Peugeot 2.0 Turbo on CardS12

[Matthijs]

As I mentioned in the introduction thread I h[*]ave a old Peugeot 309 converted to a 2.0 Turbo engine. Although the project is not really (or nearly...) finished it is always a good time to extend the scope of an already to large project

I'm software guy during the office hours with little to none eletronic knowledge so bare with me when I post my first schematics

The goal is to learn some electronics by doing, reading and stealing some good ideas from the other projects around here, just to know what I'm actually doing when controlling an engine.

Setting up this diary will encourage me to actually doing some work instead of lurking only (as I read a request for lurkers to come out of the closet and start posting... )

Ok, the project in a nuttshell:

The car, Peugeot 309 sort-of-GTI. I won't bother you with the endless list of mods as no-one really cares except for the engine part


First time running on stock ecu:
http://www.youtube.com/watch?v=ozNAeJftLz8&hd=1

So, on with the specs:
2.0L Turbo Engine from a Citroen XM. 125k km's or so, pretty healthy engine. Standard output: 150hp, 235NM at 0.5 bar or so. Can/should produce 200hp, ~300ish NM at 1.0 bar (max of the T25 turbo)
Code: Xu10j2TE, latest model from '99
60-2 flywheel VR trigger (why is this only usefull for 2500 rpm or so? Or I'm I miss reading something)
MAP sensor
intake temp sensor
TPS
wasted spark setup
PWM idle valve
coolant/head temp sensor
knock sensor

I've got a CardS12 with BDM, Codewarrior, and trying to order some (generic-) parts while reading other peoples build threads that can become usefull like 1k and 1.6k's resistors as I might need a lot of them, some Lm2940's for the powerplant, LM1815 for the crack sensor, TC4427 for driving the OEM ignition module wasted spark (from the top of my head: 0.227.100.201 or 0.227.100.200), 4 x VND14NV04 for driving the injectors, wiring for full seq. but using sem-seq. for now, some 1n5818 zeners for making the inputs monkey proof, 74HC367 and 74HC86 's, blabla. Still need to order some capacitors, opto decouplers and a spartan wideband. I'm really low on funds right now so I try to do everything on the budget which means some sampling, some waiting (ebay china stuff), some hussling.
Should keep me busy for a while figuring out what goes where and revising my schematics for a effort or 10 ...

Small steps are steps too:


Next step is building the power circuit I think.

For those with to much spare time or a boring job:
http://www.mijnalbum.nl/Album=CMSHVSKR <-- project album
http://www.youtube.com/user/Matthijs309 <-- some short clips about my projects I like to work on

[Dan]

Use a max9926 ic instead of the lm1815 ic for the vr sensor

[Matthijs]

Can you explain why one would be better than the other?

[Fred]

In simple terms: One sucks. :-)

Will read/reply more later.

Fred.

[Fred]

60-2 flywheel VR trigger (why is this only usefull for 2500 rpm or so? Or I'm I miss reading something)

Because my code is too fussy and/or too inefficient. It'll be pretty easy to provide an alternative implementation that's less fussy/inefficient, but I'd like to fix the current one, too/instead.

[Matthijs]

haha, ok, sampled some MAX9924UAUB/V+ and MAX9926UAEE/V+ as I couldn't find any at ebay =].

Regarding the 60-2: as long as it is only a coding issue it can be solved. I was afraid that it was something else I've overlooked

[Fred]

The 9924 is TINY, and even for the 9926, I'd recommend a break out board. I would recommend a BrickRPM, but I'm useless and don't have any to sell you. Sorry :-)

And yeah, the missing tooth stuff can be solved, for sure.

[Dan]

I have a brickrpm i could mail you?

[Matthijs]

That is kind of you Dan, I will keep it in mind as an option if I can't make one myself

I was doing some homework on the power supply and focused mainly on the protection part of it.
These are the original ecu's the engine came with (older and newer model)






I have little or no clue how it is designed.
2 times the big gray resistor (47 ohms)(current limitor in the input i guess, sub 300mA each, or limiting the output towards the wasted spark?),2 times the big red 100nF caps.
There is a huge 5z27 power surge protector.

My plan was to use 2 power circuits using a LM2940CT, each with a 22uF and 47uF as buffers and a big 2200uF on the input, an 1n4004(or greater) on the input, perhaps some extra smaller nF caps as filters.
Is it an idea to put a big zener or so on the input (25V, 3W or so) , somewhat like the 5z27 idea on the original ecu?

How about the topology of the vero board? Should I place the power circuit(s) near the CardS12? Or only the one feadingthe CardS12 near the Card, and the one feading all ADC and injection/ignition stuff further away?
Any ideas about that part also? Like ADC on the left of the CardS12 and the heavy ignition and injection stuff on the right?(or the other way around if it appears to be more convenient)

[Fred]

Interesting to see a MAP sensor internally mounted on an OEM ECU! :-)

For a DIY setup, you can skip the protection stuff for the time being. You'll likely do something less DIY later, and if you don't have your battery cables fall off in the mean time, it's likely OK.

Thousands of some other crappy ECUs run with almost no protection and no problems, just for example.

Don't bother with the 2200uF unit, it's not necessary or desirable.

I'm interested in the OEM power supply design, and the purpose of the 47ohm resistors. If you can trace those out and figure out what they do and how it works, I'd be interested in knowing.

For a DIY board I would use a single lm2940 regulator for both things, it's simpler and in some ways better.

Where you place the power supply isn't hugely important. How you connect it definitely is. Where you place inputs and outputs and how you connect them is absolutely important.

Fred.