I am at my wits end with this megasquirt stuff. I have tried almost everything and can't get it to fire.
Here is what I posted on the MS forum just now...
Here is the setup:
MS3.57x built by DIYautotune
Buick 350 ci – new engine, never started
Ford EDIS system with trigger wheel (set up with missing tooth 5 ahead of sensor)
MS Relay board
Injector wiring to relay board is as follows:
1&6 wired to INJ1
8&5 wired to INJ1 (second slot on relay board)
4&7 wired to INJ2
3&2 wired to INJ2 (second slot on relay board)
Power for injectors is from relay board
VR sensor is wired correctly
I have spark. When cranking and with a plug out the spark timing appears to be close.
I can't seem to attach the tune here unfortunately but I can email it to anyone who is willing to look at it.
I have the tune set for semi-sequential based on my injector wiring. If I have my main fuel outputs set to standard fuel it will sputter a bit. But I know this is wrong for semi-sequential so I have it set at MS3x fuel.
I tried increasing the cranking pulsewidths because that will supposedly help with starting. Not even a sputter. I pulled a plug and it wasn't wet, even though I had the cranking pulsewidth set around 350 and then at 400 (I varied it) for the current coolant temp. There was a faint smell of fuel on the plug.
I am wondering if my injectors are actually firing based on this result?
If my injector wiring is okay and my tune doesn't look too screwy, what should my next step be?
Thanks in advance for any help. I have been trying to figure this out for more than a week now and am getting desperate.
Chris
Chris, I'd say the first step is to revert to old timey troubleshooting steps. As we all know, you need fuel, fire, and compression. So let's see first which one is out of whack. New engine, compression should be fine. But if you haven't checked it you might do that just for a baseline.
Next let's eliminate either fuel or fire. Grab a plain old squirt bottle and spray a generous squirt of fuel down the intake. Unplug the fuel pump, crank it and see if it'll fire up on the manual prime squirt. If it does, use the squirt bottle to keep it running and see if the engine sounds like you can make it enjoy running. If you can your timing is reasonably close. If you can't, look more closely at your timing advance. Before you can ever get the injector shots right the timing has to be in the ballpark. I haven't heard you say yet that you've put a timing light on it, so if you have a buddy to help you maybe he could check to see what the timing is doing while you are keeping it running. No point in me repeating where that should be.
Once you have the fire where it should run OK you can start playing with the fuel.
Now, the only real difference between full sequential and everything else including batch fire is at part throttle and idle with the timing of the shot, and as long as the fuel gets in the cylinder the engine really doesn't care when the shot is fired. Any difference in the way it runs will be very minor and usually undetectable. So you can assume fuel is fuel.
Is your line pressure right? It needs to be rock steady at full injector saturation (all injectors wide open). You can simulate this by wiring 12v and ground directly to the injectors. That is one way to check if your injectors are doing their job. You can use a "Noid Light" to see if they are getting a signal.
Divide and conquer is the old troubleshooter's mantra. It applies here as well as it ever did anywhere else. Do not automatically assume it is the controller, more often it is something else.
Jim
I didn't have a squirt bottle so tried some starter fluid and it started immediately but died but it did sound like it was running okay for a brief bit. I think it has something to do with my injectors not firing so I am going to focus on that.
So there isn't any issue with how I have wired my injectors together and where I have connected them on the MS relay board?
Thanks as usual Jim. I used to say "I know I'll figure this out" but after two weeks I am starting to lose confidence!
I have tested the injectors in test mode using a test light in the connector. I am not getting a pulse and not hearding anything.
Are the two wires on injector connectors dedicated specifically to power and ground or does it not matter as long as you have them all hooked up the same way?
I am grasping at straws here!
I grabbed your tune and log files off that other forum to see if they'd attach properly to a message here.
cgill-tunefile.msq
cgill-logfile.msl
Chris, you should have power applied to one side of the injectors through the main power relay and the fuses, then the controller supplies the ground. So it sounds like your controller may not be grounding the injectors, but first do a test to confirm power through the injectors. You can put a clip lead on the control side, tie them all together with jumpers, and ground the lead. That should turn on all your injectors at once, full on and allow the engine to wind up with the throttle all the way open.
BUT consider SAFETY! It would be wise to disconnect the injector leads from the controller before doing this test. It would also be wise to be able to yank the clip lead off to kill the engine.
If you do this test properly, you should be able to give the engine a pulse, crank it to fire, then pulse the injectors manually and keep it running with the throttle plate at a high idle. The more throttle you give it, the longer you will have to supply a ground to the injectors, basically letting it nearly die, grounding it which should cause the engine to rev up until it starts to choke out from too much fuel, un-grounding it at which point the engine should smooth out and increase rpms before starting to die again.
This also works at full throttle except you remove the ground before the revs get too high.
This should work as described. It is also consistent with not seeing a ground pulse from the controller, so then you look at that side.
First off, without a scope you may not be able to detect the shorter firing pulses, and with the injectors disconnected you have to supply a pull-up voltage to see anything because the controller is supplying a ground rather than power. So my first question has to be, "Is B+ supplied to the opposite side of the injectors?"
Provided all is well there, the next place to divide troubleshooting is software/hardware. Both have to be set up right for it to work.
You have a guage on your virtual dashboard that shows injector pulsewidth. What readings are you getting there? This should vary from about 1-2ms (milliseconds) to something under 25ms. 25ms is 1/40th of a second, or 1/2400th of a minute, so an injector firing for 25ms in an engine running 4800rpm is on all the time. (Two crank rotations per firing pulse) You only ever want your injectors going solid at redline and WOT so if they are going solid below redline you need larger injectors.
That gage will tell you whether your problem is in the hardware (bad wiring, fried output transistors, etc) or software (setup parameters, table entries, etc). From there you can narrow it down further.
I would suspect the injector setup entries myself, it's easy to make mistakes there.
Jim
In my browser, if I click on the msq file link (above) I see a whole lot of black text on a white background. I pulled down one of the menus in the Firefox toolbar and selected "Save File As". That let me put your msq file on my desktop.
Then I opened TunerStudio, which by default showed my my own tune file. When I selected "compare tune", TunerStudio showed me a nifty comparison of your tune selections versus my own.
I didn't find a smoking gun - someone more experienced should look at your tune and log - but I did find a couple things that might be good to change.
You'll find an injector timing table under "Fuel settings". Your table currently has "360 degrees" entered for every cell. I found that my engine idles/runs a whole lot smoother with that table filled in differently: basically "-300" up to 1000rpm, transitioning to "-220" for every engine speed over 3000rpm. Also, you've set your injector size at "252cc". I've got my injector size entered as "232cc". I expect one of us probably made a typo. (Might be me.) Under "Ignition Options", in the "Cam input (if used)" box you've selected "MAP" sensor. Did you mean to do that? (I'm using CMP there. There should be other options available.)
Thank you guys!
Jim. I am concerned about doing that with my engine since it hasn't been broken in yet. I am almost convinced that you are right in that the controller isn't grounding the injectors for some reason.
When I connect the ground wire on an injector plug to a test probe and ground the probe to the chassis I can hear the other grouped injector click (I have them wired in pairs for semi-sequential). So it sounds like the ECU is not firing the injectors, but why?
Curtis, my injectors are rated at 24lbs/hr so they are a bit bigger than yours. Thanks for the information on the differences between your tune and mine. I'll change the injector timing table. As per the "cam input" I don't have a cam sensor and the only options in the dropdown are MAP sensor, MS3x Cam in, or JS10. I thought MAP sensor was the closest option. Am I wrong?
Curtis, can you post a photo of your injector timing table?
But if you did have a cam sensor, you'd connect it to the cam input pin (#32) and not to the MAP sensor pin, right? I don't know if it matters at all; it just looked a little fishy to me.
Here's that table.
Screenshot2015-03-29at11.02.15AM.png
(I don't remember who tipped me off to try these values.)
Thanks Curtis. Why are there negatives in your table? Mine are all +360 right now.
You've got 720 degrees of injection timing to play with on a four-stroke. The convention is to think of that as -360 to +360. (You could say -300 = +420. -300 isn't equal to +60!)
Chris, I still think you should take a look at your injector pulse times. That is the direct commanded output to the injectors. If the controller is telling them to fire it will show up there. That tells you if it is in the settings (probably) or in the hardware. You need to know which one it is before you go off trying to troubleshoot one or the other, because you just might be wasting your time on something that is working the way it is supposed to.
Jim
Jim. Pulse width is around 15.5 ms while cranking.
How confident are you that you have fuel pressure?
I have a bung welded in to my fuel rail for a schrader valve and have confirmed fuel pressure when the pump is priming and also when cranking using a fuel pressure gauge. I unscrewed the valve and fuel came pumping out when the pump was priming.
Can you put an LED across the injector outputs to give a visual indication of a pulse? 15ms should certainly be more than enough to run the engine at idle, usually 2-4ms is adequate.
Ideally you would scope the outputs to confirm the pulse width there. Barring that, a setup to fire off a single injector under pressure from the fuel pump would be the next best test. That can be done with scrap pieces of hose and such. An LED from B+ to the injector output can give you a visual indication, be sure to use the appropriate current limiting resistor and get the polarity right.
Did you ever put a timing light on the engine to confirm your timing advance? I know you got it to start but from here that still seems like an unknown.
Concentrating on fuel though, let's confirm fuel can get there provided the signal is output. Your injectors clicked, fair enough, but were they under pressure? Did they spray? Do they have an acceptable spray pattern? All unknowns at this point. 24lb injectors sound small for a 350ci engine, why are you using those? Are they used injectors? Have they been tested? Unless they are brand new out of the box, that's the first place I'd look for a fuel delivery problem after checking pump pressure. It isn't hard to hook an injector up to a efi fuel pump, drop the supply line in a jug of gasoline, jumper the pump and the injector to a battery and see if it sprays. (A fuel pressure regulator in the circuit is a good idea) And no more dangerous than spraying a can of aerosol hair spray provided you make sure your hoses can't come off. (You certainly want to test that at full pump pressure before powering up the injector) If your injectors are clogged no amount of signal will help.
Now, the test I suggested earlier was intended to test this, and I understand about having an engine that hasn't been run in yet. But this is a short test and personally I'd take the chance. The initial run-in is intended to give the cam the best chance of breaking in without wiping off a lobe. So far, you should still have the break-in lube on the cam. It isn't until the engine gets up to speed and runs a bit that it gets washed away. So to confirm that the injectors are doing their part of the job, I'd risk it.
Jim
Thanks Jim. I have to run off to work today but will mull over your advice once I get home again.
I checked timing last night while cranking the engine. It is approximately 5 degrees advanced.
The injectors are used and you have made a good point about testing them. I was assured they worked well but you never know.
The 24lb injectors were recommended by Jim Nichols.
Jim,
Would a test light work for that test for pulsing? I inserted a few strands of wire into each plug outlet, clipped the test light ground to one and touched the probe to the other wire. I then ran the injectors in test mode at 20ms. Nothing.
I am looking into getting the injectors bench tested but I think it is a software issue. I don't see where I can adjust pulsewidth in Tunerstudio but maybe I am just missing it.
Chris,
I may have missed it, but have you confirmed that you have 12V on one side of the injector lead with the power on? The other side of the injector goes to the controller which supplies the ground to activate the injector.
If you have a test LED or light, it would go across the two injector leads, not to ground.
Jack
Correct on the LED, and there are several ways to do that but across the injector is probably the best as that checks the entire circuit.
An incadescent bulb is supposed to be usable but I've never been that happy with the one I bought for testing injector outputs. Filaments take time to warm up and you might miss it. An LED is instant and uses very little current so you can see that when the incadescent is struggling to emit any light at all. Typical resistor values for an LED in an automotive application are in the range of 1000-1500 ohms. You don't have to be exact. Some LED indicator lights for auto use have the resistor already built in.
Your 15ms pulse shown on your digital dashboard during cranking argues loudly against a software problem. Sounds like hardware, either before the injector connection or after it. In other words, wiring or plumbing.
Jim
Jim, hit it. A regular test light won't light up fast enough to show an injector pulse, but an LED that gets a 15ms pulse will be easily seen.
Thanks guys.
I have the injectors out for bench testing including flow. I had one screw up in my fuel outputs; it was set at MS3x, not standard fuel. I am using the main relay board for my injector wiring, not the expansion card on my MS3x. The reason for this is...well...I am pretty ignorant of this whole EFI stuff and didn't even think about it. Now I need to see if I should run the injectors from the expansion board. I have asked DIYautotune techs for advice on what is best for semi-sequential.
I get the injectors back on Friday so I will try to find an LED to see if I can test for pulsing. The output fuel might just be the smoking gun for at least that part.
My second question concerns us finding 5 degrees advance during cranking. My engine builder and I were very careful when we installed the VR sensor bracket so I am surprised it is out that much, unless we had strobe light issues. If it is indeed out by 5 degrees can I advance it another 5 using the Trigger Wizard feature in the full version of Tuner Studio or do I need to physically move the sensor?
Thanks for all your patience. I am in over my head but there is no turning back so I really appreciate the advice!
Chris, the EDIS defaults to an internal setting for engine cranking, one of those not so well documented features. I do not recall what that setting is, but I don't know how you would change it. Anything between 0 and 10 should start most engines easily. More than 10 will tend to cause kickback and broken starters. 5 degrees is generally a good number, so is zero.
This is another of those things that MS3 will allow you to change if you decide to go to COP later. But it is a good system. The EDIS can operate fully independently from the MS, relying only on the VR sensor input. If that happens it reverts to 10 degrees of advance any time it does not have a SAW input. But I think it has a separate setting for cranking speed which is what you are seeing. That number may be different for different modules, I really don't know.
Once you get the engine running, you can unplug the SAW input line to the EDIS and your timing should go to 10 degrees (The factory included a plug for this purpose). This would confirm that your trigger wheel is installed correctly. There is an offset in MS you can use if it is needed, but don't try to set it at cranking speed unless you know the spec for your module.
It sounds like you've zeroed in on the problem though. MS3 is still pretty new to most of us. Mine is still in the box waiting to be installed after the new engine is running well on the old MS2. Getting the settings to accurately match the actual hardware configuration is one of the more important things you have to get right. You have taken some very big steps in your upgrade and done them very quickly, it is normal to expect a few hiccups before it all runs smoothly. I was not personally aware that you had to set MS3 for either the extender board or the relay board for the injector outputs but it stands to reason that you would since the extenter board has it's own connector to external wiring. Often it turns out to be a small thing like this. I usually bring the injector outputs directly off the main MS connector so I haven't had to deal with this issue up to now, but obviously will when I go to full sequential.
So with that setting corrected you can go ahead and test for outputs while you are waiting for your injectors to be tested. Get that going and provided they come back good you should be ready to fire it up and begin tuning.
Here is a link for an injector sizing calculator:
http://www.witchhunter.com/injectorcalc1.php
You can see that 24 lb injectors are typically good for up to about 300 Hp. However this is dependent on both rpm and fuel rail pressure, so an adjustable FPR can be a handy thing to have. Your injectors with a little bump in pressure should work out OK for your 350 provided you are running a mild cam and a stock redline. If either changes you will probably need bigger injectors. Especially the redline, remember the example where 25ms on time means the injectors go solid at 4800 rpm? At 7000 rpm that becomes about 17 ms so for the same horsepower output your injector size increased around 30% and that's more than you can safely compensate for with rail ressure in most cases. (Of course if your redline goes up that much you'll also be making considerably more than 300 hp.)This is why you see dual injectors on high rpm engines. One to run the engine at idle and low speed where a short pulse and well metered spray is required, and the second to dump fuel at high speed. However, those are pretty extreme cases, usually race engines.
Jim
Jim, His injectors are used in the 6.0 LS also and are closer to 25lb/hr at 50psi in the 5.3. Around 29lb/hr in the 6.0 recommended at 60 psi. Plus, running semi-sequential he would need half as much fuel twice as often. I think the size will be almost ideal. When we chose them we also liked the short length. I calculated up to 400HP at 85 percent duty cycle. Goldilocks, not too big, not too small, but just right. His cam may seem large in a smaller Rover 3.5 or 215, but in a 350 with 10 to 1 compression is just right for dynamic compression. I think it will be a much better combination than the Hot Rod mag build, once he gets the Megasquirt figured out.
Please forgive me if I am way off here, but I am trying to learn as much as possible before i start my Megasquirt project. I looked at the files that were sent, and find this line: <constant digits="1" name="Miss_ang" units="deg BTDC">80.0</constant>
I have an EDIS, and thought the degrees from the missing tooth to TDC was 50.
Thanks,
Dennis
Dennis, Good find. Probably a typo and everything helps.
Dennis, you are correct about the 50* (5 teeth). That is not as critical in MS as it is with MegaJolt but it still should be configured correctly as it makes everything else a little easier. Not sure what the 80.0 you are referring to is, but I haven't looked at the file yet either.
Jim, odds are good you are right, or at least that the FPR can be adjusted to compensate. I'm guessing Chris won't be going over 6 grand on his redline, so as long as his injector on time does not exceed 20 ms at that speed he should be in good shape. (actually 19ms, see below) He really won't know that though until the later stages of tuning when he's setting up full power. It's just something to watch out for when he gets to that stage. His redline may have to be set accordingly.
The number of squirts does factor in though since you have to consider the injector turn on time of about a millisecond and subtract that from your total available. For instance, firing twice with 20 milliseconds to use at 6K means you only have an actual 18 ms (or less) to inject fuel and that either cuts into your total fuel delivery or means a 2ms jump from 18ms to 20ms when the injectors are turned on all the time. I'm not sure if MS does that, though so it's safer to assume it does not. Typical injector turn-on time is usually at least 1.2ms and may be over 1.4, it varies by the injector design. Newer is generally better, peak-and-hold are definitely better but require special circuitry. A top notch injector test service should be able to tell Chris what his actual turn-on delay is. It's something you probably have to ask for.
Jim
Chris, are you still messing with this thing? Why not post a fresher log file? One thing I noticed in your old one was that your Throttle Position Sensor had never been calibrated. Only takes a second to correct that. Maybe you've fixed it already. I bet we can find more issues like that...
Hi Curtis,
Sorry, I took a break because it was getting to me. I am back on it.
I checked the coil pulg wiring and found a mix up so I fixed that. The car starts but runs rough. If you give it any throttle it immediately dies. AFR seems to be 10.3. The MS gauge shows that as well.
I don't know how to post the data log and tune on here but it is on the MS forum
http://www.msextra.com/forums/viewtopic.php?f=131&t=58126&p=436816#p436816
It's running rough because only #2 and 3 seem to be firing based on plug colour (black compared to almost new for the other ones). Why, I don't know.
Hi Chris,
You're probably firing on the other six plugs and not 2 and 3. If they are black then they are fuel fouled and won't fire.
Due to the vintage nature of my laptop I can't download your files. But I'll fix that shortly and get back to you.
Live like you mean it.
Fred
(http://i697.photobucket.com/albums/vv335/cgill/20150410_181155_zpsmmeoeudx.jpg) (http://s697.photobucket.com/user/cgill/media/20150410_181155_zpsmmeoeudx.jpg.html)
Two plugs for comparison...
An inductive timing light will tell you whether a plug is firing. Move it from lead to lead...
You can upload an MSL logfile here exactly the same way you upload a photo here. You might have to clip the beginning or ending off your log to get it down to the 600kb file size limit. The name of the logfile will appear in your message. If someone clicks on that file name, their browser will show the log in text form. No problem. If they save the file (to their desktop for example), they can open it with MegaLogViewer. Okay... that's kinda clunky, but at least the file can be stored here.
cgill-April10.msl
Hmmm...now to track down an inductive timing light!
If you just have two black plugs, check first to see if they are on the same coil. If they are, check the primary side of the coil for a signal and if that is there replace the coil.
Jim
Yes the two black plugs are on the same coil (#2 and 3).
That's my plan today. Thanks Jim.
So the second plug in that photo looks okay? I thought it would be darker given I have run the engine for two minutes. Sorry, this is all new to me. It's a great dose of reality and very humbling.
Chris
Chris,
It takes awhile for the plugs to "color in". With new plugs I would check to see if any are wet with fuel.
Also if you do have a bad coil as Jim mentioned, I would replace the 2 effected plugs as well. Sometimes once they have been fuel fouled they will not always fire correctly.
Bill
It's running finally. It was a plug order issue even though I checked it. I must be losing my mind.
It is running lean. I know I have to adjust the VE table but I am done messing with it. The plan is to get it towed to someone who has experience tuning about an hour away and get him to finish the tuning.
Thank you for all the advice and help!