KPierson

iTrader: (1)

I have a 60-70 pounder out in the shed that one of my instructors gave me when I was in college, but I've actually never used it (it is missing leads). When I started searching for leads I found it was just easier to buy a new, low frills, model. Mine weighs probably 20 lbs, but its pretty big physically, as it is a table top model.

etx

Hey Cool project. I'd say lift the car and check the sensor output. It's probably a std 32 clicks per rev methinks. Your biggest problem will probably be throttle pos control. I'd say look into spoofing the throttle pedal sensor.

KPierson

iTrader: (1)

What do you mean about 'std?' Is this an SAE standard? 32 pulses per rotation would be perfect.

Throttle position shouldn't be too hard, just intercept both analog outputs from the accelerator pedal and scale them accordingly.

My biggest hangup right now is calculation throttle output. I don't know how to implement a PID loop in to a microcontroller, but I'm sure I could get a PID routine. But, I'm not 100% sure a PID loop is needed for this application, it almost seems like over kill. My original idea was to create 4 -5 levels of slip and decrease throttle position a certain percentage based on what level of slip the car is experiancing.

Also, developing the laptop interface to do the programming and monitoring of the everything is going to take some time to perfect. I would say there is easily enough interest, now lets just see if we can surpass the infamous 60+ window mod page from G35driver! haha

AboostedG

iTrader: (6)

Wouldn't you run into some problems with the 05's and the recent changes it went through??

03Z33

iTrader: (49)

Check it out guys:

www.racelogic.co.uk

US distributor is:

www.matrixintegrated.cc

I'll be installing one of these in my Z in the next couple of months, and will use the following options:

Launch Control
Full Throttle up-shift
Drift controller

I have ridden in a 700HP Supra with this system and it's spectacular!

btw, it uses the factory abs/wheel sensors and wiring should be fairly straightforward. I will attempt to do a write-up once I have it all figured out.

G3po

iTrader: (8)

Sounds like big $$$. How much?

I would expect that KP's approach to be much much cheaper.

03Z33

iTrader: (49)

yep it's ~$900 to start out, but it works great

Don't forget this system uses a fuel cut, not your brakes or ignition retard. also the computer can make over 1,000 corrections per second, much faster than any OEM VDC/TCS including the supercars

xswl0931

It'll be interesting to read your review. If possible, get some 0-60, 1/4 mi times before and after. Even track times if possible.

KPierson

iTrader: (1)

I would imagine the launch control would cost 1/3 of that, and work just as good. It would cut fuel (based on throttle position. A fuel cut is a terrible idea, as it will run lean. I would imagine (for $900) it would actually cut spark (and maybe fuel too). The beautiful thing with changing the accelerating pedel is that we will be affecting air intake which will be metered and the ECU will adjust for it.

Using a 16mhz CPU we will have 16,000,000 instructions per second, at 1,000 calculations per second I would have 16,000 clock cycles to complete the math. That shouldn't be a problem. My only limitations will be the speed of the wheel sensors, as you can't analize data faster then you can gather it.

03Z33

iTrader: (49)

Don't get me wrong, I love your idea and think it could work great! For me I will be using this for more than just launch control, it is just a nice extra.

Also a few points:

The wheels speed sensors on the Z/G35 are simply a magnet pointed at a ring gear with 66 teeth. This has no effect on processing speed. It all depends on your processor. The WSS (wheels speed sensor) will send 66 inputs for every full rotation of the wheel.

Fuel cut: The racelogic uses a "safe" fuel cut which completely cuts ALL fuel from one cylinder at a time for one cycle at a time. When there is no fuel in the cylinder, you will not run lean, since that cylinder will not have any fuel in it to fire for that revolution.

Here is how racelogic explain it bettter (copied from http://www.racelogic.co.uk/traction/works.htm ) :

"Fuel Cut
The idea of cutting fuel to an engine sets alarm bells ringing in engine builders, as they all know of the potential disaster of a high revving race engine running lean. Running in a lean combustion mode will elevate in-cylinder temperatures very rapidly, the denser the air/fuel charge, the more heat the lean burn can generate. Therefore it is vital that a fuel cut system will not cause a lean burn.

The simplest way of preventing a lean burn is to remove more than 50% of the fuel from the pulsed delivery. A mixture will only ignite if the air/fuel ratio is within a tightly defined window, look at the efforts being put into making lean burn engines fire on very low air/fuel ratios (1:20 or more). Removing more than 50% of the fuel will cause an air fuel ratio of over 1:25 and will result in a complete miss-fire, with the unburned fuel passing out through the exhaust valve. Even if a high air/fuel ratio did manage to ignite, the energy available from the amount of petrol injected wouldn't be enough to elevate temperatures significantly. Of course the ideal system will remove 100% of the pulsed fuel delivery, allowing the cylinder to take a gulp of fresh air, and the in-cylinder temperature would remain virtually unaffected.

Prolonged fuel cut on one particular cylinder would cause scavenging of the petrol lining the inlet tracts, and when the next full fuel pulse arrived, it would be partially reduced in quantity by the re-wetting of these tracts. Therefore it is often important to manage a rotation of the cylinder cutting to prevent this situation from occurring."

Also here is there take on spark cut (:


"Spark Cut
Cutting the spark to an engine will stop any chances of a weak mixture occurring, but it carries it's own potential problems due to a large quantity of unburned fuel travelling through the cylinder and out of the exhaust. This petrol can remove some of the oil lining the inside of the cylinder, and pass it thorough the exhaust, again this only becomes a problem if the fuel to one particular cylinder is cut for an extended time. The best way to overcome this is to rotate the order in which the cylinders are cut.

The unburned fuel in the exhaust will have a catastrophic affect if there is a catalytic converter in the exhaust, as it will try to convert the unburned fuel to harmless elements, effectively burning the mixture. This causes the catalytic converter to heat up very rapidly, reaching temperatures in excess of 1000Ž°C, and possibly melting down completely. Thus prolonged spark cut is not recommended for catalytic equipped cars. "

Here are some more tech pages:

http://www.racelogic.co.uk/traction/adjust.htm (tcs adjustment)

http://www.racelogic.co.uk/traction/driftadjust.htm (drift adjustment)

http://www.racelogic.co.uk/traction/launch.htm (launch control)

http://www.racelogic.co.uk/traction/fts.htm (full throttle shift)

http://www.racelogic.co.uk/traction/datalog.htm (datalogging)

KPierson

iTrader: (1)

Thats very interesting.

If the sensor does have a resolution that high it would be very easy to do mass amounts of calculations per second, as you wouldn't sample the entire wheel revolution, just 1/66th or 2/33rds. So, the amount of pulses per revolution will have a profound affect on how fast the processor can make changes, as, like I said before, you can't analize data faster then you can gather it. I'm working on figuring out how many wheel revolutions take place at 2,000RPM so I can calculate the over all resolutiong of a 16mhz system. From what I've seen so far, 16mhz will be more then sufficient.

How did you find out the wheel sensors pulse at 66 times per revolution? I looked everywhere in the service manual and couldn't find any detailed information on them...

03Z33

iTrader: (49)

I counted the number of teeth on the ring gear

I may be wrong, but my understanding of the wss is everytime the ring gear tooth passes the magnet it sends a signal to the processor. I am not sure of how the signal is sent... it's probably seen as a change in voltage.

KPierson

iTrader: (1)

Hehe thats a good way to do it!. The sensor has two wires, one positive (above 8vdc according to the service manual) and the other one would have to be the signal. When the magnet 'sees' the tooth it would output a ground (grounds through housing I would guess). If the teeth are equally spaced on and off then this would create a perfect square wave, which the sevice manual hints at, but never specifically states.

copba1t

iTrader: (2)

Hey where's the fun in it if the car does everything for you? Might as well put a webcam in it, have a computer do all the driving, and then just sit back and watch from home

KPierson

iTrader: (1)

Shaving .2 of a second off of your 60' time is fun!

03Z33

iTrader: (49)

check out this video and tell me it's not fun

http://www.racelogic.co.uk/traction/..._Oct20.03.mpeg

Z8M

iTrader: (3)

Check this one. Scroll down to "Flatshift". Price is great too.
http://www.tekniqauto.com/CDN/tsport/tsport.htm

KPierson

iTrader: (1)

Those are both pretty cool products, and the prices are very nice.

AboostedG

iTrader: (6)

I really hope you pull this off KP. This is something I would get and hope you have a nice Group Buy.

G3po

iTrader: (8)

I note one reason these higher end solutions do not take advantage of the TB control circuit is that the implementation between vendors on Drive by Wire are very differnet and it would be difficult to make one product fits all. With the VQ35 app "only" in mind, you can possible simplify the solution dramatically. Another key important aspect is the TB actuation bandwidth. If it's too slow it may be difficult to avoid noticeble oscillation especially if wheel hop is present.

You may want to consider consolidateing this solution together with your VDC button memory module, then maybe you could use the OEM VDC button as a means of setting/arming the launch control with some form of beep feedbackma nd reducing overall cost and dash clutter.