nAst1: Progress and Concepts Thread

[robertisaar]

i can put those kind of values into my bench to test it, but it can handle really high PPM values...

BFBD (91-92 3.1 W-body calibration) is 40,000 PPM. with the FWD stuff, final drive ratio doesn't actually change PPM either.

the only real limitation to the VSS stuff is that the pulse divider setup for making the 4K PPM signal works best when dealing with increments of 4K PPM. the highest value you can use the divisor with technically, is 44,000 PPM, since it can do a /11 operation to make the 4000 PPM signal. if you go beyond that, then you'll need to "condition" the signal either before it gets to the ECM or between the ECM and the cluster.

[caffeine]

Ok I was thinking about it some more:
Stock T56 cars came with 3.42 gears, so if the 40,000 ppm signal is for 3.42 gears, that translates to around 48,000
ppm with 4.11s (what I have).

If I'm doing the math right, that puts me above the max ppm value. Is there any other way to account for rear gear changes?

Edit to add:
Changing the final drive wouldn't actually change the number of pulses per driveshaft revolution (17 IIRC), but the number of pulses per mile would change.

[robertisaar]

you'll need to do a tire size calculation to be exact on your numbers but:

the code to handle the MPH calculation is very simple... only a single value to convert it from a frequency to MPH. because of this, if you use a flip-flop or a decade counter or some other specific hardware, you can cut down the signal coming out of the VSS before it gets to the ECM and then have the calibration setup based on that.

say 48000PPM, run a flip-flop to cut the VSS pulses in half and then you can put 24000PPM in the calibration. then just set the divisor to match your speedo.

with a decade counter, you can get a lot more complicated and do things like /3, /4, etc....

keep in mind the magnetic VSS input wants to see an A/C signal, while the optical input would need a square wave. you can convert a square wave into A/C with a single capacitor. converting A/C into a square wave is a bit more difficult.

[caffeine]

Wouldn't I already have a magnetic VSS? I thought the optical VSS came from cable speedos...

I have an electric speedo in my car not cable. Firebirds got them in 87 or something.

[robertisaar]

yes, T56 should be magnetic, but you could condition the signal to make it appear to the ECM as if it were an optical setup, should it be necessary.

to calc your actual PPM, throw your rear tire size into here, grab the revolutions per mile value.

http://www.discounttiredirect.com/direct/brochure/info/tmpInfoTireMath.jsp

then multiply the rev per mile value with your final drive ratio.

then multiply that with the number of notches in the VSS reluctor. sounds like there are 17 notches on a T56?

using 245/50R16 tires with a 3.73 final drive, that's 46,378.074. using that, the ECM will get an accurate MPH signal, but the signal it would output to the cluster would be a bit off even with the /11 setting. the speedo would show 5% faster than reality, as would the odometer.

the VSS divisor can be programmed to do anything from /6 to /11, it also has a /1 option. if the 46,378 were cut in half via a flip-flop, then that's 23,189. doing a /6 on that would make the speedo/odometer 3.5% too slow. in reality, that means you would have to travel at 57MPH to actually be going 55....

it's not bad, but that's just how the divisor works on the P4 units. the P6 and P66 units can actually take in just about any value since they setup an internal timer based interrupt to toggle the signal to make the 4K PPM MPH signal adaptable to pretty much any combo of FDR and tire size.

[robertisaar]

after a rather long "hibernation" period, i'm getting back into this...

i have a beta spark cut rev limiter code setup, i'll be benching it tomorrow. assuming it goes well, if someone wants to test it(alpha), feel free to let me know. i currently don't have a running/driving vehicle that uses a nAst1 compatible PCM since i threw the original PCM back into the MC to get that stuff ironed out.

[caffeine]

As long as I can get my engine running, I might be able to test it in 2ish weeks. I just have to finish a bit of wiring, fill all the fluids; my mounts, charge pipe, and downpipe should be done by Thursday.

[LZeppelin513]

I am running on nast1 code and could test it for you.

[robertisaar]

hit a snag with the code... not sure why, but i'll investigate more tomorrow. for some reason, fuel was still getting cut and spark was not.

[robertisaar]

well, got that issue fixed....

a single byte in the patch was wrong, code was jumping to the wrong place and causing the ECM to reset under certain conditions.

now there is a switch to choose between fuel or spark cut, with fuel cut, it acts like it always did, with spark, fuel is still delivered as expected, dwell time is nothing(0.0000152587890625 mSec of dwell time, as opposed to the ~5mSec or so that would normally be run, so the coil won't charge anywhere near what is required to produce a spark). interesting result of the moment is that if hitting the cut threshold in P/N with the spark cut option selected, spark will still be cut and fuel still delivered as expected, but if you somehow are still revving up and hit the D(or step 3 if using 3 step) threshold limiter, fuel will then be cut.... i'm not sure how possible this is to achieve in a real-world vehicle or if it's a bad thing.... thoughts?

i may test some code that will literally not allow any dwell time, if that goes well, then i'll call it successful.

[Mars]

Damn sounds like progress. I may have to make the switch soon.

[robertisaar]

Forward progress even!

i should probably go through and document what all absolutely requires the extra SRAM of a 9396(soon to be compatible with a 7727 or 7730 with the NVRAM mod) inside the XDF, but on my short list that i've been keeping track of:

3/4 circuit PRNDL decoding(like the later 4T60E stuff)
TCC PWM
trans temp
3BAR stuff

anything other than those items and it will work in a 7730/7727 and to some extent, the 7749.

now i just have to get going and open up more PROM space via a rather confusing process of logic gates and address decoders.

[TurboGTU]

Quite a project you got going on here. Keep up the great work. Wish I had time to mess with my rides but those pesky diesels at work suck up my time and energy.

[robertisaar]

results and misc info:

max RPM input speed readable: 920Hz. assuming half as many pulses per revolution as there are cylinders(like a 60V6 ICM or pretty much any distributor):

4 cylinder = 27,600
6 cylinder = 18,400
8 cylinder = 13,800

that's assuming i didn't just max out my signal generator entirely.... either way, overkill is my business.... and business is good!



spark cut: still working on it, i have 3 revisions of the code going on and none of them do exactly what i want them to.... i'll have to do some digging and see why.

[Mars]

Cool was just having a peek at the files and got everything working. (I was on an old version of Tunerpro, 4.14 I think) and the .xdf's didn't work. Works perfect on the new version 5 though.

Has anyone real world tested the 3 bar map yet ? I want to know before I jump into another ECM swap...

Rob, are there any outputs on these ECM's that can push 12V ? I bought one of these boxes for my car, and it works perfect !! But I had to hack up an ignition module to get it to work. If you can get the ECM to power the coils, instead of ACC power, should be able to make it work...

http://www.code59.org/index.php?topic=8842.0