Duty Cycle Spike?

wes kiser · Post by **wes kiser** » Mon Feb 13, 2006 8:09 pm

I went over my datalogs, and think I found the source of the occaisional "missfire" I was experiencing (actually, likely another symptom, but anyway). It only seems to happen around 3000 rpm and moderate load. When looking at the datalog, there was a huge but quick spike in duty cycle, but nothing else. The duty cycles spike from around 10%, to 80%. This lasted way less than 1 sec. everything else appeared relatively steady. How is this possible? No other sensors used to calculate duty cycle appeared to move (map, coolant temp, air temp). Rpm looked steady, etc... There are also no strange irregularities in the table. Nothing else appeared to "spike" in the datalog. Thanks

Post by **grippo** » Tue Feb 14, 2006 9:11 am

This may have something to do with your choice of ignition offset or trigger angle. The code may be trying to switch between next cylinder and this cylinder when it gets to the advance you set for 3000 rpm, moderate load. When this happens it can spike the dwell - but this won't hurt anything because it is just momentary and doesn't have time to get the coil anywhere near overheated. However, it should not cause a miss. Can you tell us what your offset trigger angle is ?

wes kiser · Post by **wes kiser** » Tue Feb 14, 2006 1:22 pm

I should have clarified what I meant when I said duty cycle, the injector duty cycle is spiking, without a corresponding blip in any input data. After thinking about it, the only thing that could spike it that high is the map spiking to it's maximum value (3 bar in my case). I have only found the duty cycle spike on one datalog. I think the reason I didn't see any other spikes (in any inputs) is they are too quick, and the sample rate of the log didn't catch them (and just coincidentally caught the duty cycle spike this one time). I am going to try to burst log and see if I can get anything. Thanks for the reply. It is frustrating because most of the time I can't get it to do it. For the record, my trigger offset is 12 degrees, and my commanded timing value is between 16 and 44 degrees at 3000 rpm depending on load (32-40 in the area it is acting up). This is my daily driver, and it is otherwise running great.

m0ntecarloss · Post by **m0ntecarloss** » Tue Feb 14, 2006 2:24 pm

This sounds suspiciously similar to my problem I posted earlier except in my case it is injector pulse width and duty cycle. Also mine is happening at about 700rpm causing the motor to stall. Got one on log there and according to every sensor they are all stable, don't see why the spike is occuring.

I never noticed this happening before but I also used to have the motor idling at 900, never this low. I am going to try going back to an earlier code version for my case (i'm currently running v2.34)

http://www.msefi.com/viewtopic.php?t=15666

Bernard Fife · Post by **Bernard Fife** » Tue Feb 14, 2006 2:32 pm

wes,

Does the injector pulse width ever spike as well? DC is calculated entirely within MegaTune, MegaSquirt doesn't even know about it.

If the PW does spikes, do you feel this in the car, or is it something that only appears on datalogs (in which case I wonder if there's a comm glitch)?

Finally, if the PW and DC do jump, can you post an datalog that has this event this in it for us to have a look at?

Lance.

wes kiser · Post by **wes kiser** » Tue Feb 14, 2006 6:25 pm

I assumed that pulsewidth also spiked (as duty cycle is obviously derived from pulse width), but I will check. You can feel it (it feels just like you would expect when A/f goes form 14:1 to 2:1, feels like it temporarily cuts off, but just for a instant, then all is well). I "burstlogged" on my way home from work, and couldn't get it to mess up one time. I tried 2600 to 3500 rpm in slow sweeps and from 50kpa to 105kpa, couldn't get it to act up one time. Just out of curiosity, if MS looses the map signal, what map value does it assume (I know with coolant temp it picks 180 F, what map value does it assume?). Thanks again for the replies.

FixItAgainTony · Post by **FixItAgainTony** » Tue Feb 14, 2006 8:58 pm

Hello Wes,
This sounds similar to what I am seeing on my initial tuning efforts.
V3 board, MII daughter card v2.33 firmware. I observed no major glitches on sensor inputs except voltage, which I also log. BattV went up to 21.5V, which seems a little odd - like an A->D conversion error <unless my alternator is out to lunch>. I am able to run with the original L-Jet computer or MegaSquirt. L-Jet does not have the problem, so I think that my external wiring harness is ok.

A datalog fragment of the stumble is attached <if I am doing these forums right - this is my first post - heh - you forced me out of the closet

>

The stumble was momentary - like a 100-200 RPM drop and then picked up again.

Hope this helps.

Charles.

wes kiser · Post by **wes kiser** » Wed Feb 15, 2006 5:33 am

Here is a burstlog on my way to work. I stopped the log about 5 seconds after the miss, so look at the end. There is a short spike in duty cycle and pulsewidth. It is a huge spike in magnitude, basically it is enough fuel for 30psi at that rpm (spiked up to 77% when just .1 second previous with no changes it needed only 10%). I can also attach my msq, if there could be something helpful in there. The car is running great other than this.

mops · Post by **mops** » Wed Feb 15, 2006 12:20 pm

sounds very like the problem I'm having.
initially i thought it's a spark miss/weak spark, but it might be very short PW spike.

all tho my car is getting it across very wide rang of conditions...

idle is pretty good, afr very stable +-0.1, but even 1700rpm, steady rpm and throttle afr's go up and down from 12 to 16 oscialting about once a second, i can really feel it in the car, like it more powerfull on rich and then stumbles on lean, and then goes rich again... even tho rpm are relatively stable, i can clearly distinguish the engine note when it's happening. I'll post my datalog later on...

wes kiser · Post by **wes kiser** » Wed Feb 15, 2006 1:19 pm

When it happens it is too quick to even register on lambda (either using wideband, or the narrow band). It is just for an instant. I really don't think it is an internal hardware issue, as I can beat/shake on the MS and it doesn't so much as hickup. What baffles me is I can't see a change in any inputs that would result in that change in commanded pulse width. I also noticed in both cases I datalogged, the pw spike was equal to 32, even though because the rpms were slightly different this resulted in slightly different duty cycles. Is there any significance to this in and of itself?