1987 AMC Renault GTA 2.0 - Gone to a new home.

Quote of the day that!!!! Had me laughing that one.

Great save on the car as well, will be watching this one.

More squiggly lines!

I made an addition to the circuit tonight to add an inverter- basically now what it does could be achieved with a bare piece of wire. Put a voltage to one end (kinda) and a voltage comes out the other.



A bit more complex- needs a couple of capacitors scattered here and there.



Stuck the same 01010101010101 into the front of it. the top trace is by the green light, the bottom is by the red- one is upside down from the other. the rise time of the bottom one isn't as good as it could be, but that is still within good limits.

To prove how well the circuit works my oscilloscope has a nifty mode to add together the two traces. Ideally if 1 is present the other should present a 0- if you add the two together you get the net of zero.

This line should be flat, in an ideal world.



The upwards spikes are very acceptable- the downward ones aren't quite as nice, but I don't have a capacitor in the output of the second circuit, which would help with that.

So, with this working, I stuff about 40 Megabytes of data through it at four times the frequency I'm going to be using it at and it performs flawlessly, on the computer.

I hook it up to the car and... squat. I get allsorts of junk data.

I think I have the wrong value resistor in the pull-up circuit. I stuck a bigger one in to try and help. Tomorrow I may b able to go down to the car again and try it out.

I'm a bit worried though as there was a random bolt on the floor and some antifreeze had leaked out.

More hunting, I guess.

--Phil

Stopped at Radio shack yesterday and spend a few dollars on some odds and ends:



A few bags of assorted ceramic and electrolytic capacitors.



Started plugging them into the circuit in critical places to tidy up the waveform.



Set it all up and started throwing data through it. The dog helped.



It produces a nice clean square wave from a dirty input trace now. Might get to go try it out this afternoon. Stuck on-call working again. Bah.

--Phil

I'm tethered to the computer table so the dog's basket got repurposed as an impromptu table. the computer is sat to the right and the mass of wires in one place is due to the fact that the outlet is 3" away from the device but the cords are all 4-6' long...

I was picking up a lot of interference from the car and it was mangling the output- that circuit is intended to stabilize the output back into what the RS232 is expecting. I could probably get some decent data by hooking into the first stage of the circuit and bypassing the rest.

I'm going to scope it if tomorrow is quiet enough for me to leave.

I want to get to the inner workings of the fuel injection computer. That and I have a feeling if I figure this out I can create a smaller stand-alone device to carry in the car with me (think Arduino).

That's the plan.

Before then I need to figure out some more basic stuff like where the bolt that fell out came from and why it barfed antifreeze onto the floor.

--Phil

That is a problem with color banded things, at least the other items have the numbers printed on them.

It's been a while since I've dabbled in digital electronics- I'm missing most of my books from college too, they are all back in the UK. It's slow going compared to how some other people work with electronics, but this is all research and development for me

Still, this is what I find fun. That's why the thread has turned into computer and electronics class. I forget how many pages it's been without any pictures of the car...

--Phil

Sat at the computer the past couple days plugging away at numbers.

Pulled up some old data, compared it to that removed from a Jeep with the same ECU. Got a spreadsheet built that calculates that individual lines, and a graph that makes a bit of sense.



That was back from when the engine was not running properly, as the fuel pump wasn't working but I started it on carb cleaner sprayed down its neck- there was a similar picture a bunch of pages back.

This one has only a few parameters on it though. (Manifold pressure, dark blue line, top)(Ignition advance, green line)(RPM, light blue, bottom)(knock sensor, pink line)

It at least shows the ECU is doing what it should have been at that point- the MAP sensor detects the beginnings of a vacuum when the engine fires and dies again. The knock sensor gives some decent readings in relation to the engine RPM and the ignition does back off when the knock sensor spikes (left hand side). the green line is a bit spiky because the data was a bit corrupt.

Now I have to figure why the most recent data I have is all screwed up. The temperature sensors are being reported by the ECU as -20C, which isn't right. It does however increase slowly as the engine was warming up so I'm thinking something has got the ADC all messed up- it was also reporting the engine speed as a dead 545 RPM. I've got today (Thanksgiving, y'all) off, tomorrow and the weekend to work on it.

I might even get it figured out. Hopefully something as simple as bad contacts (I'm thinking I might know where) or a bad ground.

On the positive side, it does have a memory of sorts- and it can report if it has a failed sensor or a sensor isn't producing the expected results. Nice!

--Phil

It's both the intake air temp and the coolant temp sensor- they both come off the same chip inside the computer, and they both have the same skew (about 50 degrees) and the other readings are similarly messed up.
I think the common ground point on the back of the block it uses for the sensors may be either loose or dirty, or the B+ relay that powers and latches the power to the ECU is screwy.

I can't really bring the computer today but I'm going to bring my multimeter and check some resistances and voltages.

--Phil

Drove to the house, cleared the junk from around the car, moved it up the driveway, unpacked evreything from the truck, plugged it all in and the cord to the laptop broke, so it wouldn't power up any more.

Disgusted, I packed it all up, went to the store, bought food for dinner and a couple beers. No photos because today was lame.

Tomorrow, if it isn't raining, I am going to bring the wife's little laptop. It's due to be cooler too, which will be welcome.
I have been having to use the screen defrost and rear window heater in the truck at night (23 degrees C out!) because it's been so humid.

Edit: I took a look on Radio shack's website and to my disgust they don't stock Optoisolators. Even Maplin sells them. I'm also looking at the bias on the circuit. I'm doing something wrong. The transistor is getting unhappy the way I'm driving it. Circuit revamp again.

--Phil

I can see two elephants!
Keep up the hard work man!