Purs like a kitten...

[sblake01]

There are no fuel pump contacts in a 78 (or later) AFM.

[FastWoman]

Hmmmmm.... Are you SURE about that, Steve? I ask because mine does. Furthermore the connector to the AFM (which is almost certainly original) is wide enough for those extra pins, although nothing connects). I'm wondering whether I have the original and correct AFM. So again, are you SURE?

[cozye]

There are fuel pump contacts in the 78 AFM. I've got two of them. There are not however any wires from those contacts to pins in the connector or the harness. So effectively, it has no contacts.

The AFM should be A31 604 000, at least both of mine are

[Zed Head]

There are fuel pump contacts in the 78 AFM. I've got two of them. There are not however any wires from those contacts to pins in the connector or the harness. So effectively, it has no contacts.

The AFM should be A31 604 000, at least both of mine are

This picture tells the story. From the 1978 wiring diagram.

[FastWoman]

Phew!

I was briefly seeing dollar signs.

[cozye]

Phew!

I was briefly seeing dollar signs.

Yeah, but if you are like me on this problem you'd be willing to spend $250 to solve it 100%

[sblake01]

No pins on the connector for them in either the harness or on the side of the AFM itself. Same as my 79 810. If the AFM in your 78 has them (the pins), it's from a 77 or earlier.

Edited November 13, 2010 by sblake01

[sblake01]

Now that I look back at post #48, I see that Sarah actually said pretty much the same thing I'm saying. So allow me to apologize since all I've really done is muddy things up.:stupid: Carry on!

Edited November 13, 2010 by sblake01

[FastWoman]

No problem, Steve! Hey, if we don't bring up these things (even if we're mistaken), we can miss possible problems/solutions.

While I was sick, I hit the computer a bit and satisfied myself that I indeed have the correct injectors for the '78 Z. I decided not to flow test the injectors after all. Good grief, there'd be so much stuff to take apart and put back together, and I need my life back.

I'm now feeling 95%. No more virus, anyway. I thought I was courting a secondary bacterial infection, but that seems to have cleared. Just lots of gunk, but that, too, shall pass.

... so I got back to work on the Z.

I was first curious as to how much "boost" I would need in fuel pressure to achieve the proper mix at idle. (I was still playing with the idea of an adjustable fuel pressure regulator.) It turns out that with the AFM recalibrated per Atlantic Z's specified spring tension, I'd probably need about 60 psi. I was getting pretty close to the right range at 50 psi relative to manifold vacuum, where I topped the system out. My pump will only put out 42, and with the vacuum added, that would put me around 50. Of course there's no way I'd be able to put out even 50 at zero manifold vacuum. Anyway, this solution is out of the question, without going to a new fuel pump and pressure regulator combo. I'm not going there.

Then I said, "OK, let's figure this thing out." I cracked open the kick panel and tested out my spare ECU. Same lean idle. I then did the quick FSM check of the ECU with a miniature Xmas light bulb in the #1 and #4 injector connectors. They flashed once per rotation, as they were supposed to, and they did flash brighter when I unplugged the temp sensor. I also re-checked all of the connectivity at the main ECU connector, and everything was happy and as it should be.

I also opened up one of the ECUs for a look. It honestly didn't look bad. I was rather put off by the old can-type transistors. Anything oil-filled can possibly leak its oil over time, so I was leary. More surprisingly, even the early ICs were canned in oil! I've never seen that. Anyway, I suspect there are common failure modes in these ECUs. I do know enough to know that transistors rarely fail catastrophically. They usually fade away over time, and two ECUs of the same design would be likely to fade away in the same manner. I did an exhaustive/ing search on the internet for any info about MTBF in these sorts of transistors, common failure modes, etc. I really couldn't find much at all. I take back any previous assertions that one can find ANYTHING on the internet. Apparently those can-type transistors are too prehistoric (like me).

Anyway, from what I can gather, these ECUs don't really have many failure issues at this point in their lives. I'm still somewhat leary of all that oil-filled can stuff, but I'll just hang with what I've got, with the expectation that it might work, but still develop inaccuracies over the years. In other words, my current operating theory is that the ECU will be responsive to changes in sensor readings, but otherwise a bit out of whack in its overall response to those readings. That is, some adjustments may be necessary.

I'll mention here that I'm rejecting the Atlantic Z approach to solving this problem at the AFM. That's because I would have to alter the spring tension so much that the vane would too easily peg out at wide-open, leaving the mix too lean at peak intake. That of course would not be good. I wouldn't mind tweaking by a tooth or two, but certainly not by 15 or so.

As I see it, the most obvious place to richen the mixture is with the coolant temp circuit. I was curious to see just how much an open sensor would richen the mix and was delighted to find that my poor engine was choking and gagging on fuel, barely able to run. Plenty of adjustment room there! I'll note here that the PO or his mechanic had inserted a resistor in series with the sensor, I presume to richen the mix. I suppose this is a time-honored bandaid approach. I had removed the resistor when I rewired, refreshed, replaced, etc.

Anyway, I ran upstairs and grabbed a 5k variable resistor and plugged it into the temp sensor connector. That let me tweak the mix manually. I found that the best mix was achieved in a warm engine with a resistance of 2,550 ohms, while the temp sensor's resistance had dropped to 240 ohms. Thus the mix could be richened about right by adding 2,310 ohms in series with the sensor. Interestingly, I think I remember the value of the previously added resistor being 2.2k, so apparently someone else had been down this road before me. I'm sure at one time it made the engine run very well, and then it got wonky again with age and deterioration.

Anyway, that's the solution I propose: I'm going to put a resistor in series with the temp sensor to richen the mixture. I'm going to try a more refined approach than simply splicing a fixed resistor into the wiring harness. Instead I'm going to install a mixture tuning potentiometer directly inside the ECU housing, provided there's some good place to mount it. Before finalizing the mount, though, I think I'll run a couple of loose wires into the cabin of the car, so that I can play with mixture with the car in motion. I'll have my lovely assistant drive my car down the freeway, with her foot blocked against the transmission tunnel to keep a constant throttle opening, and then I'll tweak the mix to achieve the maxiumum speed. I'll note the resistance, and then set that value when I mount the pot back inside the ECU. Further tuning will be done on the basis of plug readings and mileage calculations.

Oh, and when I'm done, I think I do probably need a couple more deg of advance in my timing (which is currently set to 10 deg BTDC, per factory specs. I'll probably do that too, and I expect that will tweak me to a perfectly healthy engine vacuum. I just have to fix my timing light first.

So there... That's my plan! Any thoughts or comments?

BTW, I'm very bad about taking all the pics I should (ironic for a photographer, eh?), but I'll make sure to document my ECU mod.

Edited November 14, 2010 by FastWoman

[Zed Head]

If you haven't read these "tips" from Paul Rusch of Rusch Motorsports, they are worth a look. He goes in to detail on how the sensors work and has some good examples of experiments he ran in the past.

http://forums.hybridz.org/index.php/topic/95316-braaps-l6-efi-induction-advice-and-tips/

[FastWoman]

Hey Zed, thanks for the link! Very useful info. He has basically the same approach to what I'll be taking, except that I've now learned the AFM pegs at 4500 RPM, and fuel delivery after that is based on RPM and other factors. Per his advice, I'll be adjusting the mix for WOT first and then tweaking the AFM for smooth running at cruising. Finally I'll tweak the almost useless little bypass screw on the AFM to get the idle right. Good to have these refinements in mind. Thanks!

[cozye]

Interesting that you are arriving at the same conclusions I have. I re flowed all the solder joints on the entire ECU last night and it didn't help. I too was wondering if degradation of the analog ecu components were to blame. I had thought of wiring a potentiometer in line with the water temp circuit, but worried that it was too broad of a fix and might make the WOT too rich.

I can tell you that I've continued to drive, and tweak on mine. I recently adjusted the idle air mixture, I put it three turns in (mine was 6.5 turns out from full rich stock). This with the 6 teeth advance and the timing at 13 is working pretty good for me. It's not 100% and I'd like to find a better work around, but it's close.