It's not predictive.......
It's just a BIG DELAY from the previous cylinder.........
| Slant Six Forum https://slantsix.org/forum/ |
|
| Distributor Phasing. https://slantsix.org/forum/viewtopic.php?t=43289 |
Page 2 of 3 |
| Author: | Sam Powell [ Sat Jan 22, 2011 7:05 pm ] |
| Post subject: | |
Quote: I guess it must use the previous vane. I thought MS used a high number like 50 deg and then retarded timing, but my memory is murky since I don't have it running on a car yet.
Wow, So the crank timing would be off by 120 degrees. That's right there are only 3 pistons entering a firing stroke for each rotation of the crank. That would definitely keep it form starting. I do not know if I am right about my assumption though. I must learn more on this. Keep in mind that 60 deg of dist rotation is 120 deg of crank rotation, and timing specs are listed in crank rotation. So, 40 deg of (crank) advance is only 20 deg of dist advance, so that rotor/cap can handle that. Whenever I start an engine, I just assume I might have goofed on timing, and I just turn the dist around until it fires up, if it doesn't fire right away. Lou Sam |
|
| Author: | 65CrewCabPW [ Sat Jan 22, 2011 10:46 pm ] |
| Post subject: | |
Quote: How is it possible for the ECU to advance the timing? How can it predict when the reluctor will pass the pick up before it actually does so? It seems like it would be a simpler logic for the ECU to retard the timing. It could simply withhold the signal to the ignition module for a specified number of ms. The only thing I can figure is that it uses the previous reluctor vane for the next cylinder. Does anybody know about this? Sam As to how it can advance the timing, it DOES "predict" when the next spark should be. All the pickup from the distributor does... is tell MS the rpm of the engine and the engine's relative location. Period. Your engine can't actually start until you've passed enough timing events for the computer to ascertain the engine cranking speed, so it can extrapolate where and when to spark. Everything I've read about MS, says to me it uses predictive timing - yes, it actually predicts and fires ahead of time, in terms of thinking of the next ignition signal it gets - and in a slant six, you have three timing events per engine rotation to keep the rpm up to date, and be able to predict when the next spark should occur. Instructions tell you generally to set your distributor at 0, and use the software to advance your timing to where you want it. As to whether the timing event is properly phased to the rotor, that is determined by whether you use the rising or falling event as the trigger. MS documentation will tell you which, and that in some instances, it requires tuning a bit to work out right or be stable. EDIS, for instance, uses "past" events, the timing trails BEHIND the sensor, so that it doesn't use predictive, but measured "after the fact" ignition events that follow the trigger. The "#1 signal", which is the missing tooth, is not fired until after the wheel has rotated another 60 degrees (6 teeth) on an EDIS6. Your EDIS equipped motor cannot start until after a rotation where it learns the missing tooth on the 36 tooth wheel. This is nothing new in the automotive world. AFAIK, no factory computer predicts timing, it controls it "after the fact" by measuring events and calculating when to fire AFTER the "controlling" signal has passed. Only MS apparently uses software to predict future events. In reality, however, the two systems operate almost exactly the same, one "counting down after the event" to fire, one counting ahead to fire, but the logic is precisely the same, the only difference being the timing relationship between the trigger event and the actual "fire" done. I believe MS uses assumptions about cranking speed in order to calculate the timing when you start cranking, based on the sensor events, but it updates after you've gone past a couple of events to "real time" rpm. I could be all wet, but that's how I "think" it works. That's deduced from many conversations I've read about how to deal with ignition timing and the MS. Either way, your phasing should not be off, IF you have the proper signal event chosen as the trigger event, that being either the rise of the voltage spike, or the fall after it spikes. I could probably find it by searching the documentation, but I suspect you know it better than I do. Each specific sensor type produces a signal or multiple signals, which must be analyzed digitally. From what I understand, the Chrysler reluctor based system has a variable voltage depending on speed, and the actual predictable event is not the rise of the voltage, but when the spike starts to fall. Please don't take this as gospel, but that's my analysis of the waveform and how to get a consistent reading from it. |
|
| Author: | emsvitil [ Sat Jan 22, 2011 10:56 pm ] |
| Post subject: | |
I think most ECUs don't predict timing during startup. They fire on the ignition signal. Once started, they'll predict the timing. Firing on the ignition signal is also used as a 'limp home' mode. |
|
| Author: | 65CrewCabPW [ Sun Jan 23, 2011 12:24 am ] |
| Post subject: | |
Quote: I think most ECUs don't predict timing during startup.
That's certainly true on some. This fact is illustrated by why most new motors do not start instantly. My 93 Dodge w/Cummins will start with just cranking it over the first full compression stroke. Even cold, at 20 degrees, that's often true. My 78 W/Slant six often did when warm, as well. But none of my modern, sequential injection engines will start anywhere near that fast. Often it takes 2 or 3 full rotations before the computer knows exactly where the motor is at, and can inject and fire meaningfully. An EDIS based system can't start and will not fire until it "sees" the missing tooth go by, so it knows which coil to fire.
They fire on the ignition signal. Once started, they'll predict the timing. Firing on the ignition signal is also used as a 'limp home' mode. |
|
| Author: | Sam Powell [ Sun Jan 23, 2011 6:24 am ] |
| Post subject: | |
Very interesting post. For what it is worth, I feel it is healthier for the engine to fire after a few rotations. This gets a little oil pumped into the bearings. Someone will likely tell me I am wrong about this, but I still prefer a few cranks prior to firing. Peter did respond to an e-mail I sent about this and said yes, MS predicts timing based on the previous VR vane. But if, as Em's says, it starts on the ignition event, this would make it try to start 120 degrees out of time. So MS must be predicting something even in start mode. So Crewcab, your description of things, which was very helpful by the way, still leads me back to a central question, How do you tell if your VR wiring is producing a a rising a falling signal? I know it has both always, but wired one way, one is sharp and wired the other way the other is sharp. In each case one is vague and slow in either its rise or decent. Sam |
|
| Author: | Wizard [ Sun Jan 23, 2011 10:56 am ] |
| Post subject: | No no not predicting!! |
Guys, ECU do not predict! The trigger event is set specifically to the static timing/base timing (remember you had to time the engine with vacuum plugged or unplugged manifold pressure and plugged that hose during setting up static timing?) but yet the rotor (if using distributor) have sufficient overlap with the contact in the distributor cap. Every trigger event takes certain of time for computer to calculate BASED on rpm and manifold pressure also engine's temp. Think about it, computer cannot do this predicting. Cheers, Wizard |
|
| Author: | Sam Powell [ Sun Jan 23, 2011 4:10 pm ] |
| Post subject: | |
The ecu sets the timing for cylinder 1 from the VR vane before the one that coincides with when the rotor is under the number 1 cap button. Of course you are right when you say the timing is determined by the timing map, which takes all those things you mentioned into account. But, the precision of when it happens is all set up by the previous VR bypass event. Then it knows precisely when to fire the ignition, at some predetermined time before the number one VR vane crosses the pickup. Sine this is Peter's professional field, I trust his understanding of this. And, I think I am relaying his information accurately. Sam |
|
| Author: | Wizard [ Sun Jan 23, 2011 4:26 pm ] |
| Post subject: | |
Powell, The main issue is terminology confusion. ask him what does Peter meant on several certain items and clear them up. Make common terms and make this clear then I hope you would see we are talking about same thing. Cheers, Wizard |
|
| Author: | 65CrewCabPW [ Sun Jan 23, 2011 9:44 pm ] |
| Post subject: | Re: No no not predicting!! |
Quote: Guys, ECU do not predict!
I tried to explain this... The EDIS system fires AFTER the DESIGNATED NUMBER ONE timing event is 60 degrees past, fires the number one coil. It uses the continuing passing of the teeth on the wheel to be able to "guess" at where the engine is via knowing how long it takes to get from tooth to tooth. The trigger event is set specifically to the static timing/base timing (remember you had to time the engine with vacuum plugged or unplugged manifold pressure and plugged that hose during setting up static timing?) but yet the rotor (if using distributor) have sufficient overlap with the contact in the distributor cap. Every trigger event takes certain of time for computer to calculate BASED on rpm and manifold pressure also engine's temp. Think about it, computer cannot do this predicting. Cheers, Wizard In that methodology, ok, the event that triggers firing number 1 cylinder is long past. In a megasquirt, the magnetic point passing the reluctor for 1 happens AFTER the firing for that cylinder has already happened. It was predicted by prior events, but since the MS has no idea what cylinder is firing, it is merely using the signal as a reference point, and extrapolating (predicting) engine position. I hope I was clear that the logic - the mechanism by which the systems knows when to fire is the same - but only our perception, which is based upon knowing that the actual firing event for said cylinder occurs before the reference point signal happens. Sam: Does this help? http://www.diyautotune.com/tech_article ... engine.htm I believe the part(s) of this that are causing you difficulty is here: Megatune Ignition Configuration: Trigger offset = 10� (this will vary, depending on the distributor orientation, see notes at the end of the article) Ignition Input Capture to 'Falling Edge' Cranking Trigger to 'Trigger Rise' Coil Charging Scheme to 'Standard Coil Charging' Spark Output to 'Going High (Inverted)' You will need to set the dwell to match the particular coil you are using. You're talking about trigger offset and whether to use rising or falling, etc. These are the closest things to official answers I can find. Also, the trigger offset is not "set in stone". Rather, you will adjust it to achieve the required timing behavior. Your "trigger offset" is the cultprit, as far as having distributor phasing off, but in this case, your phasing is is off by a few degrees, not dramatically. |
|
| Author: | 65CrewCabPW [ Sun Jan 23, 2011 9:59 pm ] |
| Post subject: | |
Ok, now let me explain a NON PREDICTIVE system. MSD makes a "timing controller" box that lets you very crudely construct a timing curve for your engine, using a programmable rpm only curve, and then a vacuum sensing timing adjuster. In the case of these devices, your distributor is set to a specific timing setting equal to the highest advance you will have. The timing control devices then DELAY the spark event to achieve timing control. This is a 'real time delay' mechanism. Sam was asking if that's how megasquirt knows when to fire. The answer is no. The predictive mechanism is often used in Europe, by CANBUS control systems, where a small digital circuit is put inline, which can be used to "adjust" timing of sensor outputs, to fool the OEM computer. Yes, one of them can "advance" timing, but it works by obtaining some signals, then extrapolating how long before the next expected timing signal to send the "modified" timing event. It works by calculating how long to WAIT after the timing event before the one in question. ALL of them operate by extrapolating engine position by means of clocks and timing events. Whether we call it predictive is merely how we view the effect of trigger sequence in question vs what the device causes to happen. |
|
| Author: | emsvitil [ Sun Jan 23, 2011 11:38 pm ] |
| Post subject: | |
It's not predictive....... It's just a BIG DELAY from the previous cylinder.........
|
|
| Author: | Sam Powell [ Mon Jan 24, 2011 5:31 am ] |
| Post subject: | |
Crewcab, I had that linked article printed out and on the bench, and had followed every step and every set up instructions to the letter. Things were wired right, and every configuration was set properly in tuner studio. And it would not start. This is why I said it makes me feel really dumb when I am told to simply" read the Manual" (in this case an internet article). And I mean depressingly dumb, when you have been over and over it, and people still talk as if the answer is as simple as reading the manual. Just read the manual. At that point, I have read the manual, asked on the internet what is wrong, and am told "read the manual". By then my only recourse is another professional, several of whom have failed at the task, and then finally, find another hobby. Something was going on there that has escaped everyone so far. Especially me. And, it is probably something fairly simple. But, it is not in the manual. One option is to accept the fact tht I have wasted a ton of money on the MS II stuff, and just suck it up and drive the car with MS I. But, it could work better with the MS II, and optimism eventually rears its head again. I suspect some of the argumentative nature of this discussion could be laid to semantics as Ems slyly pointed out. However, semantic subtleties can lead to misunderstandings, especially when reading the printed word. And this is where this ignition things can get really tricky(read "confusing"). Controlling fuel in a batchfire system is a piece of cake compared to controlling an ignition system with an ECU. I think you need someone who has "been there, done that" working with you to trouble shoot when things don't work. I really appreciate the contributions to this subject. You never know what will trigger the light bulb to go on,. Sam |
|
| Author: | Wizard [ Mon Jan 24, 2011 9:17 am ] |
| Post subject: | |
My apologies. Could you detail what you have done to set up distributor and details in the megaquirt program set up? Cheers, Wizard |
|
| Author: | Sam Powell [ Mon Jan 24, 2011 7:35 pm ] |
| Post subject: | |
Apologies accepted.  That was just to let you know how frustrating this has been. And why. Wizard, I set the ignition up just as described in the manual, and in that article. At first I set it up to run off the MSD box, using the manual's wiring diagram, and ignition settings. VR pick up to tach and ground, and the MSD box white wire to s-5. The msd box ran fine with the VR pickup, and it fired when the white wire was grounded. It passed those tests. There was a tach signal reaching the computer screen. When that did not start, I changed the tuning page settings and the wiring as described in the link Crewcab linked us to and tried it that way. Still no start. Things really went to hell when I switched the coil setting without unhooking the coil. I know I was told not to do that. That is why I am sending it back for warranty work. I pretty much take responsibility for any real damage they find. Maybe just the firmware got scrambled. Sam |
|
| Author: | 65CrewCabPW [ Mon Jan 24, 2011 9:41 pm ] |
| Post subject: | |
Quote:
I really appreciate the contributions to this subject. You never know what will trigger the light bulb to go on,.
That's all I'm trying to do here, Sam. Seeing if I can jog something. Sam I've been in that kind of situation MANY times over computerized stuff, networking, software, scripting, code, you name it. Sometimes people come up with exactly what I've done, we look at it and indeed, nobody sees anything different. Other times, I've found some tiny stupid thing I overlooked, just in walking someone ELSE through, since I then have to explain it all in detail, and I catch my own inconsistency. I know exactly the frustration you feel, as I've done the equivalent you have, and had things not work. Having followed the written directions, double and triple checked and see nothing, others look at it and see nothing. I have repeatedly, however, found my own errors just by trying to walk someone else through what I did. Or someone points something out and says "what's that mean?" and like you said, the light bulbs go off. I don't mean to seem like I'm belittling or in any way suggesting you're not capable, nor anything fo the kind, just intrigued at how you're having an issue with something that SHOULD be relatively straight forward. At this point, I, too, would probably defer to someone who has done a bunch before and see if they can spot whatever the heck it is that's busting your butt. If I were in your neighborhood, I'd head over see if I could help you out. It's just the neighborly thing to do. At this moment, maybe we could all just take a up a collection to pay overnight shipping to his choice and back and see if that helps! 
|
|
| Page 2 of 3 | All times are UTC-08:00 |
| Powered by phpBB® Forum Software © phpBB Limited https://www.phpbb.com/ |
|