Build Thread - MegaShark

[ianwood]

Our team flirted with idea before, but the LeMons budget rules were too tight to make it work. Well, I found an incomplete MS2 3.0 board for close to nothing on CL. The soldering looks downright awful but they didn't get very far and I am pretty sure I can clean it up, replacing damaged/missing components as needed. Ordered a JimStim for full testing! I will be very happy to say goodbye to the finicky BEGI, troublesome AFM and super expensive 111 fuel! As a contingency, I am going to run a totally separate harness and leave the original in place until we get the MS fully dialed in.

I've been studying for the past week or so but will likely have LOTS of questions for the experts! This weekend, we will go hunting in the junkyard for a 60-2, VR sensor and a TPS. Ordered the GM IAT and CLT as well.

The setup will be M2 3.0 running MS2Extra speed density with 42lb injectors, 50 trim Garrett T3, stock coil and distributor, NGK BPR7ES spark plugs, initially targeting 10psi boost for about 250hp. No IAC. Tune to 100 octane under racing conditions. Will add boost control and fan control once it's up and running. Engine is M20B27 regular eta bone stock. Head was cleaned up and new gasket put in about 3 races ago. BTW, our race is June 30th at Buttonwillow. It's a true 24 hour race this time (our first) which means night racing!!! Totally unlit track!!! Yikes!

Some initial questions:
1. I've read on e30tech.com that the E36 TPS should work on the M20 throttle body. Can anyone confirm this? Apparently the E34 M30 auto TPS is kind of hard to find.
2. I've also read the VR sensor can be a little tricky to dial in and may need a 10kOhm resistor to clean up the signal. I don't have a scope and was wondering if there is an easy way to make sure there is a distinct rising edge for timing.
3. Grounding sensors, centrally or locally? I think centrally near the MS unit.
4. Injectors. I am thinking 42lb EV-1 "green tops" but should I consider 34lb instead for better resolution?
5. Process: From all the reading I've done, it seems the process is: build MS unit with needed circuits, verify the MS unit works with JimStim, install harness and sensors, put in base settings in TunerStudio, calibrate and test sensors in TunerStudio, tune starting parameters, tune idle for min kPa, tune VE and timing. Enjoy! Did I miss anything?
6. I am pretty clear on everything up to tuning VE and timing which is where I still have some learning to do. Are there default maps for this? Or do I borrow someone else's with a similar set up to get started? What's the best method? Does tuning with the WBO2 make this easy or is there still a lot of fiddling, guesswork? Pointers to threads on tuning maps appreciated. I still have to read this: http://www.e30tech.com/forum/showthread.php?t=33028.
7. Should I change the FPR to a 3.0bar unit or keep the 2.5bar stock? How does that influence my injector settings?

Wiring harness (link to larger version):




Some gratuitous pictures to get us started:

[mooseheadm5]

You just need a little adapter plate to make the E36 TPS fit. I have the drawing if you want, or can make the plate for you if you need, but it is really, really simple.

Fuel pressure:
http://www.rceng.com/technical.aspx

Basically:

New pressure * original flow rate / old pressure = new flow rate.

Ground all sensors at the same point very close to the MS. Use shielded cable for any AC signal wires, like VR sensors.

[ianwood]

You just need a little adapter plate to make the E36 TPS fit. I have the drawing if you want, or can make the plate for you if you need, but it is really, really simple.
.

Please send me the drawing and I'll see if I have the skills. How much do you want to make one? Email is lemons at ianwood dot com

[marc79euro645]

Your really making us all so proud.
Commit to the squirt wiring you're never going back.
consider a mid 90s nissan pathfinder tps, it has the same d shaft ,just strip you're old tps for the back plate,install new tps onto that with two self drilling sheet metal screws. I'm using 55# low z injectors with stock fpr, squirt will tune so no need for high fuel pressure,larger injectors will support more power later.
good luck
marc

[ianwood]

The timing may be right to make the jump to MS. I just built a home made smoke machine to test for vacuum leaks. Turns out the BEGI is leaking like a sieve in two places! Confirmed it with the car running and a vacuum gauge too. Always wondered why our idle vacuum was so terrible. Now I know! And now I have a headache too from burning rags soaked in WD40!

Highly recommended BTW. $35 in parts from Home Depot and as result I found the BEGI leak and 3 other small ones. Much easier than spraying carb cleaner all over.

[nerd of nerds]

The timing may be right to make the jump to MS. I just built a home made smoke machine to test for vacuum leaks. Turns out the BEGI is leaking like a sieve in two places! Confirmed it with the car running and a vacuum gauge too. Always wondered why our idle vacuum was so terrible. Now I know! And now I have a headache too from burning rags soaked in WD40!

Highly recommended BTW. $35 in parts from Home Depot and as result I found the BEGI leak and 3 other small ones. Much easier than spraying carb cleaner all over.

Got any kind of a link to instructions? This is relevant to my interests.

[ianwood]

Got any kind of a link to instructions? This is relevant to my interests.

Best bet is Google it. I pretty much built the first or second thing I came across. Note, you'll need a compressor or electric tire pump to get it work properly. I bought a paint can and lid, two brass fittings, some hose and a plastic piping end cap. The end cap should be slightly smaller than the inner diameter of your intake boot so you can clamp down on it. Make a hole in the middle of it for the brass fitting. Put a shop cloth around the other end (to prevent burnt shop cloth from entering your intake) and clamp it in. Do the same with the paint can lid and run the hose between the two. Make another hole in the lid for your compressor/bike pump and connect it. Grease up a shop cloth with WD40 or some other slightly flammable material, put it in the paint can and light it up (obviously observing fire safety!). Seal the top on with a mallet and then turn on your pump (slow steady air flow is best). Voila, all your vacuum leaks will be emitting smoke.

Back to the main story... Went to the junkyard to get an E36 TPS, 60-2 and VR sensor. Installed the 60-2 and VR sensor today along with a new timing belt and thermostat for good measure. So, based on some crude eyeballing, the sensor in it's stock location appears to be 107.0 degrees BTDC. Is that correct? I assume however there is more to it based on whether or not I trigger on rising edge or falling edge, etc. Obviously I need a timing light to confirm but is 107.0 a good base number to start with or is there more to it?

EDIT: After more reading, it appears 84.5 degrees is the right trigger angle for the rising edge of the 60-2 (to be timing light confirmed). Can someone tell my why that is? Big difference between what looks like 107.0 and 84.5 so I am missing something fundamental here! Maybe the sensor bracket is inverted.

[ianwood]

So who wants to help me here? 84 degrees is the right cranking angle for the rising edge of the 60-2 trigger for TDC. There should be 14 teeth between the TDC mark and the sensor with each tooth being 6 degrees. Well, after mounting up the VR sensor, bracket and trigger wheel pulled from a late model 325i, there are 18 teeth between the TDC mark and the sensor which is 108 degrees. 60-2 wheel is mounted properly and aligned with the dowel. The timing cover with the TDC mark on it is installed properly. And I've double checked the VR sensor and it only mounts one way. I changed the timing belt when I installed the 60-2 and the timing marks on the cam and the crank lined up perfectly. So, what am I missing. See picture above for what it looks like. Someone please tell me why I see 108 when I should see 84.

Megasquirt board assembly is almost done. I've built a truly hacked Lemons-style adapter plate for the E36 throttle position sensor. JimStim and sensors should be here by the weekend. I've read a whole lot about spark advance and now understand the fundamentals about when more advance is needed vs when it needs to be dialed back. And even better, it actually is making sense. I am looking forward to getting this thing installed and doing some actual tuning but I need to figure out why my crank angle isn't right first!

[mooseheadm5]

Sorry I missed your post. It sounds like you have the adapter covered. It really was simple, eh? If you want a more "professional" one in the future, let me know.

As for the crank triggering, I'm pretty sure you can set an offset up in Megasquirt. Your sensor is properly mounted.

[turbodan]

I'll check my laptop later and post my trigger offset. Megasquirt is designed to factor in a trigger offset. Its one of the basic settings you go through during the initial setup.

[turbodan]

Mine is set to 60 degrees. I believe thats pretty much universally used for M20 Megasquirts.

[ianwood]

OK. I think I might have figured this mystery out. There is indeed 107 degrees between the TDC mark and the sensor. The difference (and all my confusion) has to do with where the O | T mark on the 60-2 is in contrast to the tooth that MS triggers off of. If you rotate the crank so that the O | T mark lines up with the mark on the timing cover, the tooth that MS triggers off of (the first tooth after the gap) is a little over 3 teeth away. At 6 degrees per tooth, that's about 20 degrees. So take 20 degrees away from 107 and you get 87 which happens to be the MS trigger angle for the falling edge (VROUTINV).

So, it comes down to this: at TDC, the sensor is 87 degrees (or 84 for the rising edge) past the tooth that it triggers off of. Ah, I feel better now!

Turbodan, 60 is a number haven't encountered in my numerous searches on crank angle. I haven't even powered up my MS unit yet but is your sensor installed like mine is? I am not sure how you would get 60. Have you checked it with a timing light? From everything I've read on e30tech, it is 84 for the rising edge trigger and 87 for the falling edge trigger.

Sorry I missed your post. It sounds like you have the adapter covered. It really was simple, eh? If you want a more "professional" one in the future, let me know.

For Lemons, a beaten up home made bracket helps keep the judges from thinking you've overspent your budget. But, if this MS install goes well, I may be doing another on a street car and I will take you up on it.

[Shadow]

It's going to be in the high 80s. I used 89 for the old m20.

[ianwood]

Here is the E36 TPS and the adapter plate I made from a Home Depot metal strap (I'm a big fan of these!). I also discovered the fuel tank was getting rain water in it. Our fuel sender bracket was a big rusty mess. We pulled the tank to be cleaned and I adapted a new bracket to our Wallbro 255gph fuel pump. New gaskets going in today. Anyone else get a little freaked out by an exposed 12V wire going right into the fuel tank? I know it's normal and it's like that when stock but still a little disconcerting. Apparently fuel is non-conducting. I put three layers of heat shrink on the connections just to be extra safe.

Also today, I am getting started on the MegaSquirt wiring harness. MS2 unit is almost ready. Hopefully bench testing it tomorrow with JimStim.

[M. Holtmeier]

Don't run that stock connector between the pump and pickup with clamps. It will only split open. I did the same thing on mine only to go back and install fuel hose with an adapter.

[ianwood]

Progress! V3 board is done and working! Miracles do happen! JimStim shows injectors, spark, VR is set up correctly. Harness is installed. I haven't tapped the existing harness and instead am wiring the MS harness to be completely independent. Did this for two reasons:

1. In case I can't get MS dialed in before the race, I can simply revert back to the BEGI.
2. If I do get MS dialed in, I can remove the horrible nasty old stock harness and be done with it forever!

Picked up some pretty beaten up 42lb green top injectors. Did some very basic testing on them using the MS test mode. One leaked but thankfully I picked up 8 of them so swapped in a spare which worked. IAT and CLT sensors installed which completes the list. Initial calibration and base settings are done. Just finishing up my first spark and fuel maps. Both are composites of maps from similar setups to get me started. Been doing a lot of homework and I have a much better understanding of timing than I ever did before. Tomorrow I am testing the coil, CPS, O2 and other sensors and then it's time to start it up. Hoping to tune startup and idle pretty quickly.

Does anyone have a cranking PW table I can study? I really don't want to flood the engine. Priming PW I am setting to zero.

Don't run that stock connector between the pump and pickup with clamps. It will only split open. I did the same thing on mine only to go back and install fuel hose with an adapter.

Agreed. That little stock hose is super flimsy. Used high pressure fuel line there.

[turbodan]

-10---207
8------198
28----181
50----155
70----137
100--118
120--110
140--105
165--103
188--102

This works pretty well for me.

[ianwood]

She cranked right up on the first shot! It was lumpy and wouldn't take any throttle without getting really unhappy but I soon found out that it was because I was losing sync. After a number of tooth logging sessions and some guess work, I realized I had inverted my CPS signal while the board was using VROUT and not VROUTINV. A quick wiring change and now it can rev. Idle is still not pleasant and is oscillating which is what I will work on next.

I've temporarily set REQ_FUEL to 6.8 and all of my start-up tables are pretty low to get it started without flooding. Next step is to set REQ_FUEL to 7.2, confirm timing, tweak the tables, and dial in the idle. My start-up tables so far:



EDIT: Start up tables now reflect more realistic settings.

[skip535i]

I admire you guys who do your own tuning. Hot damn.

[ianwood]

Drove it around for a bit. Very lightly. AFRs are out of whack but at least my initial timing hasn't broken anything and engine revs nicely. My start up settings need work. That one crank and instantly running first attempt was a fluke. I have to give it throttle input to get it to start. I've increased cranking pulse and ASE in the tables above and will be trying them today.

I've also built a new timing map composited from some existing eta maps which I think is pretty conservative. I'd appreciate some feedback on it. Note, it's a bone stock eta and head (maybe with some slightly better than stock head bolts! Shhh!) with a smallish intercooler but we'll be running on 100 octane. Head was redone about 3 races ago. Plan is to street tune (tricky when your car isn't street legal and looks like a big fish!) and then hopefully put it on the dyno with a couple of maps to test. Initially, we'll run at 7psi and then re-introduce 11-13psi when we're confident about the tune (and my ability to tune!). The race is 24 hours non-stop and in the desert on July 1 (think hot) so I am thinking conservative timing. 111 fuel also an option if need be. Feedback requested!



Fuel map I will let VEAL (great acronym!) take care of for me. TunerStudio is pretty amazing. One thing though. My WBO2 is about 9ft from the engine (3ft of stock'ish piping to turbo then 6ft of 3" pipe, no cats and barely any muffler on the other end). Is there an easy way to determine the lambda delay? I think mine might need more delay but I am only guessing. Also, what does a good AFR log look like? Mine seems very bouncy.

[ianwood]

Drove it around last night for a good 20 minutes with TS cooking the VE table in real time. It took almost no time for it to dial in the target AFRs. Really impressive! Built up to some moderate pulls (60% throttle up to 4,500 RPM at about 5psi boost) and all is good so far. No odd noises. Nothing that remotely sounds like detonation. Smooth revving. No sync loss. I don't know why I was so daunted by MegaSquirt before. Very happy with it so far. We might be taking it to the track as soon as Friday. I think I can safely start removing the old harness now!

[ianwood]

Fuel supply issues have sent my tune in a loop-di-loop. The stock 2.5 bar FPR was only delivering 24psi so I got a junkyard 3.0 bar one (this is Lemons after all) to replace it. We did that on our way to Buttonwillow for some on track tuning. When that FPR also indicated 24psi, I started to scratch my head. But, we were at the track and so I tried to tune it to 24psi and would deal with the fuel supply later. Well that didn't work so well. A couple of runs with Tuner Studio fixing my suddenly lean VE table -- spec E30s flying past me like rockets -- me lifting every time I think I hear detonation, as I kept hitting SEND to update the VE table hoping that Tuner Studio will work its magic and get it dialed in. But the car kept running lean and not much later I noticed the fuel pressure started to drop to 22psi then 20psi. Baffled, back in the pits, we noticed the brand new gasket on the fuel level sender had squeezed itself out. That's when the penny dropped. The fuel tank was getting the missing fuel pressure! A shredded fuel level sender gasket was the proof.

Ironic considering the warning a couple posts up, the high pressure hose I used to bridge the 1.5" gap between the Wallbro fuel pump and the feed pipe split open. 300psi hose! Albeit probably stretched too much to bridge the difference between pump outlet and the brass pipe. The shop at the track had better hose that we rigged up to fit without stretching. $10 a foot!

By the time we got it sorted out, the track was closed for the day. With the tune totally invalidated by a busted hose, we decided we could still make the day count with a little country road tuning. The new fuel pressure was 39psi (more on that in a sec), so I adjusted the REQ_FUEL accordingly and went about re-tuning VE. Armed with the radios, I tuned the shark while Clay drove 30 seconds ahead of me looking for anyone who might not like an unregistered shark screaming down a country road. Initially very rich to the point it bucked under boost, I made some manual updates and Tuner Studio did the rest. After 30 minutes, I was getting consistently decent AFRs. About d*** time!



But what's the deal with 39psi?!? It's not the 43.5 it should be and when I turn the fuel pump on before I start the car, the fuel pressure is 48psi?!? Once it starts, it drops to 39psi and pretty much stays there maybe going up a few psi under load. Is this normal for a 3.0 bar FPR??? Should they be rock steady at 43.5 or do they actually go up and down based on the vacuum signal? If I unplug the vacuum line, I get rock steady 48psi (engine off, idling and revving without load, didn't test under load). After some research, I understand that the vacuum signal to the FPR allows it to maintain a consistent pressure difference between manifold and injector by varying the pressure on the diaphragm. This means a 3.0 bar FPR actually changes rail pressure based on manifold pressure (if that's obvious to everyone else, well, I am just learning). So what pressure do I use to calculate REQ_FUEL? 39? 48? Or 43.5? My REQ_FUEL is set to 7.7 which is based on a Bosch 42lb injector (apparently it's really 40lb at 43.5psi) being used at 38psi. Note that the digital fuel pressure gauge that is wired into the car may need some calibration so could be up to 3psi off.

I don't know if my FPR is normal. Because the fuel pressure is higher when cranking, I've had to turn ASE almost completely off, and my cranking PW is 15% at 160F. Much higher and it's filthy rich on start. Mind you there is no ICV so air is only what gets through the throttle plate. But still, 15% seems odd!

I've got a Megasquirt headache. I need to tune this thing for racing this weekend!!!

UPDATE: I think I've got it figured out. Guess I just had to write it all out and then sleep on it. It's starting to make sense. Subtract manifold pressure from fuel pressure and you should get 3.0 bar. I get 48psi but I think my gauge is off by a few psi so I will recalibrate it today and reconfirm.

[ianwood]

On a separate note, I think I've got some heat soak to tune out of my IATs. Either way, my IATs are high! Ambient was about 85 and the peak IAT was 140! Pretty sure the sensor is heat soaked which I didn't know much about before this. It's about 6 inches from the the throttle body. Steel charge pipe (more Lemons makeshift engineering on zero budget). Thinking that wrapping it in ceramic heat wrap might help. How do I tune MAT correction? Definitely do not want to lean out from heat soak!

BTW, I am writing the names of anyone who helps me on the side of the car for this race! Just imagine the fame!



Note the max CLT was me forgetting to put the fan on while idling earlier. Pretty steady between 160-180 normally. Also chasing a hunting idle as you can see. And i will be rebuilding an alternator that has questionable output below 2,000RPM which isn't helping.While we're at it, how does this look?



One thing that I noted was the lumpiness of the TunerStudio tuned VE table. I've just moved the WBO2 from 9' to the just after the turbo exhaust (about 3') so hopefully that will help correct AFR lag. Manually removing the lumps.

[turbodan]

A lumpy AFR table may be indicating insufficient acceleration enrichments. AE has a pretty significant effect on driveability.

[marc79euro645]

Flatten out the ve map in the idle area, spark map too, this will give you a steady idle.
My fp starts and idles at 28# & goes to 40# under boost.
My ve is lumpy too, like Dan said AE will play in here,attempts to lean out the lumps will probably lead to lean stumbles under acceleration. I've yet to retune with richer AE, but I intend to in the future, this should allow for leaner cruise ve.
I use a 6ms prime pulse which helps mine start quickly.
good luck
marc