The MegaSquirt Project has experienced explosive growth other the years, with hundreds of new MS installations occurring every week - a phenomenal success! MegaSquirt has been successfully used in all aspects of Internal Combustion engine applications including R&D, Industry, Race, and Research. The MS project has transformed itself from a simple R&D project into a full-featured mature engine control system. To reflect this the support structure has also changed to meet the needs of MegaSquirt Users.
Moving forward, the R&D forums for MegaSquirt project are in a read-only mode - no new forum posts are accepted.
However the forums will remain available for view, they still contain a wealth of information on how MegaSquirt works, how it is installed and used. Feel free to search the forums for information, facts, and overview.While the R&D forum traffic has slowed in recent years, this is not at all a reflection of Megasquirt users, which continue to grow year after year. What has changed is that the method of MegaSquirt support today has rapidly moved to Facebook, this is where the vast majority of interaction is happening now. For those not on Facebook the msextra forums is another place for product support. Finally, for product selection assistance, all of the MegaSquirt vendors are there to help you select a system, along with all of the required pieces to make it complete.
Forum rules
Read the manual to see if your question is answered there before posting. If you have questions about MS1/Extra or MS2/Extra or other non-B&G code configuration or tuning, please post them at http://www.msextra.com The full forum rules are here: Forum Rules, be sure to read them all regularly.
TheMonkey wrote:Bruce- any progress on getting your wired up to your motor?
I have everything ready to go - I had a custom set of plug wires made up by Magnecor - they have the special LS2 coil terminals and can make up special plug wire setups.
Now I just need time to test it out, I am targeting the weekend if I can peel off the time to finish the install..
Last night I methodically went thru all the Delay Teeth from 0 to 20 using your 6 cylinder setup. Everything worked fine except when I set Delay Teeth = 18 or 19, meaning the tach pulse comes in on tooth 19 or 20, I got the phase shifting you described on both spark and fuel. In fact some cylinders gave no output and others just kept phase shifting. Also there was no resynching or missing/extra teeth detected. These 2 teeth mean that a tach signal also comes in on tooth 39 or 40 and tooth 59 or 60 - which are missing teeth. And that is the problem - you can not have Delay Teeth such that this happens. The reason is simple - you never see the tach pulse and everything gets off by one cylinder insider the router processor. There is no fix for this, because many calculations and flag settings occur on each tach tooth and you have to have a real tooth to trigger it. What I can do is figure out a test for when this happens and trigger a resynch. This will happen over and over until Delay Teeth is set properly.
While your finding sounds related, this does not explain some of the other settings I had problems with. For instance.... when I had 29 delay teeth, -90° offset, that was really bad, and it was RPM specific at 2300 revs. This would have put tach pulses at 9, 29, & 49. Something else still happening.... hopefully problem will be re-created when you guys install it to help pin it down.
I have been having loads of fun and success driving this car with the new ignition. Just as a side-note... I checked TDC reference w/ offset wizard and it did not require any additional offset.
Still need to configure for the cam sync, I think I'll have some time tomorrow.
I will check your specific configuration. This sort of thing should show up on the bench as well as on a car. Car testing will wring out any noise issues.
Last night I went through all the trigger offset settings from 0 deg to -260 deg. All it did was move the spark to the left (advance) more and more. When it went over roughly 200 deg, the dwell started being reduced, as it should because dwell + spark advance must fit inside 2 tach cycles which is 240 deg for a 6 cylinder. By -260 deg the dwell was 10 us, but nothing fishy happened with fuel or ignition.
When I was able to recreate the problem on the stim, it would only do it as I passed through the RPM 'trouble zone'. Not only that, but it would only do it while RPM was being changed, and it would only do it when using the 'coarse' pot, and not with the 'fine' pot. Perhaps your stim has better soldering? Actually I'm sure it does Anyhow, it seems the RPM changes on the coarse pot (even when trying to do it slow and smooth on the stim) was emulating whatever was happening on the car, because the symptoms were the same.
I think it's peculiar that I have found a setting that works by using the very first tooth available after the missing gap tach pulse (as you noted 18, 19 delay teeth won't work, but I'm using 20).
it's running on sequential now. in the driveway, i hooked up the biased VR- as Bruce described. i tapped into TPS to do it, so i'll need to clean it up to take it for a spin.
transition to sequential went super lean. similar routine on my Pontiac, to find proper opening time (or technically 'net open time') for smooth transition. i had to dial back open time from 0.80ms to 0.15ms.
adjusting the open time requires an adjustment to the entire VE map. i thought i could just do the math....
i know my fuel map worked great on wasted (2 squirts per cycle). if i reduced my open time by 0.65ms, then i need to increase every map point by 1.3ms. if my req fuel # is 19.8, then shouldn't i just be able to add 7% to each VE bin? i tried it, and the idle oscillates wildly.
i think it is possible that by having such a small open time, that any changes in VE bins become highly levered? for instance, if my preferred wasted pulse width is 2.2ms, and i have 0.8ms open time, then changes in VE are only applied to to the 1.4ms while the 0.8ms open time remains fixed. when i swap to 0.15ms open time, then the changes in VE is applied to a full 2.05ms making it more sensitive. anyhow, tough to explain.
but the point is, it sequences now, transitions at same AFR (after open time adjustment), and i'm going to have to retune the fuel map. hopefully have road report soon.
attached is a datalog of the transition from wasted to sequential. note that there is some kind of lean hiccup, but smooths back out to same AFR. ignore TPS... it was unplugged for this.
i've been running some errands in sequential. seems to work fine. still have some tuning to do. seems like i just finished tuning the wasted map to be real nice. it will be interesting to see if the open time that makes idle fungible between wasted / seq will work through the whole map. i think i'm going to setup a small LED light for an output to be on in wasted, out in seq (just so i know when laptop is not hooked up).
i'm pretty much done with install. i'll start keeping track of mileage and any changes i make or quirks i run into. as a matter of reference, i put several hundred miles on the car in wasted mode after converting ignition.
I did chop down my accel enrichments in half to compensate for 1 squirt vs. 2 squirts (oops - meant to say double). will this adjustment be built into code in future?
Last edited by TheMonkey on Fri Jul 17, 2009 7:03 pm, edited 1 time in total.
Given a couple of difficulties from wasted (at 2 squirts per cycle) to sequential (at 1 squirt per cycle), could you setup wasted to run on 1 squirt per cycle? Once setup, the sequencer will run in wasted only a very short period of time after starting before it gets cam synch and it switches to sequential. Will that give you a "one-tune" solution that eliminates problems induced by opening time inaccuracies and acceleration enrichment -- and anything else that is impacted by the switch from 2 squirts to 1 squirt per cycle?
My thoughts are that the opening time should reflect actual 'net open time' as it does now. I just put too much tuning effort in while it was wasted only. Only way to dial it in is to check idle iterating using multiples of squirts.
If accel enrichments get automatically adjusted, I think that's the only thing leftover. There are 'cranking PWs', but generally those happen before it syncs with cam, so it's always wasted cranking. I think Al already noted that he was going to adjust the accel into the code before production.
The lean hiccup on transition is weird though. That did not happen during transition on my Pontiac, but I'd have to bring up old datalogs to confirm. Could potentially be related to that car having a positive offset, and this car having negative?
I will work on modifying the accel so it works as it should in wasted or sequential. But as far as the pw, the code does and always has corrected for this. If you are in wasted spark mode, but you selected a sequential dual spark mode - which is what you did, then the code calculates pw as if you were in sequential, but before it is applied it sets pw_wasted = topen + (pw_seqtl - topen) / 2
The first term on the right is open time. This should be the same in wasted or sequential, unless you can open injectors quicker in sequential because it is only firing one injector at a time instead of 2. I doubt this would show up on a datalog, since the drop in voltage would only occur for a ms or so. But it would be interesting if someone scoped voltage when firing one and firing 2 injectors from the same battery.
The second term on the right is the time during which fuel is actually being injected. For wasted the amount of time during which fuel is injected is cut in half, because there are 2 simultaneous squirts.
There is also some stuff going on with x-tau and changing wall wetting characteristics during the transition, but I don't believe you are using x-tau.
So this should work and you should see a smooth transition without changing the topen time. Now the second term is an approximation because in reality you have a breakpoint time in which no fuel flows (equivalent to trying to move a heavy piece of furniture and nothing happens for a second or so, then it gives way and moves), then a ramp up time, then a flat rate squirt time, then a ramp down time. All of these factors are being put into the model based code, but we will need data for how much fuel actually flows during the ramps and how long they last.
Anyhow, it seems the RPM changes on the coarse pot (even when trying to do it slow and smooth on the stim) was emulating whatever was happening on the car, because the symptoms were the same.