I just plotted the pw during the transition to sequential and get 1.7 ms bewfore and about 3.3 ms after. If you used .15 open time, that would give:
pw_wasted = .15 + (3.3 - .15) / 2 = 1.725, so the code is doing what I think it should be doing. But of course the .15 ms is way too small - it is a patch for something else that is going on. Possibly it needs x-tau enrichment because there is a longer period of time between squirts and this would affect the wall wetting balance. You would need to get a little more fuel out to maintain the balance.
More Data
Moderators: grippo, grippo2, Bruce Bowling
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.
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.
Re: More Data
The more I look at this the more it seems a transient. There is definitely a real effect - rpm dips, map spikes and even spark dips about a deg, probably from the rpm and map change. But the AFR comes back to the same AFR after the transition, so it isn't something you can tune out.
- Attachments
-
- monkey_wasted_seqtl_2.rtf
- (1.3 MiB) Downloaded 253 times
Re: More Data
In thinking about this last night, here is what is probably happening:
................................cam sync.........
........|........|........|........|........|........|........|........|.......|........|........|........|
.....1,4w.....2,5w....3,6w.....1w.....2w.......3w......4w......5w......6w......1w
The vertical tics represent tach cycles. Prior to cam sync we have simultaneous wasted injection of a short pw into 2 cylinders, then at and after cam sync we have full pw injection into one cylinder. In wasted each cylinder goes 3 tach cycles on a short pw. At and after cam sync, cylinders 1, 2 and 3 each get a full pw after 3 tach cycles, BUT, cylinders 4-6 will each have been without fuel for 6 tach cycles before they get their full pw. That in itself is not a problem, because it is only when they get the injection that they need it, but what has happened to the fuel on the walls. It has had 6 cycles to evaporate, so it needs more than a full pw to make up for this. I thought of this when I first wrote the code, but forgot about it til now. This should take care of itself if x-tau is on and tuned because the code keeps track of the time between squirts for each individual cylinder, and this time is used in the x-tau algorithm to calculate how much fuel has evaporated.
If I get a chance, I will try to duplicate your results and then turn on x-tau and see if it adds fuel to cylinders 4-6 right after cam sync.
Another thing to note, cam sync will normally occur during the startup process and any glitching will likely be less noticeable (but maybe worse - I'm not sure). It occurs much later with Scott's setup because he has a VR sensor on the cam and the voltage on this is rpm sensitive, so it takes a while before it gets sufficient strength to register. Most installs use a Hall sensor for the cam sync.
................................cam sync.........
........|........|........|........|........|........|........|........|.......|........|........|........|
.....1,4w.....2,5w....3,6w.....1w.....2w.......3w......4w......5w......6w......1w
The vertical tics represent tach cycles. Prior to cam sync we have simultaneous wasted injection of a short pw into 2 cylinders, then at and after cam sync we have full pw injection into one cylinder. In wasted each cylinder goes 3 tach cycles on a short pw. At and after cam sync, cylinders 1, 2 and 3 each get a full pw after 3 tach cycles, BUT, cylinders 4-6 will each have been without fuel for 6 tach cycles before they get their full pw. That in itself is not a problem, because it is only when they get the injection that they need it, but what has happened to the fuel on the walls. It has had 6 cycles to evaporate, so it needs more than a full pw to make up for this. I thought of this when I first wrote the code, but forgot about it til now. This should take care of itself if x-tau is on and tuned because the code keeps track of the time between squirts for each individual cylinder, and this time is used in the x-tau algorithm to calculate how much fuel has evaporated.
If I get a chance, I will try to duplicate your results and then turn on x-tau and see if it adds fuel to cylinders 4-6 right after cam sync.
Another thing to note, cam sync will normally occur during the startup process and any glitching will likely be less noticeable (but maybe worse - I'm not sure). It occurs much later with Scott's setup because he has a VR sensor on the cam and the voltage on this is rpm sensitive, so it takes a while before it gets sufficient strength to register. Most installs use a Hall sensor for the cam sync.
Re: More Data
thanks Al.
that datalog is when it was idling steady state in wasted, and then plug in the cam sensor. the transition was noticeable. but, when the sensor is plugged in prior to startup it transitions right away and nothing is noticeable (maybe wall wetting effect of transition gets diluted during ASE). in the datalog i posted, the AFRs seem to be flatter after sync.
also, MT seems to round open time to one digit, so 0.2ms is what is kept in memory.
when you said: << So this should work and you should see a smooth transition without changing the topen time >>... i don't completely understand...
are you saying that transition should be smooth only after topen has been adjusted, or are you saying that transition should be smooth without having to adjust?
for me, the effect of changing topen was profound. starting at 0.8, it transitioned wayyy lean (pinned AFR at 20 and tried to stall), and i kept adjusting by 0.1ms until 0.2ms created a bit lean transition, and 0.1ms topen created a bit rich transition. but it definitely kept scaling down while topen was being adjusted.
Scott.
that datalog is when it was idling steady state in wasted, and then plug in the cam sensor. the transition was noticeable. but, when the sensor is plugged in prior to startup it transitions right away and nothing is noticeable (maybe wall wetting effect of transition gets diluted during ASE). in the datalog i posted, the AFRs seem to be flatter after sync.
also, MT seems to round open time to one digit, so 0.2ms is what is kept in memory.
when you said: << So this should work and you should see a smooth transition without changing the topen time >>... i don't completely understand...
are you saying that transition should be smooth only after topen has been adjusted, or are you saying that transition should be smooth without having to adjust?
for me, the effect of changing topen was profound. starting at 0.8, it transitioned wayyy lean (pinned AFR at 20 and tried to stall), and i kept adjusting by 0.1ms until 0.2ms created a bit lean transition, and 0.1ms topen created a bit rich transition. but it definitely kept scaling down while topen was being adjusted.
Scott.
Re: More Data
You should not have to adjust topen time to go from wasted to sequential. The code should take care of it. Since it doesn't in your case, either the open time you are using is wrong or the simple model I use in the code is inadequate, or the mixing produced by the 2 squirts per cylinder vs 1 bigger squirt per cylinder is different and hence requires a different factor than just multiplying the sequential fueling by 1/2 to get wasted fueling. We may need to make this factor a user input which defaults to 1/2. Then you could adjust it that way and leave the 0.8 ms open time. But I suspect such a factor would also be a function of rpm because that would affect fuel mixing, and maybe even load would be a factor. Then we get into another table. I have not seen any models of 2 squirts vs 1 squirt per cylinder, but I am sure they exist. We will have to take a look. But right now it is not worth another table to make the transition smoother, because as you say it is not noticeable on a normal startup, with cam sync occurring immediately thereafter. Any fuel shortfall/ excess can be tuned out with crank pw and ASE.
It would be nice to get some data on pw vs afr at several steady rpms running wasted then switching to sequential and letting the transient die out. This all needs to be done with a single topen otherwise it gets too complicated to analyze the data. Having the AFR we can calculate what pw we need to achieve it and then we can try changing topen and then try just changing the 1/2 factor, and see which one gives the best results across all rpms. What I am saying is that if you let the car idle at 3000 rpm and then insert cam sync, I don't think you will see as much, if any change in afr after you go to sequential. I can do all this through simple math if I have pw vs afr vs wasted and sequential for idle and say 2 or 3000 rpm.
It would be nice to get some data on pw vs afr at several steady rpms running wasted then switching to sequential and letting the transient die out. This all needs to be done with a single topen otherwise it gets too complicated to analyze the data. Having the AFR we can calculate what pw we need to achieve it and then we can try changing topen and then try just changing the 1/2 factor, and see which one gives the best results across all rpms. What I am saying is that if you let the car idle at 3000 rpm and then insert cam sync, I don't think you will see as much, if any change in afr after you go to sequential. I can do all this through simple math if I have pw vs afr vs wasted and sequential for idle and say 2 or 3000 rpm.
Re: More Data
<< It would be nice to get some data on pw vs afr at several steady rpms running wasted then switching to sequential and letting the transient die out. >>
i can definitely do that. just turn off ego correction, go for a spin in wasted and duplicate same route/speeds again in sequential. using same topen that works for idle transition would be good starting point.
BUT... i ran into a bit of a stumble (pun intended). i was datalogging, in steady state on highway, and there felt like a rev limiter at 3,000 rpm. i couldn't throttle through it, a few tries later i just downshifted and it seemed to shake it off. at first i thought the problem would be the resistor inline on crank VR (that limited at 1300 rpm before i put resistor in), but that would have shown a tach dropout on the datalog. the datalog does not show any tach dropouts or trigger +/-.
see attached datalog (trimmed down for just period of interest). you can see the problem start at timestamp 8046... it starts to go lean, and then if you watch the TPS, i try to throttle through but it just goes super lean and the RPM won't go through 3,000. keeps doing it as the car slows down to 2200 revs, i just can't accelerate. timestamp 8075, i shift from 5 to 4 and at 2800 rpm same problem, i can't accelerate through. timestamp 8078 i just take foot out of it, coast down to 1750 RPM, shift into 3rd.... and away i go.... everything was okay after that.
it happened again on the way home. once at low speeds between lights similar RPM, and then again on the highway similar situation as above, but not for as long.
i can definitely do that. just turn off ego correction, go for a spin in wasted and duplicate same route/speeds again in sequential. using same topen that works for idle transition would be good starting point.
BUT... i ran into a bit of a stumble (pun intended). i was datalogging, in steady state on highway, and there felt like a rev limiter at 3,000 rpm. i couldn't throttle through it, a few tries later i just downshifted and it seemed to shake it off. at first i thought the problem would be the resistor inline on crank VR (that limited at 1300 rpm before i put resistor in), but that would have shown a tach dropout on the datalog. the datalog does not show any tach dropouts or trigger +/-.
see attached datalog (trimmed down for just period of interest). you can see the problem start at timestamp 8046... it starts to go lean, and then if you watch the TPS, i try to throttle through but it just goes super lean and the RPM won't go through 3,000. keeps doing it as the car slows down to 2200 revs, i just can't accelerate. timestamp 8075, i shift from 5 to 4 and at 2800 rpm same problem, i can't accelerate through. timestamp 8078 i just take foot out of it, coast down to 1750 RPM, shift into 3rd.... and away i go.... everything was okay after that.
it happened again on the way home. once at low speeds between lights similar RPM, and then again on the highway similar situation as above, but not for as long.
- Attachments
-
- datalog200907181229 trimmed.txt
- stumble, fuel rev limited 1846->8090
- (375.54 KiB) Downloaded 227 times
Re: More Data
With regard to testing pw vs wasted / sequential - I would first do it in the driveway where things are much more controlled. Plus I want to first see if it is a function of rpm, then we can see if it is a function of load.
As far as the rest, that is scary. I will take a careful look at the datalog. But was this the first time you drove it like this or did it work before and now doesn't. If that is the case - what is different ?
As far as the rest, that is scary. I will take a careful look at the datalog. But was this the first time you drove it like this or did it work before and now doesn't. If that is the case - what is different ?
Re: More Data
OKay... i'll do a couple driveway logs in wasted & sequential without a load and post them.grippo wrote:With regard to testing pw vs wasted / sequential - I would first do it in the driveway where things are much more controlled. Plus I want to first see if it is a function of rpm, then we can see if it is a function of load.
As far as the rest, that is scary. I will take a careful look at the datalog. But was this the first time you drove it like this or did it work before and now doesn't. If that is the case - what is different ?
the rest:
Not sure what is different, other than it seems to be intermittent. This happened once a few days ago also before cam sync. I thought I was running out of gas because I don't trust my gauge so I floated to gas station. But it was same sensation, so I'm sure it's the same thing.
Car runs fantastic, except this suddenly happened. It's not too scary because I'm not doing any WOT stuff, and it behaves just like a fuel cut rev limiter. It just slows down on the highway, and then it suddenly 'snapped' out of it. I guess an intermittent problem (although does have to be figured out). Only thing that really changed before this happened was changing EGO correct from 10% to 15%, but that seems pretty benign (except you'll notice that my correction oscillates; different issue - tuning only). It's not a fuel supply issue because I can get through the gears pretty quick and drive with a heavy load on the highway with no issues (when problem is not occuring). For reference, I've driven about 80 miles with cam sync on, and the problem has happened only briefly a few times - I don't think it's conditional problem, it acts intermittent.
I'll post the 2 long logs (4mb each) on media fire if you want to take a look through them. The earlier time stamp log has the problem I already shared, but it's the full log, and the second log is the drive home where it happened again - once at low speeds, and once on highway.
If it was a bad signal from crank VR (the resistor needed inline)... it WOULD show up as RPM dropout & reset wouldn't it?
Drive out: http://www.mediafire.com/?xtizmy2znek
Drive back: http://www.mediafire.com/?0wjzz2zjzqn
Re: More Data
Just ran another errand (is that all I ever do?).
Car was pretty much undriveable for a period on the way there. I stopped car at my destination, and on way home it was great. Right through the gears.
Log of there and back attached: http://www.mediafire.com/?cnnmi1yqnh1
175 - 328 no response from throttle.
limped to destination see very low TPS until processor reset.
return home, i ripped through the gears at time stamp 638.
I'll try to think of a way to monitor fuel pressure on the fly to make sure it's getting gas. other thing i can try is to scope an injector on the fly.
goes real lean like it's running out of gas. seems to me that it's either not getting the gas, or it's not sending the signal to injector that the log says it is.
Car was pretty much undriveable for a period on the way there. I stopped car at my destination, and on way home it was great. Right through the gears.
Log of there and back attached: http://www.mediafire.com/?cnnmi1yqnh1
175 - 328 no response from throttle.
limped to destination see very low TPS until processor reset.
return home, i ripped through the gears at time stamp 638.
I'll try to think of a way to monitor fuel pressure on the fly to make sure it's getting gas. other thing i can try is to scope an injector on the fly.
goes real lean like it's running out of gas. seems to me that it's either not getting the gas, or it's not sending the signal to injector that the log says it is.
Re: More Data
Good news... A clue....
Check out battery voltage on the datalog. When it goes bad, the voltage jumps from ~13.7v to ~17.8v.
Hopefully this is a bad regulator on my car. I'll just hook up a voltage meter and watch actual voltage. I'll report back if voltage is actually jumping up. I swear this project throws me more curve balls....
Check out battery voltage on the datalog. When it goes bad, the voltage jumps from ~13.7v to ~17.8v.
Hopefully this is a bad regulator on my car. I'll just hook up a voltage meter and watch actual voltage. I'll report back if voltage is actually jumping up. I swear this project throws me more curve balls....