Acceleration Wizard question

[gboone]

I'm struggling with how to set the Rate numbers in Accel Wizard. I'm using 100% TPS based. The manual says the following:

By watching the bar graphs in the accel wizard, you can see how much noise there is in the respective sensor signals. try to set you threshold(s) above these levels to avoid unwanted accel enrichment, which can make you engine run very poorly.

How do I interpret the bar graph relative to noise in the respective sensor, what sensor is producing noise? The bar graph changes so fast when I stroke the throttle I can't get much info from it visually. And after I establish the rate values, how do I set the corresponding values?
Thanks for any help.

[loren]

The bar graph is telling you the TPS rate of change at any given moment. The TPS signal shouldn't have a lot of noise on it. If it did for some reason, that noise might trigger as a false "rate of change". I've not seen that on TPS, more on MAP... the normal MAP fluctuations WILL show up as a rate of change and if you use MAP for AE, you'll need to be sure the threshold is set above that "noise" level.

What you should be looking for on the bar is where the bar peaks when you blip the throttle. What is the fastest rate of change that you can induce in your TPS? Whatever that is, that's about what you'll want your highest bin for AE to be. Fill the remaining bins below that one with lesser values, so that when you dip into the throttle with "less gusto", you still get some AE, just not as much.

Best way to see if your TPS (or MAP) threshold level is set too low is to get AE somewhat tuned, then run a log and see if AE is kicking in at idle or any other time that you're not intending it to. Getting AE at idle does really weird things, you'd probably notice it, anyway.

[sawnuoff]

I was just going threw some simular troubles, it maybe easier to watch in the realtime display. I was getting lean spikes between shifts and had to adjust the tps rates and there value. To get an idea if you look in the datalogging section under the thread usng mlv for acel wizard or something like that, its one of the first few threads there, thats were I started posting for help. The last 2 pages maybe of some help to you.
Watching the realtime display with lite blips of the throttle they only go up to 10 or something that would be the lowest cell in the rate column, then blip it a little harder and so. Mine are posted in that thread what the are but have been playing with them now that I have somewhat figured them out, still fine tuning to do though. Also if you look in your logs at your tpsdot graph those spikes, that is your pedal speed and also tells you what cell you are getting your enrichment from. I set all mine pretty high so it would go rich and thats what I am trying to get tuned a little better now.
Do some logs with the car not running and just sit there and litely blip the throttle and then get more agrressive and then look at them spikes in the log and you might be able to get an idea of were your beens need to be.
Hope this helps you a little bit.

[devastator]

To get an idea if you look in the datalogging section under the thread usng mlv for acel wizard or something like that, its one of the first few threads there, thats were I started posting for help. The last 2 pages maybe of some help to you.

This is the thread sawnuoff is referring to: viewtopic.php?f=25&t=34109

How do I interpret the bar graph relative to noise in the respective sensor, what sensor is producing noise?

If the bar graph is bouncing around without you pressing the pedal, then there is noise. The noise is probably not produced by the TPS, but it is interfering with the TPS signal.

The bar graph changes so fast when I stroke the throttle I can't get much info from it visually.

I suggest using MLV to set your accel rate settings. You can turn on a custom graph just for tpsdot, making it easier to find the highest and normal rates. As far as the added fuel values, watch your WBO2 and see if you get a lean or rich spike right after accel, find the tpsdot value that triggered the accel event at that point, (using MLV), and adjust the fuel value accordingly. Also, Tuner Studio has an easy to adjust bar graph with immediate AFR feedback. My accel rate values are too clode together to use it, but yours might not be.

[gboone]

Using Tuner Studio, I did several throttle stabs without the engine running while watching the %/sec values. The numbers got extremely high, up to 1700. I plugged some higher numbers, corresponding to what Tuner Studio was telling me, into Megatune Accel Wizard and it just got worse with AFR when running the engine and stabbing the throttle. The AFR was all over the place. I'm keeping the Accel slider set to 100% TPS based. So I set the accel TPS bins back down to this:
200 25.5
100 20.0
50 10.0
10 2.0

When I run the engine with these numbers, It seems to be better with AFR but still sporadic. When I run the engine and without touching the throttle, I get no bar graph bouncing or any red at all on the bar graph at all on the TPS side bar but get a lot on the MAP bar graph. But since I'm using 100% TPS based accel, is this a problem?

I've attached corresponding msq and datalog files. Thanks for your help.

megasquirt200901311740.msq

datalog200901311555.xls

[devastator]

Using Tuner Studio, I did several throttle stabs without the engine running while watching the %/sec values. The numbers got extremely high, up to 1700.

Hmmm. That's strange. I dunno enough about TS to help you there.

When I run the engine and without touching the throttle, I get no bar graph bouncing or any red at all on the bar graph at all on the TPS side bar but get a lot on the MAP bar graph. But since I'm using 100% TPS based accel, is this a problem?

No. This is exactly what you want to see if you are using TPS only for accel. You can usually raise the MAPdot threshold to keep the bouncing MAP from causing random accel events, but that's only if you are using MAP based accel in any percentage.

So I set the accel TPS bins back down to this:
200 25.5
100 20.0
50 10.0
10 2.0

Your highest tpsdot reading is 492 in your datalog, and just barely "touching" the pedal seems to be between 20 and 55.
Try something like:
380 25.5
260 20.0
140 10.0
25 2.0
These added fuel values seem pretty high, and show up as rich spikes on the datalog.
With your previous settings, you were using the first and last accel values, for most of your accel events. Hopefully these values will give you more tunability.
Let me know how these settings work for you.

[gboone]

Thanks devastor.
I put your numbers into the accel wizard and it seems to be working much better in the garage at least, while watching the AFR gauge. But I still need to test them while driving the car so there is some load on the engine. It's been cold and snowy here in Wyoming the last few days. My car is topless.

Thanks for your help. I feel more confident it will be driveable now for tuning. Last time I tried a drive a few days ago, it fell on its face, way too lean to drive safely while trying to accelerate. Didn't want to burn pistons.
Gary

[devastator]

Glad I could make some sense of this for you Gary. Hopefully my posts make enough sense that anyone can examine their datalogs for tpsdot highs and lows and extrapolate an appropriate accel table to start with. Now if only I could do the same with mapdot.....
Dev

[gboone]

Thanks Dev,
I've tried some percentages of MAPdot but have experienced worse engine running conditions. So I'm using 100% TPS. But I'm wondering if I should pursue using MAPdot again since my elevation is 6200 feet above sea level. Do you think it's worth pursuing for that reason alone?
Gary

[devastator]

I've tried some percentages of MAPdot but have experienced worse engine running conditions.

Which values are you changing, the rate or the added fuel amount, or the MAP/TPS accel ratio?

So I'm using 100% TPS.

Were you always?

But I'm wondering if I should pursue using MAPdot again since my elevation is 6200 feet above sea level.

I live at 4700 ft, but do most of my driving at sea level. My accel works the same at both altitudes. The altitude variable affects the VE table value that you are using, since it is part of the fueling equation to begin with. If you are at the same AFR at an altitude, the accel will react in the same way.
My advice is to set the accel table to 100% tps, set the values to the numbers I gave you earlier, (not that they are perfect, but a good starting place), and dial in the accel added fuel amount, (the number to the right of the corresponding pedal speed). Using the WBO2 gauge in MT is a good way to tell wether or not you are adding too much or too little fuel. You can also manipulate the pedal speed settings, and/or tpsdot threshold, to isolate them individually. A datalog of this whole process helps too, as long as you remember when you made changes to the added fuel amounts.

[gboone]

The weather was warm enough today to take a drive in my topless car here in Wyoming, first time since I made the recommended changes to acceleration enrichment (see previous posts). I did a 10 minute drive and tried to drive thru several rpms but not much over 5000 rpm. I took the laptop in the passenger seat and watched the AFR gauge. It ran fairly well except for when I tried to accelerate hard or even moderately hard. It went off scale lean according to the AFR gauge on Megatune (I have a WBO2 sensor) and it stumbled very badly when it went lean when I tried to accelerate. But it cruised very well. I did a VE analyze with Megalog Viewer afterwards and it told me I need to fatten the numbers in some areas, but I'm still not convinced that the Accel Wizard does not need some adjustments. Maybe someone can advise on what changes I should make. I attached the msq and datalog files that were in effect during my drive today. Is it really true that if the Accel Enrichment has no effect on the VE Analyze in Megalog Viewer and it doesn't overcorrect or undercorrect?
Thanks for any help.
Gary

datalog200903011650.xls

megasquirt200902280030.msq

[devastator]

ran fairly well except for when I tried to accelerate hard or even moderately hard. It went off scale lean according to the AFR gauge on Megatune (I have a WBO2 sensor) and it stumbled very badly when it went lean when I tried to accelerate.

Try fattening up your accel values, (not the tps/s values though). Remember that the accel is added fuel to the PW, so if you increase the AFR with the VE table, you will also yield a richer accel value.

Is it really true that if the Accel Enrichment has no effect on the VE Analyze in Megalog Viewer and it doesn't overcorrect or undercorrect?

Yes, this is correct. (This is the reason I like it over Autotune.)

[gboone]

I took 2 drives in the car today. The first one was with my revised VE table from yesterday's MLV VE Analysis. It ran good but not when trying to accelerate, still went way too lean when trying to accelerate. So I came back and fattened the accel values per your recommendation. It wouldn't even run decent at all even steady throttle, a much worse condition than before. AFR was all over the place with a steady TPS. I've attached files.

datalog200903021529.xls

datalog200903021556.xls

megasquirt200903022224.msq

[devastator]

Your VE table still needs some more tuning, as your AFR's are moving all over the place, even when the accel, (and decel), is not activating. Also, you changed the pedal rate and not the amount of fuel added, which takes you out of all but 2 of the accel bins. Put the pedal rate, (the left side numbers), back where they were and try changing the added PW values, (the numbers on the right). I would be wary of that lowest number of 25 that I think you had earlier. It might be making the pedal "touchy". Try a lowest value of 50. You can always change it back if you don't like the result.