Megasquirt Progress, Main Board complete Options

[DNHunt]

One piece of advice. Plan your cable from the ECU to the relay box carefully. Try to avoid routing that leads to bends at the DB 37 connectors. This IS the weak spot of the system. The DB 37 offers very little stress relief and the solder joints can break. I had this happen in a Friday commute. Not fun. Luckily I had my laptop and I reprogrammed the VE table so MS ran with no MAP input. One of the things holding me back from building my V 3.0 board is a search for an alternative to the DB 37. A trip to Boeing surplus is in order. It is undoubtedly the best junkyard in the world.

I would go so far as to plan on having to repair that cable. In other words, think about how you can get to the ends of it to solder. Mine goes through the firewall and the hole in the firewall is not big enough for a DB 37 to go through so I have to solder with the cable in the car. The advantage is that my FI harness can be dropped with the engine or can be removed for changes.

Dave

[yarin]

Dave,

I will be sure to strain relief both sides of the DB-37 cable properly. Actually I have the perfect solution.. I have tons of these little things available at work. Just 1 screw into any surface and a quick ziptie.

I visited Boeing 7-8 years ago when i was on vacation in Seattle. I would KILL for a trip to their surplus yard. I work at a semiconductor company that cleared out a bunch of their old stock, a lot of stuff was chucked. Sooo sad.

Since only about 1/2 the pins are used on the cable wouldn't it be nice to build a MS system with two nice big sturdy LEMO connectors? (IMG:http://www.914world.com/bbs2/html/emoticons/blink.gif) Those things aren't cheap, but they are bulletproof. Maybe a combo of that and a big 8 pin COAX Sub-D for the injector signals, dizzy pickup and power signals.

[DNHunt]

As a guideline I have a 180 degree turn a the relay board end of the cable. I have it tied down so the bend is about a 6" diameter and I have had to repair this twice in about 3 years. Definitely not fun.

Dave

[bd1308]

BNC connectors!!!?!!?!?!

those things are nuts.

[yarin]

I bought the 12' cable from DIYautotune.

The finished end is crimped, the other end is do it yourself solder cup. Should I connect the solder cup end at the MS and the crimped end at the relay board?

[agrump]

Yarin,

I am using a pertronix. I did not use the Ed capacitor, I just changed the resister. I know what you mean about changing something and then having to change it back. I was pleased that I did not have to change the R57 back once I started controlling timing.

I did get the cover. I got it a long time ago when I was running MSI/V2. Godd idea about attaching the cover. Be warned that if you do decide to mount it low it is tough to get the wires into the terminals especially if two or more need to be connected to the same one.

I am using the stock fast idle device. I stole one of the connections for the IAC to use for my MAP sensor which I have mounted in the relay box. SInce I started with MSI/V2 I built a resister box for the injectors and am still using it so no PWM. For warm up enrichment I am using the stock head sensor. I manipulated the table so that it reads about 100 degrees lower the the actual temperature. It works ok but still cuts off to soon. There is a limit to what you can do by changing the tables, I think I got it about as good as it is going to get with the stock sensor. I am thinking about mounting a sensor next the the stock thermostat and see if I can get better readings there. With MSII you don't have to mess around with loading thermal tables there is a menu item under tools in megatune that will handle it for you. I am currently using:

5280 ohms -40 degrees f
2000 ohms -6 f
200 ohms 130 f

Again it is not perfect. Interestingly MSII limits the lowest temp you can enter to -40, I was using -48 with MS1 for the first entry.

I found a tps that fit within the plastic cover of the stock throttle switch. I wired it into the stock connector, it looks factory. I will try to dig up where I got it, it was online somewhere. Attached is a screen print of my warmup settings. I used VEXME software to tune my tables with a LC1 WB. I am playing around with the PID algorithm currently, the simple WB algorithm seems to oscillate to much.

I am with Dave in not trusting the stock connector although I have not had a problem YET. The minisquirt has a much nicer one but I can’t imagine changing it now, what a pain that would be!

[DNHunt]

Also on the injectors and fuel pressure. I believe the stock 2.0l injectors are designed for 28-36 psi. On the low side atomization is poor and on the high side the pintle has problems closing. I'm kind of fuzzy on that.

One thing to look at Borg Warner makes replacement injectors for some if the later Bosch L-jet and K-jet systems. The output is way less so they are easier to control. These are barbed fittings in some cases and they all use the improved l-jet electrical connectors. Also, they are cheap. One problem the hoses are not swedged to the barbed fitting well at all and may leak. It's best to redo them with good FI clamps.

The resolution with the MSII chip is 100X more sensitive that the MSI chip so you will be able to tune pulsewidths very well. I plan on using my 2.0l injectors and I expect it to be very smooth indeed. I ran a fuel only program that had 10X resolution before (Hi-res code) and it was very smooth. Unfortunately MS extra doesn't have the room for this. The 100X resolution was a big consideration for me to upgrade.

Dave

[bd1308]

If i do this on my 1.7, is there someone that could help me out?

b

[sgomes]

LEMO (IMG:http://www.914world.com/bbs2/html/emoticons/drooley.gif)

[yarin]

Check these out, FCI D-sub Mixed Power DW series

Not cheap.. but would also work for a PCB mount all in one cable. Plenty beefy with the right mating connectors and strain reliefs.

[DNHunt]

Britt

There is plenty of support online. Meuller could do it and it worked until he spilled coffee on it. I'm a dentist that can't tell you how a diode works or why it is in the circuit. I can't program in assembly or C, but I read and absorb things. MS just needs a couple of things to really work.

Research, plan your instal before you start. Once you decide where you want to go resist changing things midsteam.

FOLLOW the instructions. It's like cooking, if you follow the recipe it will work.

Take your time and ask questions if you are confused. It's no different than school. If you ask enough people someone will explain it in a way you can understand.

Be ready to problem solve in a LOGICAL step by step sequence. Something will go wrong and you will have to figure it out.

When you get done you will have a basic understanding of FI and something that you can be proud of.

Dave

[yarin]

QUOTE (DNHunt @ Feb 1 2006, 06:57 AM)

Also on the injectors and fuel pressure. I believe the stock 2.0l injectors are designed for 28-36 psi. On the low side atomization is poor and on the high side the pintle has problems closing. I'm kind of fuzzy on that. One thing to look at Borg Warner makes replacement injectors for some if the later Bosch L-jet and K-jet systems. The output is way less so they are easier to control. These are barbed fittings in some cases and they all use the improved l-jet electrical connectors. Also, they are cheap. One problem the hoses are not swedged to the barbed fitting well at all and may leak. It's best to redo them with good FI clamps. The resolution with the MSII chip is 100X more sensitive that the MSI chip so you will be able to tune pulsewidths very well. I plan on using my 2.0l injectors and I expect it to be very smooth indeed. I ran a fuel only program that had 10X resolution before (Hi-res code) and it was very smooth. Unfortunately MS extra doesn't have the room for this. The 100X resolution was a big consideration for me to upgrade. Dave

Dave,

If the 2.0 injectors that are on the way are good i'll keep em for now. My original plan was to build a intake plenum, maybe individual throttle bodies.. but that has been put on hold.

Thanks for the info on the Borg Warner injectors. I'll hold off till I can build an intake manifold with all the goodies and pickup better suited injectors.

If there are atomization problems with the 2.0L injectors at low fuel pressure and pintle closing problems at higher pressure why are you going to use them? What pressure should I run them at? I'm going to use Summit's modified adjustable manifold referenced fuel pressure regulator. link here I don't want to end up with low fuel pressure(manifold vacuum) causing atomization problems. Is there any type of feedback I can look for aside from volume flow bench testing?

[yarin]

QUOTE (bd1308 @ Feb 1 2006, 06:59 AM)

If i do this on my 1.7, is there someone that could help me out? b

There are enough people on this forum that have squirted their 914s that can help you. Also check out the MSefi.com forums. There is tons of info on building and tuning your MS system. That's where everyone on here got most of their MS specific questions answered. I'm in the build process right now and have received plenty of great information.

THANK YOU 914 MEGASQUIRTERS!

The real tricky part coming with tuning, i'm about a month or two away from task. See my link earlier in this thread. I created a diagram for all of the MS wiring on a 914 with the accessories. At least it's what i'm planning on doing.

Next step is finding a TPS from a junkyard. Gotta get my throttle body first, 2.0 parts(injectors, plenum, runners, etc) should be here Friday.
(IMG:http://www.914world.com/bbs2/html/emoticons/driving.gif)

[lapuwali]

IMHO, Dave's experience with solder cup DB connectors is exactly why you should NOT use them in a car. Use DB connectors with crimp-on pins. They're more expensive to buy up front (esp. as you have to buy the right tool to crimp the pins), but they're much more reliable in a high vibration environment.

The other thing I intend to do, and I'd suggest for others, is to mount the board in an oversize box, and run the DB connectors to some other connector that connects to the box side. My actual plan is to put the stock metal box with endplates, but the connectors not tied to them (metal box for EMF protection), and place THAT in a waterproof plastic case. A short cable will connect the DB connectors to waterproof connectors that pass through the plastic case walls. My setup will go in the engine bay.

I'll also rubber mount the metal case inside the plastic case to reduce as much as possible the vibration passed to the board itself.

Mark Henry has had good luck mounting his SDS inside the stock ECU case, and that's also an excellent option. I may still end up doing that myself, since I have a couple of useless 1.7 ECUs. The MS relay board and the MS itself should fit inside the stock ECU case quite nicely.

2.0 injectors will work, just run them at stock D-Jet pressure: 29psi (2 bar). Most injectors now are run at 3 bar (43.5 psi), which is too high for the D-Jet injectors. The main difficultly with 2.0 injectors is they're HUGE for the stock engine, presenting some problems with idle pulsewidths, which are usually down around 2msec. The MS resolution of 0.1msec means you have a minimum step of 5% at such small pulsewidths. Use a manifold pressure referenced fuel pressure regulator (see posts by airsix on how to convert a stock FPR) to help significantly here. You can also sell the 2.0 injectors and find a set of 1.7 injectors instead, which are still plenty large enough for even a modified 2270, yet small enough that idle pulsewidths are a good bit longer, giving you more headroom to get the idle tuned right. Since D-Jet is all analog, it has no "resolution" on pulsewidths, so it can run the injectors right down to their minimum pulsewidths (about 1.7ms) with infinite control.

D-Jet injectors are mostly used because they fit the stock intake setup well. There aren't many hose-barb injectors to choose from. Most need a fuel rail. 1.8 Subaru injectors work (and are smaller than 1.7 D-Jet injectors), as do some L-Jet injectors, and both of those injectors will work at 3 bar, which will give better atomization than 2 bar D-Jet injectors. I intend to use 1.7 D-Jet injectors, because I have several sets on hand, including a NOS set still in original boxes.

[yarin]

James,

Thanks for you feedback. I will run the injectors at 29psi. However the fuel pressure regulator sold by summit claims 36-45Psi. Someone on her said their regulator is a modified stock D-jet regulator. I would hope that I can turn it down to 29psi. I guess time will tell. It's pretty difficult finding a manifold referenced adjustable fuel pressure regulator for the 29psi range for under $200. Do you think I will be ok?

I'm looking at Digikey p/n HM131-ND for the plastic enclosure for the relay board along with the clear lid 1591EC-ND. It's not water tight, but water resistant. I will secure some form of hood over the cable entry/exit holes to defect any water that may happen to enter the area. I was also planning on using rubber mounts to dampen road vibrations.

I was afriad that the solder cup connections would become unreliable due to stress and vibration. Crimp is the way to go, assuming you have proper tooling. Pliers and a screwdrivers will yield a less reliable solution than a solder-cup connection IMO.

Can you explain your EMF solution? Are you avoiding grounding the metal case? Is this the way to go?

Thanks for the info!

[jhadler]

The stock D-jet regulator is pretty good. You say "manifold referenced", do you mean rising rate? Why do you want the fuel pressure to change? You've already got an ECU that can be programed to compensate for manifold pressure. Seems you'd be adding yet another variable that you don't need.

Are your injectors failing to work at idle or redline? If they're good there, why do you need dynamic fuel pressure changes? Maybe I'm missing something?

-Josh2

[yarin]

Josh, Yes rising rate. Or in our case when the manifold is under vacuum the fuel pressure is regulated to ambient - vacuum. If the manifold is boosted, Fuel pressure increases proportionally. Unfortunately the latter condition doesn't come in to play for me (IMG:http://www.914world.com/bbs2/html/emoticons/smile.gif)

From the MS site:

"The vacuum referenced fuel pressure regulator is essential. It provides constant pressure differential between fuel at injector nozzle and manifold air pressure [port EFI] or atmospheric pressure [TBI]. This makes the injected fuel quantity solely a function of the injector open time.

If you were to 'cap off' the manifold vacuum port on the fuel pressure regulator, you are reducing the dynamic range of the injectors. This means you will need lower pulse widths at at (giving less control over idle mixtures) and lower flow under boost (restricting the maximum horsepower).

So, in general, for port injectors, have the fuel pressure regulator connected to the manifold vacuum is a good thing. There is very little reason not to do it (though some have argued against it for individual runner port EFI set-ups). "

Since 2.0 injectors are rather large I was under the impression that I would run into minimum pulse width limitations at idle, therefore lowering the fuel pressure at idle would help. Correct? Or would PWM take care of those issues?

Thanks

[lapuwali]

You can modify the stock D-Jet FPR to reference it to manifold pressure yourself. Search here for posts by airsix, he covered it in detail. It's pretty simple, basically just drilling a hole and fitting a small tube with epoxy will do. If you don't have a stock FPR, you can pick them up on Ebay pretty cheap. I may even have a couple lying around in the parts bin. I'd send you one for shipping if you're game to try the conversion yourself.

On EMF, I was planning on running the ground connections from the DB connector to a bolt through the outer case, then one big ground wire from that bolt to a case bolt. Another connection from the inner (metal) case to the bolt will ground the case, and allow me to run a lot fewer wires through the outer case connector. If I put the whole mess in the stock ECU case (which is sounding more and more attractive every day), I'll not have an inner and outer box, just the metal ECU case, but I'll still do the ground bolt trick. I'll also make a plate to cover the stock connector hole and run my waterproof bulkhead connectors there. One for EFI signals, one for serial signals (with a cap for when the serial cable isn't connected).

Josh, you want a manifold referenced FPR to keep the fuel pressure differential across the injector the same. If the rail pressure is fixed, then the pressure differential across the injector goes UP as manifold pressure goes DOWN. Manifold pressure is low at low loads, and high at high loads, so the pressure differential across the injector is high at low loads and low at high loads, which is the opposite of what you want. You have to compensate for this by making the difference in pulsewidths wider, but you can hit physical limits at either end. The injector cannot open and close in less than ~1.7ms, so if you need t make it shorter than that to compensate for the high pressure differential at idle, you can't get the idle to be lean enough. Injectors also can't be open longer than 80% of the time without overheating, so you may not be able to make the pulsewidths long enough at high loads to overcome the lack of pressure drop across the injector with 1.0atm in the manifold.

With a manifold referenced FPR, however, the differential across the injector is fixed regardless of manifold pressure, so you end up with longer required pulsewidths at the low end, and shorter required pulsewidths at the high end, for the same fuel flow through the injector. Thus, you can run smaller injectors than you might be able to otherwise, and thus get more tuning flexibility at the extremes.

[lapuwali]

QUOTE (yarin @ Feb 1 2006, 01:22 PM)

Since 2.0 injectors are rather large I was under the impression that I would run into minimum pulse width limitations at idle, therefore lowering the fuel pressure at idle would help. Correct? Or would PWM take care of those issues?

PWM is there to limit current through low-impedance injectors (and the D-Jet injectors are low-impendence).

Basically, since current = voltage/resistance (so Mr. Ohm said), if you lower the resistance, and the voltage stays the same, the current increases. For our purposes, impendence = resistance, so low-resistance injectors will allow too much current to flow through them (overheating them) unless you provide some way to limit the current. L-Jet uses external ballast resistors to do this. MS gives you the option of using external resistors, or using PWM to limit the current. PWM works using a "peak and hold" strategy. It delivers a full 12v at first to get the injector to pop open, then it starts to switch the voltage on and off rapidly (too fast for the injector to react), and the duty cycle of the switching lowers the average voltage seen on the circuit. Lower the voltage, with fixed resistance, and you lower the current. D-Jet does this, too, but it can actually lower the voltage directly, since it's an all analog ECU. It doesn't have to resort to digital tricks like PWM.

Using peak and hold instead of just limting the current with external resistors means you get faster response times from the injector (particularly in opening), so, again, you gain a bit in injector dynamic range at the idle end of the spectrum. L-jet just copes with slower response times for a simpler circuit, and L-Jet can't handle large injectors easily, as a result. L-Jet is famous for leaning out on top, as a result, as it must use pretty small injectors to idle well, and that's where all emissions testing is done.

[DNHunt]

I understand the rising rate FPR but, I haven't found it necessary and I'm not going to mess with it.

I've only had a few grips. The idle flucuates with alternator load (I can't imagine running AC) so idle control would be nice. Over run is difficult to tune but was very acceptable with the greater resolution of the Hi-res code. I suspect that will also be the case with MSII. The IAT compensation is off due to heat soak of the sensor in the plenum. Both idle control and air density corrections are being worked on by other people with similar problems.

The basic B&G codes for fuel only give a nice replacement for D-jet with the benefit of tunability. Adding ignition control helps even more but, the system is not perfect and is still short of what most of us are used to with modern FI.

All in all I've been very happy with my stuff. I've had to repair a couple of solder joints, replace 1 transistor and 1 MAP sensor in 3 years and about 8000 miles. The same board has been used with 3 separate engines and in 2 forms (feul only and fuel with wasted spark distributorless ignition). I can't begin to remember how many different code versions I've used.