Hmm. Perhaps it was time for the 30yr old logic pulser to die. One too many abuses over time. Probably kickback voltage along the ground path. I wish I had more time these days. Don't know how you do it during the holidays, but this is good info.
Hmm. Perhaps it was time for the 30yr old logic pulser to die. One too many abuses over time. Probably kickback voltage along the ground path. I wish I had more time these days. Don't know how you do it during the holidays, but this is good info.
I made some progress on ls1 disassembly.
There are 3 dwell values calculated but it is not clear where they are used. It must be again some TPU usage.
The first value is the BASE dwell. It is derived from the IGN vs rpm table.
It is multiplied with the coolant vs ign multiplier and than multiplied with the map multiplier.
Than it is stored as BASE_dwell.
From the BASE_dwell a maximum and minumum allowable dwell variation is calculated.
The MAX_dwell is base_dwell * max dwell vs rpm table and is stored.
The MIN_dwell is base_dwell * min dwell vs rpm table and is stored.
The coolant scalar is a coolant value scaled for the table. There isn`t any fancy conversion there.
I am really interested in a hot coil test. Warming the coils to 80-100 degrees and run the test. I am still puzzled with the correlation from dwell times and coolant temperatures but they seem to be very important.
I also have a running code for 68hc11 to calculate dwell based on the ls1 tables. I used it for another GM pcm that was retrofitted to run ls1 coils. On that pcm I can change dwell via obd and can run some experiments with engine running. The code can be transferred to lt1 pcm and based on my calculation the pcm can run dwell upto 1.5-1.8ms at 6000+rpm without overlapping coils.
If you are willing to make a more close integration between LTCC and PCM we can setup dual mode. A cruise mode where the pcm calculates dwell and a power mode where LTCC commands dwell. The trigger will be from pcm pulling low a pin on the LTCC. In that case we can safely run the more dwell needed on startup.
Since I am short on coils now, I will need to buy a new set. I made a furhte investigations on different coils and there are roughly 5 different types.
1. square ls1 car coils made by denso.
2. square truck coils made by mitsubishy form 99-06
3. round truck coils[d585] made by Delco 00-06
4. square coils from 07-18, I managed to find 3 different stamps on them through out the years.a delco stamp , a GM stamp and one with p/n stamp.
5. round coils 07-18. couldn`t find much info on them.
I suppose you have option 5 in your setup. Is that the best coil to have. Based on earlier round coil design I think they will have the best output with the shortest dwell times.
I found somewhere that aftermarket coils have some crappy igniters that can cause all kind of problems. I also found some reviews that running too much dwell will damage the igniter and make it to preignite on shorter times. I am sure these are the myths that needs busting just like the few dozen milliapmps.
What we might miss with the few dozen milliapms is that they might run the coils without the 430ohm resistor to limit the current. The igniter might be hungry for more amps if it is not limited.
Did I mention that I managed to run the coil through a led and resistor. If we put leds on the circuit that might reduce the noise on the avr.
That was basically what I gathered from looking at the calibrations, minus the min / max value.Originally Posted by kur4o
For me, the takeaway from looking at the factory dwell tables was:
1) the factory calibrations for all intents and purposes implement zero dwell compensation for MAP
2) they seem to be reducing dwell as rpm increases
3) temperature matters
A little research led me to this page [link] which contained some information I hadn't found previously.
I recalled from ee 101 30 years ago that temperature changed the resistance of copper but had long since forgotten the direction and magnitude.
And I assume this is why the GM cals are reducing dwell with rpm (though at a relatively marginal rate).
As am I.
Which is why I ran all the tests I mentioned previously with the coils at room temperature (~21c) and then repeated after letting them let them sit in my oven at ~80c for a half hour. I didn't post any of that information because I really want to get a more complete picture at different voltages. But since you asked, here's what I observed.
This is compelling evidence that temperature compensation would be a good feature to add. But as always, there's a catch - depending on mounting location, underhood airflow conditions, etc. coolant temperature simply can't be relied on as an accurate representation of coil temperature.Code:data format: [dwell time in ms],[system voltage],[coil temp deg c] D580 - dwell time before current limiting evident 4.5,12.8,21 5.0,12.8,80 D581 - dwell time before current limiting evident 4.6,12.8,21 5.2,12.8,80 D514a - dwell time before current limiting evident 5.3,12.8,21 7.3,12.8,80 8183 - dwell time before current limiting evident 3.6,12.8,21 3.8,12.8,80 D585 - dwell time before coil fires 5.0,12.8,21 6.0,12.8,80
Again, I mean this in a light-hearted way, but I think you might have a patching addiction. Also, 1.5-1.8ms of dwell at 6000 rpm should quite clearly explain the stern way I'm going to answer the next question.
I'm no more interested in discussing this than I was when you mentioned it previously. It seems very pointless. If you want to control dwell in real time you could more easily modify the controller code, or implement a feedback loop, etc. Integrating this into a 25 year old ecu via patches just seems like the wrong direction. But, as I shouldn't need to mention again the source is freely available. Fork your own version and call it something else if you so desire.
By the way, "LTCC" is a product of Bailey Engineering. Let's stop calling what we've done here that please. Lest a cease and desist letter arrive in my mailbox for something I'm doing in a non, or more accurately, "negative" profit manner.
Part numbers make that list meaningful. Without them you open the door for assumptions to be made, which breaks the scientific method.
If you look hard enough you can find pretty much anything you (may or may not) want on the internet. Use much discretion in what you choose to believe, or believe nothing.
Let me jump off on a slight tangent here - and rant a bit. When I publish my test data I fully intend to disclose the source of the coils I tested. There is one that came in a Delphi box, but could easily be counterfeit. The rest are asian knock-offs with the exception of the 8183 coils which I hope are the genuine article, but were sold as "scratch and dent" and didn't come in original packaging. My point here is that I'm funding this adventure with my own money, and not one of you has made a monetary contribution to the project in any manner even after I've asked repeatedly to borrow coils to test. So while I'd like to be testing only the genuine article, they're goddamned expensive. Testing salvaged parts of unknown condition introduces another variable in the scientific method. So I've bought the cheapest stuff I can, because at some point the notion of spending money from my own personal toy budget is going to get sour. In fact, I'm starting to taste it now. Stepping off the soap box now.
I'm currently running 100 ohm resistors in place of the 470 in my car. I think I posted the igniter current at that configuration, but if I didn't I have it on paper somewhere. The difference was negligible, still well under 1 milliamp. So the 20ma number and "a few dozen" were clearly complete bullshit derived from non-scientific methods.
I wouldn't use an LED inline with a system critical circuit. LEDs are made to emit light, not protect the driver circuitry. Not to mention, you're going to create a huge voltage drop across them. There are no diodes inline in the ecu pics you shared, so I think you're inventing a fix for a problem that is imaginary until proven real.
Last edited by spfautsch; 01-03-2020 at 10:11 PM.
That seems to defeat the whole point of allowing overlapping dwell between coils if/when it's necessary.
You can easily find out by checking voltage drop across the resistor.
Unless it increases spark energy or decreases dwell time, there would be little point. It should be easy to test so I will - if dwell limiting happens at the same time then there would be zero benefit to changing the resistor size / igniter current.
Tested a D580 and a D514a coil today with a 100 ohm rc network resistor.
Side tangent (again). The purpose of this resistor isn't technically aimed at limiting igniter current. It's the buffer component of an R-C network, designed to absorb induced or reflected voltage spikes when the igniter circuit turns off and the comparatively massive amount of current running into the coil's primary winding suddenly halts and then attempts to change direction. The capacitor, which is fully charged in what we'll effectively call a "positive" direction quite effectively absorbs the reversed voltage spike when the circuit turns off. The resistor is a protection and buffer mechanism that limits positive charge current into the discharged resistor when the circuit initially turns on (goes from 0v to 5v) and then provides a similar buffering function when the negative spike is received.
Anyway, the D580 coil's igniter current increased from approximately 0.514 milliamps to approximately 0.600 milliamps. Likewise, the D514a coil's igniter current increased from approximately 0.471 milliamps to approximately 0.514 milliamps. More importantly, at the same system voltage of the previous test (12.8) there was no measurable change to observed dwell limiting with either coil.
In summary, it appears there's absolutely no benefit to messing with the value of the 470 ohm resistor. Going forward I'm considering this avenue of investigation a dead-end.
Funny story... I started to test this setup with the D585 coil until it occurred to me I'd have to re-arrange leads to observe dwell limiting with it and instead picked up the D514a. Anyway, in my haste forgot to disconnect the clamp from the positive battery terminal when I plugged the 4 pin connector into the coil. Sonofabitch did that "tickle"! Got the palm of my hand real good. I felt my teeth tingle, and when I began to gather myself started to wonder why I wasn't smelling my flesh burning. I do not recommend anyone try this unless you really hate life. It's a damned good thing the coil body is wrapped in a piece of grounded sheet metal, or instead of typing this I might instead be lying in a puddle of drool and urine, dead on my basement floor.
Last edited by spfautsch; 01-05-2020 at 12:52 AM.
So in summary, with this resistor, the D580 drew roughly 0.580 milliamps...and the D514a drew roughly 0.514 mA. Yay, fun with numbers!
I am about to get the SPFcop system tomorrow and start playing with the code. I doubt I will go anywhere but will give it a try.
You`d better start adding code for temperature compensation since my test vehicle run 100 times better with stock gm dwell tables with temp compensation, than the initial average settings with the stock code.
I wonder if the ls1 pcm can monitor ignitor current and readjust on the fly with the MIN and MAX dwell settings.
Posting Permissions
Reply With Quote
Bookmarks