that I am building this to run up to 30 psi on E85, but it does introduce some challenges.

My FPR is adjustable, I may look at tweaking it a bit. It seems to be responding in 'steps' though, not sure if that is normal or I am just overdoing the adjustment a bit.

Vacuum pump is a good test, I have a vacuum bleeder for the brakes I will hook up to the compressor and try.



This snapshot shows my concern with the fuel pressure - yellow is the calculated FP (i.e. what injectors are seeing), green is the bosch PTS pressure on the fuel filter, a fair ways upstream from the rail. White is the MAP. This is over a drive I did a couple of days ago. I am concerned by how much the yellow line is jumping - some of the worst min/max I am seeing is 240-330kPa, which seems extreme. Often if looks like the FPR isn't even responding until the MAP gets above 80kPa, which doesn't seem right to me (but I am not an expert), which is indicated by the oscillation in the yellow line while the green (fuel pressure relative to atmosphere) just stays constant.

[JehTehsus]

Well well well... here are my extended lunchtime test results:

First off, vacuum: I tried a few things after I got the car to a semi-stablish idle, such as dumping water on vac fittings, a propane torch, etc. No noticeable change at all. Not totally ruling out a vacuum issue, but so far it doesn't seem too likely.

Next up, injectors: Went through the harness, everything seems to be connected. Next step was some logging and tweaking. I ended up enabling individual pulse width control for the injectors and going through and bumping them one by one by 20%, after once again reaching a stableish afr at idle.



There definitely seems to be something up with injector 6. 1-5 all immediately and noticeably dropped the AFR considerably when adding 20% more fuel, as expected. 6 just .... didn't. I think it moves a bit, 'on average', but definitely not the same behavior.

This makes me wonder if it is partially plugged or something. Only thing I can't get my head around is that the plug looked pretty rich for cylinder 6 - if adding extra fuel has no real noticeable impact then why is it running rich? Maybe it is just 'stuck' partially open?

I am leaning towards pulling and checking. I may also throw the scope on it if it looks fine, could be an electrical issue where a poor connection is limiting power available.

Here is the logfile:
https://drive.google.com/file/d/18tA...ew?usp=sharing

Also, some really exciting youtube video action, in two parts just for additional enjoyment:
https://youtu.be/JWHQ8qW3hv4 (terribly blurry in this one)

https://youtu.be/pofXutxlVKQ

And here's one more video of it idling while I screw around:

https://www.youtube.com/watch?v=c9EO0-NuiwM

[varg]

I'd pull the rail out and fire the injectors in test mode to see what's up with number six.

This snapshot shows my concern with the fuel pressure - yellow is the calculated FP (i.e. what injectors are seeing), green is the bosch PTS pressure on the fuel filter, a fair ways upstream from the rail. White is the MAP. This is over a drive I did a couple of days ago. I am concerned by how much the yellow line is jumping - some of the worst min/max I am seeing is 240-330kPa, which seems extreme. Often if looks like the FPR isn't even responding until the MAP gets above 80kPa, which doesn't seem right to me (but I am not an expert), which is indicated by the oscillation in the yellow line while the green (fuel pressure relative to atmosphere) just stays constant.

Well the calculated fuel pressure will vary wildly if the MAP does, since it is a MAP based calculation. That's not unusual. That MAP oscillation I'm assuming was some bucking, common in an untuned car. The regulated fuel pressure won't be able to vary as quickly as the MAP does, it's a fairly well damped system, and there may be some sample rate or averaging limitation for the fuel pressure input as well.

[JehTehsus]

I'd pull the rail out and fire the injectors in test mode to see what's up with number six.



Well the calculated fuel pressure will vary wildly if the MAP does, since it is a MAP based calculation. That's not unusual. That MAP oscillation I'm assuming was some bucking, common in an untuned car. The regulated fuel pressure won't be able to vary as quickly as the MAP does, it's a fairly well damped system, and there may be some sample rate or averaging limitation for the fuel pressure input as well.

Going to pull the rail tonight and do some testing. It definitely seems like something is going on with six, hopefully just a wiring issue.

I may be over analyzing fuel pressure at this point. The Bosch sensor is a fair ways away (a good 2-3 ft of 8AN line, plus about a foot of dual 6AN where it passes through the e85 sensor, and two Y fittings). I did also see the pressure gauge on the regulator moving in response to small throttle blips during my idle test. Maybe I will think about moving the fuel filter (where pressure sensor is mounted) much closer to the rail at some point, but for now I am going to assume it is a bit damped given its location.

[JehTehsus]

Made it out for another short cruise on the weekend. Had the lady drive while I worked the laptop. The idea was I could keep an eye on things, maybe run a bit of autotune, etc.

I have attached the logfile and where the tune ended up, but long story short I am becoming even more convinced there is an intermittent wiring issue with either my wideband or one of the injectors, or (very probably this) my fuel map is just wayyyy too far out and given the non-linearity of the injectors and my inexperience I am having a hell of a time getting something reasonable out of it. That said, my ~1200 RPM idle does seem somewhat stable once she warms up.

I tested the injectors individually with the rail pulled. They all *look* to be operating consistently - although obviously my eyes are a poor substitute for a flow bench, number 6 did not look or sound any different to the others with a few different (1.2-3ms pulse widths were what I tested) pulses. Not ruling it out but it seems ok. Obviously this doesn't mean there couldn't be a wiring issue with it - but the power wire is shared with 5 with a couple of joins and the other wire is totally new/no joins all the way back to the MS.

The AFRs I am reading are driving me nuts though. My point of reference here is the M20 3L stroker my buddy has that we recently started up - the AFRs were (relatively) stable, and behaved as expected for an untuned car during a first drive. Obviously there are some appreciable differences between our motors - but not enough (IMO) that we should be reading AFRs in an 11-16 range on his motor, generally in the 12-14 range, whereas on mine the engine repeatedly and consistently will fly all the way out to < 7.5 or 22+, i.e. the absolute limits of the gauge, even if it only stays there for anywhere from 1-2 samples to a second (which is obviously very concerning). If I look at the wideband I am seeing lots of dangerously lean spikes and lots of stupidly rich spikes.

The other interesting thing is the extreme instability in my readings, even at relatively stable conditions. Sometimes AFR outputs will look absolutely normal, and respond to me adding/decreasing fuel as expected, but then out of the blue will start jumping around again. Which is what makes me suspect wiring, since I am running out of ideas at this point. Maybe there is nothing wrong mechanically/electrically at all at this point and I just need to admit my ignorance is biting me in the ass.

I guess this leaves me with two options - use the data from the last run to try to construct a better VE map that will avoid the dangerous spikes, (i.e. get it in a reasonable range), or make the next drive one to a dyno, which does feel a bit like giving up and I hate the idea of, since I want to know how this thing works inside and out. I am definitely not comfortable 'cruising around with autotune on' right now - it may have helped a little bit during the drive where I was watching it, but often times it would seem to get off into the weeds and I would end up running dangerously lean or rich. It is also a real PITA to 'street tune' when I can't even seem to get the numbers into vaguely the right ballpark. Admittedly this may just be due to non-linearity in the injectors, which are spending a ton of their time down in the weeds.

I had to dramatically change the VE map from the template I was using (which seemed to work fairly well on my buddies car).


Note that there were a couple spots in the log where you can see fueling go absolutely crazy, this is my fault as I accidently entered 255 in a cell. Any recommendations on a good way to adjust cells while being driven around? Is there one? Should I migrate over to the MS forums? lol
The big drop in fuel temp is me filling up the car.

Tune and logfile:

https://drive.google.com/file/d/1b1DNd0Txfff2miABCcbI88VVd8ai2nHz/view?usp=sharing

https://drive.google.com/file/d/1sqaayS4pogVNrX9Z-XFlQdC5q-4feDP_/view?usp=sharing

Anyway, some other minor details:

• Had to swap out my lower rad hose, when it came off earlier during my second (attempt) at a drive, turns out the alternator fan rubbed a nice big juicy hole in it. Old hose is in place now, and cable tied just in case.
• I know everyone seems to have a hell of a time bleeding the M20... but I have done it multiple times now, unfortunately, (without a bleeder screw, mine broke a few days ago when I went to use it naturally, about 95% rust) and don't seem to have any issues. Am I doing it wrong? Just been raising the reservoir and filling to the top then blooping the hoses till all the air seems to be out. After a short drive topping it up and it seems good. Note that this is the early cooling setup.
• My fuel pump mount/bracket leaks like a sieve when I fill the tank. That was a fun discovery to make. Suspect I need a new o-ring or I haven't tightened it properly, or both.
• Low beams seem dead or burned out, haven't check the relays yet.
• (lsd)Diff seems to work, although I really should change the fluid. First drive I accidently lit up the tires (in the wet) and she was pleasantly stable.
• My brake fluid reservoir was about 1/2" too high... I obviously suck at measuring. Twice. Need to pull the plate and move the holes lower, for now cable ties (safety first!).
• Brakes do feel OK. I think *maybe* a bit more travel than I would like (on a street car - 100% too much for hard driving) - my partner complained they feel kind of weak (which IMO they do in most all stock M20s with old pads) and was a bit concerned, but I had her do a few harder pushes and the car absolutely stops without issue, she said they were 'OK but a bit weird' after that, which I mostly attribute to them not being at all grabby. That said, I am not yet sure if I can lock the wheels or not, so maybe a slightly bigger cylinder would have been a good choice, or maybe I just need to dump those old pads and rotors that are not in great shape and have been sitting for almost 2 years. The autocross pads on my merc, which I am more used to, are extremely grabby, whereas the E30 is like the exact opposite. Finally, the seats in the car are absolutely *fucked* and make the entire driving experience terrible, so while I wait for new seats it may all be a moot point.
• Seems to be a bit of noise on the oil temp and fuel temp signals... nothing major, probably not enough to worry about. No noise in the pressure signals from the same physical sensors, so maybe the NTC element is just more sensitive/low current.

[moatilliatta]

Late style cooling setup helps keeping that radiator hose getting destroyed. Plus E36 radiators are easy to come by.

Something I learned with time with fresh standalone and harness and options, Install on a crap stock motor and debug the system. It can be very hectic trying to start and break in a fresh motor with a bunch of debugging.

Staying tuned in to see how it goes!

[JehTehsus]

Late style cooling setup helps keeping that radiator hose getting destroyed. Plus E36 radiators are easy to come by.

Something I learned with time with fresh standalone and harness and options, Install on a crap stock motor and debug the system. It can be very hectic trying to start and break in a fresh motor with a bunch of debugging.

Staying tuned in to see how it goes!

Totally. Given how many modifications have been made though I really don't see how I had much choice haha.

It is what it is. I am fine having to teardown and replace rings in the future - but I certainly don't want to melt pistons. I knew it was going to be a real PITA so I have purposely avoided the blower so far; once I have something debugged and running reasonably well it will be time to push for excessive power.

[digger]

i'd be tempted to throw in a stock or smaller set of injectors to get the shape of the map ironed out then use injector scaling to get the big boys in

[JehTehsus]

i'd be tempted to throw in a stock or smaller set of injectors to get the shape of the map ironed out then use injector scaling to get the big boys in

I am going to mess around a bunch this weekend and try to validate dead times... I did manage to get a stable idle (at like 1200rpm lol) without much effort. I also plan on making some significant VE table changes (some more lower rpm bins) and doing a few quick pulls around the block, taking time to review and change between runs. I can definitely move the afrs around, but it seems like the sweet spot is really small.

[JehTehsus]

So, good news: AFR issue is/was either a wideband calibration problem or a wiring issue. I recalibrated the LC2 yet again (and was also screwing around with a bunch of wiring) and it is now working exactly as expected... quite stable, obviously needs tuning, but no extreme crazy min/max jumps. Autotune actually seems to be able to work with it.

The sad news is I can't actually drive the car at the moment, ripped the (badly damaged) old seats out and am getting ready to get some old recaros into it instead.

In other fun news, I wired in the wheel speed sensors on the front two wheels to the spare VR inputs I have. These seem to be working well, up to an apparent speed of ~19km/h (which is as fast as I can readily spin the wheel with the car hanging on the lift). Since I don't have anymore VR inputs, I ended up wiring a spare digital frequency input to the speedometer out signal from the ECU. This hasn't been tested yet.

Here is my really exciting passenger wheel test: https://youtu.be/Yyxst7UHBBQ

My hope is that the speedo output reflects the actual ~ average rear wheel speed, which with the LSD is hopefully not too bad for traction control. Later on I would like to just grab the other two rear wheel sensor outputs and get them into the MS3 directly (either via swapping those sensors to hall or VR->digital pulse train conversion). Worst case I now at least have vehicle speed, so I know how fast I was going right up until I hit a tree.

I also wired the clutch switch, I have to actually test it still but I did ring it out and it was working, so unless I fucked up the soldering it should be fine. This is actually the wire that ran to the cruise control ECU - so it is really a brake/clutch combination switch thing, not 100% sure it will work for my traction control needs, but I can easily rewire the other end to just the clutch switch if I need to. I do need to properly adjust said switch, it has obviously shifted a bit or something and the un-pressed clutch does not quite activate it at the moment.

Lots of other minor things that have been bugging me got addressed - airbox is replaced, cleaned up some bits with cable ties, tightened my squealing belts, smoke tested intake (fucking idle valve boot was leaking, brand new boot. Glued it, fuck it. My AN fittings all appear to be leak free on the ~1psi smoke test, which is good if I expect them to hold up to 30psi) and exhaust (big exhaust leak where muffler joins midpipe I haven't bothered getting to yet, a stromoung replacement is on the way). Also idled the engine for about 10 minutes without the VC vent hose in place, thankfully mess wasn't too bad.

Finally, there appears to be some oil weeping around the front of the block, and I suspect the MLS. It is very little oil - at least at the moment. It could also possibly be the rocker shaft rubber caps, but I don't think so. If it is just a bit of weeping I will likely leave it till I yank the motor again. Zero oil/coolant mixing or mysterious coolant loss so far as I can tell, and I will do another oil change this week just to be sure.

Here are a few photos:




Goodbye, old friends (not really, these were FUBAR)







'Test fit', no rails yet




I lied earlier, I actually hit 20.9km/h spinning a wheel by hand, not 19

[Nader393]

Fun thread. Glad to see you sorting out the teething issues. I also just built up a 2.9, but milder tuning.



Here's your photo of your flywheel (minus the missing pilot bearing). I'm curious about the ARP flywheel bolts you used. Did you get at least 12mm length thread engagement? I bought those same 28mm ARP bolts as well, and found that they only engaged about 6mm or so of thread length into the crankshaft. (I'm using an M52B28 crankshaft for the longer stroke). So I went with longer bolts that I trimmed to get at least 12mm thread engagement.

[JehTehsus]

Fun thread. Glad to see you sorting out the teething issues. I also just built up a 2.9, but milder tuning.



Here's your photo of your flywheel (minus the missing pilot bearing). I'm curious about the ARP flywheel bolts you used. Did you get at least 12mm length thread engagement? I bought those same 28mm ARP bolts as well, and found that they only engaged about 6mm or so of thread length into the crankshaft. (I'm using an M52B28 crankshaft for the longer stroke). So I went with longer bolts that I trimmed to get at least 12mm thread engagement.

Are you using a dual mass flywheel? I did not measure the actual thread engagement but I would have to say it definitely seemed to be more than 6mm. I threaded them in a comfortable distance by hand before torquing.

**edit** NVM, just found your thread. I think I had more than 6mm (pretty much certain ...) but again, I didn't measure.