Because I haven't really looked into sequential I'm not 100% familar with the "best" tuning practices on it. My logic was for WOT, but I can definitely see what you are getting at for low rpm. It makes sense. My guess is that logic has been programmed into MS3 already but I honestly don't know. With my "alternating" firing set up there is definitely fuel sitting by the intake valve when it opens up, because the injector will fire at least 4 times before the next intake cycle and the WBO2 averaging the cylinders show it being pretty lean, it would have to be a relatively small quatintity of fuel, and I think the real gains would be in the AFR balancing, which get's back to checking each cylinder, one's running richer and other's running leaner to make a "normal average" my WBO2 is picking up. My off the shelf injectors aren't flow balanced, and intake tracks have all sorts of inherent imbalances on air flow, plus slight differences in ring seal for pulling air, etc to throw a couple wild card variables in there I have no way of quantifying each with out individual exhaust gas monitoring. It's just $$$$ versus benefit, the increment might be there, just "likely" too small for me.
Let me throw some real world facts at you guys. When it comes to sequential, you've got it a bit backwards Silver buick. You should be more concerned about cyl to cyl imbalance with batch fire not sequential. Here's why, and you touched on it. If your cyl to cyl flows, compression etc are not balanced then who's to say you'll want them all the same air/fuel ratio anyway? Maybe 1 cyl runs hotter than the rest because of slightly higher comp ratio or unequal cooling etc. Ultimately you should run that cyl fatter to deter detonation as you're governed on how much advance you can use by the cylinder that's the most detonation prone. On the other side of the coin that will be harder to tune out with batch because unavoidably your dumping fuel into some cylinders at exactly the wrong time. And peak power is just one aspect of sequential. I worked with the Univ of No. Carolina Charlotte on a back to back test of a well known sequential system vs a Holley 950 batch fire. In even cyl to cyl mode it always had 2 cylinders that ran richer and 2 cylinders that ran leaner than the rest. I'm sure you can figure out why, you're firing at least a few injectors at again, exactly the wrong time. And don't think that won't change with rpm and load, it will. That's why you can't just dial in more fuel overall and expect it to be right. When they did the back to back the sequential only made about 4-5 hp more at peak, but a whole bunch more in the midrange, to the tune of 25-30, on a 300 hp motor. That's a bunch, and that's what you feel cuz that's where you drive it. When I tune injector timing, I look for a change in O2 correction. Normally when you hit the right area in injector timing you'll see a minor change in positive correction but moreso you'll see a noticeable reduction in transient fueling needs. Conversly if it's the wrong timing it will show up bigtime. I'll reserve my opinion of the MS system, I know it's cheap and probably has a niche. But right now you're giving up something for that cheap price aren't you? And more importantly something that effects drivability. And that general notion that assembling it yourself helps you learn EFI doesn't hold water either. I'm not saying you subscribe to that but as you know that is a common argument for the system. But did you actually have to build your house in order to understand how it stays up? Probably not. And conversly does that nice women in China automatically understand how a tv works just because she's screwing circuit boards together? Probably not huh. Either way actually building something doesn't automatically make you understand it. If you have a chance to run sequential do it, if for nothing else than the learning curve. Jmo.
I'm pretty sure that's what I was getting at, and more to the point the inability to measure and quantify the imbalances even if I did go sequential. So if I can't quantify the imbalances how in the world can I correct them with sequential injection? With Carbs, TBI's and with Batch fire you are ALWAYS tuning to the leanest cylinder if you are running a safe tune. And unless you have a way to measure each cylinders output then even with sequential you are tuning to the leanest cylinder, hence my argument for wanting an O2 sensor or EGT sensor on each exhaust port if I do sequential. And I highly doubt that you've reserved your opinion on MegaSquirt. You've been pretty open about your dislike of it in several threads. I'm not even going to debate YOUR mis-conceptions of any of the megasquirts. Looking it up, it'll cost me around $350 to go sequential on my Buick from the setup I have now. That's ordering the MegaSquirt 3 chip, the expansion board and $60 in miscellaneous parts (which there may not be any as I have a ton of wire on hand).
Most certainly. I will be going Sequential on the Pontiac OHC six I have, and will update the thread on the "Other Bench" when I make real progress on it. For now I'm working on putting together a multiport intake manifold for it.
Good stuff guys. I've run MS for quite a few years now on multiple vehicles and feel it's pretty good. On this latest build I'm planning on upgrading in a graduated manner. Initially I'll keep the crank trigger EDIS, run batch, and plug the individual egt ports in the headers and just rely on the single WB-O2 and collector mounted egt gage I was using with the smaller engine to get dialed in, but I'll have a cam position sensor ready to go. I'm still a couple months away from having to decide exactly how I want to proceed. Anyway the egt probes are around fifty bucks each or more so those will come after the board mods, and those will be pretty extensive. Not only do I plan to run sequential with the egts, I'll be adding megashift control of the tranny and this time around it's gonna get kinda involved. That's probably what I like most about the MS, it isn't a one size fits all solution. I don't have programming skills, but it's very clear that if I did there is a great deal that I could do with that, and at some point it even may make sense to bone up on it. But really that's a struggle I'd just as soon avoid. Anyway, since you guys are messing with this stuff I thought I'd say hello. I am too, or more correctly will be soon. Again. It makes ya a little crazy but is so worth it. Incidentally, I'm guessing I'll probably end up making a new cam sensor/oil pump driver (again) but I saw a new GN piece for about $139. Is that pretty much the going price or are there any deals out there? Jim
Nice. On my OHC L6 turbo build with the MS3, if I grenade the 2004r transmission I'm putting behind it I'll likely be stepping up to an electronic shift transmission using the MegaShift. I just happen to have a 2004r in the basement so I'm going to use it. I've heard EGT's can be a double edged sword, they'll let you know something is different in the running of a cylinder but not necessarily if it's lean, rich or the timing is fubar'd. Using the WBO2 average down stream though should help. I'm thinking I'll dyno tune my OHC Six with 6 WBO2's, then install EGT sensors to monitor it after it's in the car and running for any small trim corrections. At least that's the current plan, may not go that way, we'll see =P That $139 is a good price for a new cam sensor. Scouring e-bay used ones occasionally pop up for ~$50. I didn't see any checking tonight though.
What I learned about egt on the last engine was that it is very useful for tweaking timing and fuel maps. IIRC (it's been awhile) as the engine leans out it shows up first on the egt and then the water temp starts to follow it up. Pretty much the same I think on the timing. But it's been several years since I tuned that engine for power and cruise and I'll probably remember the finer details when I get there on this one. I particularly recall a long cruise and a very long hill being really helpful to get it dialed in. That was with just a single dashboard egt gauge though. I've been told the autotune on the MS3 is pretty good so I might see what that does once I'm at that stage. There is still quite a bit to spend money on before I get to the probes, and more than anything else I just want to drive the car. Even batch fire, in poor tune, power will be very good, but I'd like to get the mileage up over 20. Jim
I made some recent changes to my Thunderbird, and rather than manually re-tune it I've kicked on the Autotune function of the MS1/extra and it works pretty dang good once I opened up the range it was allowed to tweak. Cleaned up the drivability quickly from a "from scratch" tune. MS3's I'm sure works as well or better.
Those 'autotune" setups sound interesting, wonder how well they work. I'm sure individual cylinder tuning of a sequential setup isn't included. And I don't thing the ignition timing is "auto tune" either. I tried sequential on my 79 Eldo, because it already had port injection, and the original batch triggers could be used as a cam sensor. What are the merits of a completely SEPARATE programed ignition (like EDIS) and some EFI? Bruce Roe
I'll agree with you there, I don't know if any systems autotune the timing. I don't know what the feed back to the ECU could be that would be reliable. Knock sensors only work if the valve train is wisper quiet and most people don't like drilling holes in their blocks (I have, but have valve train noise issues I think). As for a separate programed ignition setup, this is my take. When people compare cost on carbs and EFI, it's an apples and orange comparison because with the EFI cost usually comes some pretty good ignition control on top of fuel. I don't know what MS3's timing controls look like, but as far as I know some separate MSD box's let you do individual cylinder settings for advance. One guy I know was looking at that to cruch a hurt cylinder when racing, being able to pull some timing out of a cylinder that either has rings, headgasket or rod bearings going on it so he could make the next round of racing, and MSD sounded like the ticket item. But if you aren't looking for individual cylinder trims, I don't see any advantage. It also goes back to my arguement of how do you measure and quantify needing individual cylinder timing trims? From what I gather combustion chamber pressure sensors are pretty much what are required if the engine is healthy.
We had questions concerning how to measure the work each individual cylinder was doing and if it could be equalized by individual spark timing and/or fuel. The answer was to pack up all of our stuff and head to the nearest 3000hp inertia dyno which was in Calgary, Canada. An inertia dyno measures torque by measuring the amount of acceleration of a cylinder or "drum" that has a known inertia. It is similar to a chassis dyno except the drum is driven directly by the engine. The beauty of it is, you can look at the acceleration every 90 degrees and see how much torque each individual cylinder is applying to the drum. The final printout shows the difference or "scatter" between each cylinder. Then both the individual fuel and timing can be worked with to get a tight grouping on EGTs and scatter. Now if it sounds like changing both timing and fuel to accomplish this would drive you crazy, you would be right. Taking this down a level, if the EFI system is receiving a cam sync pulse along with an accurate crank trigger signal, then it has the ability to examine the exact time interval between each 90 degree rotation of the crank. When under load the system could then manage individual cylinder tuning similar to tuning on an inertia dyno. If detonation occurred it would be seen by the sudden change in torque between cylinders and could be compensated for before the next compression stroke. Just some more stuff for thought. Paul
The ECU would need to be able to quantify that load. An inertia dyno works well due to it being a calibrated mass. Today's high resolution crank triggers coupled with the cam sensor can already quantify which cylinder is mis-firing based on the slow down of the crank in that range of rotation, so that technology is here, it's just left up to quantifying the load. The ECU doesn't know if the wind has shifted and lessened the load on the cylinder causing the acceration or if the cylinder's efficiency has increased. After lots and lots of data gathering and statistics a computer might be able to sort out the background noise from cylinder efficiency (for both fuel and spark), but I'm not qualified to say how much or little data that would require while driving the streets instead of a steady state calibrated load.
I was looking at the relative output from cylinder to cylinder within any 2 crankshaft revolutions as long as the throttle position wasn't changing within the time frame of the sample. No need to quantify if it is just a matter of getting the cylinder outputs even. Changes such as wind or road grade are not likely to effect the load sampling within such a short time period. With thousands of 2 revolution samples each minute there should be no problem figuring which cylinders are operating above and below the average. Anyway if you still don't think there is merit in this, suppose a way of actually measuring the engine output torque was added to the system. This is just an example. http://www.aemelectronics.com/dyno-...no-shaft-on-vehicle-dyno-sportsman-series-74/ Paul
That's a beautiful piece of high tech instrumentation. I'm sure it has its place. However, Paul, as you noted, simply monitoring the variations should be adequate. At 2000 rpm that's 33 engine rotations a second or 30 milliseconds each, about 7 milliseconds for each 90* rotation. Any modern processor can easily keep track of that time interval and do comparisons to other 90* phases. Plus, most of our driving transitions are in seconds, not milliseconds so even driver inputs, wind, or road conditions are not much of a hurdle. So the pieces are there. For someone who was handy with this sort of programming, it's a wide open arena for development and could very quickly explode into an entire new market segment. Too bad I'm not a programmer! Jim
Interesting thought on using the short sample time interval for the statistics. I hope you don't mind, but I asked on the MegaSquirt forums if this cylinder to cylinder time between ignition events could be measured off the crank sensor, and if possible tied to cylinder or crankshaft degrees with a cam sensor. I paraphrased your above/below average line. The ignition calculations try and predict when the next tooth will show up, so that I would think it could have a value recorded of when it actually did show up. This is an MS2 setup, the MS3 might have more resolution since it can handle 8 separate fuel and spark points. http://www.megamanual.com/ms2/wheel.htm Doing reading on the sequential part (while looking this up), it looks like MS3 programming starts injecting fuel when the exhaust valve closes (during overlap so the intake is open already) until the point it reaches the timing of when the intake valve closes, until the point the fuel demand is high enough that the injector has to spray longer than the intake valve is open for and then the timing locks with the closing of the intake valve and the fueling time is then added to before the exhaust valve closes but intake is open, then of course as more fuel is required it starts before the intake valve opens. Interesting.
No problem involving other forums. The more heads together on a project the better. Yes, when the engine rpm is changing (Acceleration) the calculations have to predict the arrival of the next tooth based on the rate change of the time between the previous teeth. Knowing the rate of change, the calculations can not only predict the time for the next tooth but the time for any interval in between teeth down to a tenth of a degree. If the software (Read Software Writer) wants to take this a step further, during steady state conditions (No acceleration) the time between each tooth can be examined and if the tooth separation or spacing is not equal, the error for each tooth can be calculated and used in the calculations for even higher resolution for timing events. Now you can really zoom in on the cylinder to cylinder time with more accuracy. The object here is not to spray any fuel out the exhaust during low rpm, low load conditions like during highway cruising when gas mileage is the main consideration. This is one of those areas (Valve overlap) where EFI gives you the control a carburetor cannot. With this advantage you can use a larger overlap performance camshaft and not suffer a big gas mileage penalty. It would be better during high fuel demand, if the fuel injection period didn't include any valve overlap time. Using a larger injector would decrease the injection period for WOT but as the injector size increases, so does its ability to inject a linear or consistent short squirt for idle and low load conditions. One solution is to run 2 injectors per port and phase them so when the first injector gets to 40% duty cycle, the duty cycle drops back to 20% and the second injector is put into play. Now the injection time is short enough to injected fuel after the exhaust valve closes and only when the intake valve is open for most of the engine load. Plan "B" could be a linear increase in fuel pressure based on load. Of coarse Plan "C" is go to Direct Cylinder Injection. Paul
Interesting thought. Reading the MS3, it's not quite up to par to do that on it's own, though it is very possible (easy?) to use a CAN link between two MS3's and pull that off with each one controlling a set of injectors. One MSIII will probably be able to do it for a 4 cylinder, and I wouldn't doubt if sometime in the future the programming is adapted to have the spark outputs converted to fuel for 16 fuel outputs (staged or otherwise) and a single spark output (for a standard distributor). An assembled yourself set of MS3's and expansion chips would be $926, or pre-assembled $1,280.
We ran two injectors per port with our twin 88mm turbo setup using gas and three 160 lb injectors per port when using methanol. The reason here for staged injection was to have better fuel control for idle and during the time the engine was not under boost when the fuel requirements where low. We staged the second injectors to come in at 50% duty cycle. The BigStuff3 was the only system at the time that could handle 24 injectors and be able to stage them. Now there are several systems that can handle multiple port injectors with the ability to stage. Paul
The spark outputs of the MSIII are logic level, so drivers would have to be added if you wanted to run injectors with them, and you'd have to get someone to modify the code for you but I suppose it could be done. For my part, I'm trying to decide between 36 and 42 lb injectors for a blown 340 so I don't think staged injectors are in the picture here. I doubt the output will be much over 500 but it could creep up there a bit with the planned 7K redline. Won't really know until I get there. Paul, you have some really good thoughts on this stuff. We really should have a code writer working on it. I've rubbed shoulders with a few pretty good ones but unfortunately I've come away with the impression that our thought processes are very different. I've not learned the key to turn which brings them to concentrate on the task, and therefore am helpless to get the ball rolling. Along with which, I've developed a secondary impression that they are possessed of simply enormous egos, or alternatively have an impenetrable defense mechanism. In either case, getting them to listen to good real world suggestions has not been a pleasant experience for me. How to you present the world's greatest idea ever to someone who already knows everything and is going to do it his way regardless? Only, I'm just guessing here, if you control his paycheck, which I obviously do not do. And even then only maybe. I wish my experiences had been better. But wishes don't help. Jim
