Announcement

**yunvmyegt** · 07-30-2011, 02:47 PM

i no help here, but its nice to see someone actually getting this far into electronics, most people are way to scared...lol...

**jay** · 07-30-2011, 03:43 PM

I've been trying to make sense of the problems you're having and it just ain't comin. I'm nearly sure the ecu just does batch fire. So you should be getting 4 injector pulses, or at best 2 pairs / revolution. Maybe that's what you are getting and I'm just not reading carefully enough, but in any case that's all you should be getting. If you have some injectors firing, and some not, and swapping injectors around doesn't change anything and you have continuity from the injectors to the ecu, then by golly you have one or more fried injector driver circuits. If it turns out the ecu is bad there's no profit in fixing it. You wanna turn up the boost, get rid of the afm, and run a standalone anyhow.
It's been too long since I've messed with the stock box, but are you certain you have the injector pairs right? I'm pretty sure mine are set up 1-4 and 2-3, but would have to go check again, and maybe it's cuz my ecu fires sequential....

**gofanu** · 07-30-2011, 04:18 PM

CC's scope pic is:
"92 mazda protege 1.8L SOHC "
That is in fact a BPE, but is most commonly referred to as a B8 or 1.8 SOHC, NOT as a "BP"
The 1.8 DOHC engine is a BPD, commonly called "BP"

the turbo engines have a "T" suffix, as "BPT" or "B6T". There are sometimes other letters, but these are basic and universal, I think.

It is extremely unlikely that the firing order is different between any of these engines. Firing order is determined by basic crankshaft geometry, and there is little choice. Mostly the only 4 cyl engines that are different are those few with reverse rotation, which does not include any of the ones we are dealing with.

Possibly related, and most puzzling: I came across a Mazda TSB that related a case of a code of "No G signal", but the distributor checked OK. It said that it was found that an aftermarket cap & rotor had been fitted, and when replaced by a factory cap and rotor, the G signal came back. No explanation was offered, but that was from Mazda. I do not understand the mechanism here, but it indicates that something about the cap influences distributor signals, possibly at the ECU. I can only see this as some kind of HT leakage or inductive phenomenon interfering with "G". Is it possible that the ECU is filtering out a spurious signal, and also taking out the second injector signal as it does so? This may relate to the plethora of caps, rotors, and distributors that I mentioned finding in parts listings. Also to the matter of different rotor fitting positions and your diagram of rotor geometry at TDC, which seems wrong to me. Rotor orientation has to be taken relative to cap contacts, such that at any possible spark timing the rotor is pointed at the right contact. For most engines, this would cover, at minimum, the range from 40BTDC to TDC. As I said before "The one (rotor) I'm looking at appears to be about 30deg dist or 60 deg crank in effective width, so that's OK if it starts out in the correct place."

Q: Are your plug wires fitted as shown in the WSM for your engine, and does the rotor position match any data in that book? Given that there are 4 Ne and therefore 4 spark events per cam rotation, but maybe three possible rotor positions as with the BP distributor I have in hand, the whole setup acts like a vernier which would alter the relationship of the rotor to the cap contacts and the signal generator. Different caps and distributor signal generators might give even more possibilities.

FRM

**gofanu** · 07-30-2011, 04:27 PM

Jay-
At least on non turbo engines, injectors are in fact paired 1/3 & 2/4.
Al has all injectors firing, but only once per two crank revolutions, or once per cam revolution. So it seems he is missing a set of injector pulses that should appear halfway between the ones he has. This appears to be an ECU fault since all injectors do fire.

FRM

**jay** · 07-30-2011, 06:17 PM

Yeah I ain't certain of turbo inj. pairing, it's been to long since I messed with my car in that area. So yeah he is looking at a bad inj. driver if all are firing but not often enough. Not worth trying to fix. I mean sure, go ahead and look at the mainboard through a magnifying glass and look for a cracked solder joint, specially where the opamps are plugged in, ya might get lucky and be able to touch up with a smalltip soldergun, but I won't be surprised if one o' them op amps gave up the ghost. Does the board smell like some smoke got let out of it?

**sano86323** · 07-30-2011, 06:22 PM

you are right about the meaning(less) of the rotor pos in my geo drawing... needed to include a cap reference which I didnt. Still dont get the valve timing part, but I just need to do more work on that...

I dont understand why you ask the Q. The engine runs reasonably well... when we couldnt keep it running I was looking into causes and ran across the ECU anomaly, and so I am still perusing it. We havent driven the car enough yet to know if all is well... and its a new set-up so we dont really know just how much power it should have... eventually it should show up on the plugs I guess....

The wires are per the firing order and ign timing check says the cap is in the right place.

really puzzled at this point. I cant see how the scope could be missing those pulses, or, conversely, how the engine could run as well as it is with them absent. I cant stand having something like this "out there". Drives me nuts.

For the scope to be missing the pulses they would have to be much shorter/lower voltage and either of these would be similar as them being absent as far as the engine is concerned... even if I couldnt see the traces the scope would be triggering on them and I would see jitter on the G trace... and no, there is no way I am screwing up on how to use the scope. I built it back in my soph. year in collage and have been using it ever since.... well, there is always a way... but I cant see how... and I have been looking.

FRM, just how in the hell could the thing be running decently (driving around town) with this problem? It is in a little 86 323 DX hatch which was meant to have a 85 hp engine in it, so there is lots of room to spare as far as power goes... but should this problem be showing up in some noticable way?

**jay** · 07-30-2011, 07:34 PM

Oh now waitaminute. I musta missed that bit. If it runs and drives and ain't misfiring, then it's OK. I'm no expert but I've never heard of a bad ecu to injector signal that still allows the car to run on all 4. yeah go ahead and drive it, specially if you can monitor A/F, and read the plugs.
There are a whole bunch of more likely reasons for the car to start, run but not keep running... best of luck, I hate problems like this.

**gofanu** · 07-30-2011, 08:59 PM

OK. you never told us what you did to make it run??

I expect the problem will show up as lack of fuel on heavy load/high RPM. Should act just like any other car running out of gas under those conditions. Until that point, the ECU will adjust injector pulse width to make O2 happy, but it will likely run out of available time and everything will go lean. Not good under boost!

We know that there is only one G signal per cam rotation, so that could trigger one set of injector pulses, but the second set would have to be interpolated by the ECU. Or, it could just use the Ne signal to give two sets of injector pulses, triggering the pairs alternately, which is evidently what the SOHC cars without the G signal do. Or, it may use the G as index and then work off Ne, in order to relate injectors to valve events; that is in agreement with the statement that G failure gives a default of "eliminating alternate pair injection".

There is something in the ECU called the "fuel cut control circuit", used to "improve fuel economy, prevent engine bucking under deceleration, and prevent engine overspeeding". This takes a Mazda SST to test but it appears to check pairs 1/3 & 2/4 separately. Since this might be different for **T engines, look it up in your book. Seems to me that this is maybe permanently (partially) triggered in your case.

Possibly now in the wrong thread but:

Well, the Chilton book shows #1 wire as down and back for all 4 cyl engines, the Haynes showed #1 as down and forward, and:
I looked at 2 BP, #1 wire tower is straight up, and B6 & B8 #1 is slightly forward of straight up - all 4 run fine. If there were only one possible rotor position on the shaft, all would be well. However, I have one rotor that fits in three positions, so the rotor can be indexed in 120deg (dist) steps, while the wires can be indexed in 90 deg (dist) steps. If you have #1 plugged into a different tower, the rotor would have to be moved to make the spark come out on #1 wire, but then the rotor would be out of phase with the spark and the cap terminal and the spark could start jumping around inside the cap, causing all sorts of odd signals. If the various rotors/caps/distributors have different phasing, there might be more possibilities. Curious point: I have a BP & a B6 dist here, and both rotors fit on the BP one in all three positions, but both only fit the B6 in one position, though I have no doubt some fool could force it. I can see no difference in the shafts without setting up some precision measuring stuff.

FRM

**JB2sowner** · 07-30-2011, 10:43 PM

Its me... The guy that has been posting under his son's (sano) account.

I havent told you what we did to get it running cause it is too embarrassing how long it took me to find. Flex coupling to lower intercooler inlet was not sealed... It was pushed on enough to look like it was and the clamp was tight on the coupling... hard to see and harder to get to... so I just looked at it and said ok. We found it with starter fluid spray. It really did look like it was on there and tight.

I was suspecting that the O2 loop might be opening up the injector "dwell time" to try to compensate, but that was only a theory in my head... I dont know enough to tell if it actually could happen. Having you through it out there makes me take it a bit more seriously.

On the rotor and distributor positions... As one should, I noted both before messing with any of it... so I actually know what they were when we got the front clip... No 1 is where you describe, and the rotor aligns with the the TDC mark on the exhaust cam sprocket. We have owned the car (that we have put the BPT in) forever... Wife bought it new in 86. I started driving it around 18 years ago... it had a B6 in it, and that rotor would only go on one way (if I am remembering right.) I am pretty confident I do not have a plug wire, rotor orientation, or distributor orientation (ign timing) issues.

I tried to post the question on the MS (ecu replacement vendor) site but the damn thing said I had to wait for a moderator to approve the msg... which probably wont happen until Monday. P*sses me off.

I traced through the NE, G input circuits and the 1/3 and 2/4 output circuits about a week ago. I found the output transistors (2) and the driver(buffer) chip... but then we got it running and the son wanted it in his car for some reason..

To go further with the debugging I planed on making a very simple car simulator on my work bench... which is why I started looking at the dist. geo... I need to know the location of the G relative to the NEs so I could generate them with enough fidelity that the ECU would not barf... then I can check the buffer chip... I guess now that we have it running I could do it in the car... Wish I had written down that part number for the buffer.... that was the plan, but I keep getting distracted by don wanting me to help fix other stuff or by him actually wanting to use it.

I have see some stuff on fuel cut and dismissed it as a possibly cause long ago... but cant remember any details... will definitively go check it again.

I also want to force the ECU into limp mode and see if the injectors start firing simultaneously... should see 4 pulses (total) when I really should be seeing 8 per cam rev.

**jay** · 07-30-2011, 10:58 PM

"Flex coupling to lower intercooler inlet was not sealed."

I bet most everyone here has that Tshirt. It wouldn't hurt to make a tool that lets you pressurize the whole intake tract and see if you have any other hidden eastereggs.

**gofanu** · 07-31-2011, 12:12 AM

No hiding dirty laundry!

"I was suspecting that the O2 loop might be opening up the injector "dwell time" to try to compensate, but that was only a theory in my head... I dont know enough to tell if it actually could happen. "
Open loop reads basic info - Airflow, rpm, TPS etc and looks up a table. Closed loop uses O2 to modulate that info for better fueling, so if it reads lean the ECU widens the injector pulse. I repeat that the pulse is the little low blip on the injector traces, when the injectors are switched to ground. If your scope can keep up, you should be able to see them widen as you snap the throttle open. It needs a lot of fuel to accelerate, just like the accel pump in a carburetted engine.

"I need to know the location of the G relative to the NEs "
Well now you know from the traces.

Fuel cut is likely a "soft cut", not a complete & abrupt shutoff. It would be staged in some way, maybe several steps. One way (stage) to do that would be to eliminate half the injector pulses, which is pretty well what you have now.

My feeling is that the injector drivers are OK (since both pairs work), it is the logic unit that triggers them that is deranged.

"force the ECU into limp mode "
For this purpose, all you need to do is interrupt the G signal, preferably by unpinning it someplace, but there's always wire cutters!
Trouble is, you only have 2 pulses (1 per pair) per cam rev (=4Ne)now, and you'd still have 2 but maybe with all 4 injectors firing together. But it is possible that "limp" should have 4 pulses with all four firing together.

Another random thought: I have found that TPS anomalies cause all kinds of bizarre behavior, because they scramble the logic unit(s). Apparently, if the TPS readings do not make sense, the ECU tries to force them to do so by tweaking other things it can control. Check TPS signal output very carefully, looking for dropouts and jumps and erratic readings as you move the throttle slooowwwly through the range, several times up and down. Worst places are usually right around idle and at cruise conditions. Had a Taurus that was very weird, and read lots of accounts of people including dealers with computers going nuts trying to fix them, right up to selling or scrapping the car after thousands of $$ spent. A month's research plus 5 minutes with a meter and a $28 TPS cured it, and its successor Taurus, and then both my 323 stick and auto (I cleaned the Mazda units since they were not so bad and cost much more).

FRM

**JB2sowner** · 07-31-2011, 12:58 AM

I got that the output transistors for the two injector signals are working... I was (correctly???) remembering that I had trace there input back to a uPC1771C op amp chip and thinking that the op amps might be being used to combine the signals... looking back at my notes I am pretty sure that was wrong, and that the output transistors are being driven (through a bit of passive circuitry) by the uP. The op amp chip was used for sensor input, in particular, the G and NE signals. My memory sucks these days.

BTW, the damn thing is a multilayer board. I hate trying to trace things on them. Some of the stuff we work with at work has 9 layers on it. I have no idea what modern computer boards are up to these days... I remember when the first 3 layer came out...

Pulling the O2 doesnt require cutting a wire... which was why I chose it.... you doubt it will put it into injector limp mode I take it. Why cant things be simple.

"I repeat...". Got it this time.... of course it is that part of the trace... hadn't really though about it before.... just counting them.

At least I can get a duration easily from that part of the wave form. Thanks for forcing it trough my thick skull.

"Trouble is..." Not following what the trouble is. I have now
1: X - - -
2: - - X -
Limp mode should get me
1: X - X -
2: X - X -
What I think is correct (but am trying to get some one to confirm is):
1: X X - -
2: - - X X
And yes, I think that limp mode should be:
1: X X X X
2: X X X X

Hope that is not to cryptic... left to right is passage through 1 cam rotation. Up and down are injector 1/3 signal (1

and 2/4 signal (2

.

I checked the TPS twice but only as far as the manual says to go... Will look harder at it per your recommendation.... I would not have thought of this as a cause.

Ever take apart our AFM? the parable resister that is in there is a very very strange creature... took the one from the B6 apart about 2 decades ago to see what the heck was going on.

**JB2sowner** · 07-31-2011, 01:59 AM

Read through the stuff on fuel cut again... the testing section is cryptic as hell, and I dont get it.

The theory section (Dechoke control system) is much clearer, but only seems to address fuel cut as it is used to unflood an engine... Which I once again thought was stupid as I have every time I have read it... The conditions for fuel cut here are
1) ign=start
2) rpm < 500
3) throttle all the way open
4) coolent temp BELOW freezing
WTF?

Anyway, when I read it before I thought the fuel cut function was only functional when the above conditions were met.... but reading the "fuel cut control system" test procedure makes clear I was wrong on this... but they sure as hell dont give me enough info to do anything useful with.

Will spend some more time on it. I agree with you that this could be the source of the problem WITHIN the ecu.

**gofanu** · 07-31-2011, 03:44 AM

I am not an electronics guy. I have a fair idea of the basic principles of operation, and of what they can theoretically do, but the details are beyond me. I frequently give them more credit than they deserve, that is they are more primitive than my basic expectations. I do know pretty much everything about what an engine needs, wants, or can live with, and I know that the electronics have to provide it. In order to do that, they need input, and I am fair at figuring out where the input comes from, and what general logic might be used to do the job, but the logic is often beyond what the electronics can do in a simple way, so there are big gaps between me and "IT".

Pulling the O2 will put it in open loop, but may or may not do anything with the injector firing. "the (WSM) statement that G failure gives a fail safe of "cancel two group injection" is one of the few unambiguous things we have, so I think it should be followed, though you might want to check that your book agrees (I corrected the quote). I suspect there are a bunch of different default states that people refer to (sloppily) as "limp mode". There are many "fail safe" conditions in the "self diagnosis function" section.

Without knowing what 'injector limp' is supposed to do, or on what basis, I don't think you can make any useful predictions. It gets you into expecting things, and you then don't see what is happening. I can think of arguments for several different states. maybe depending on what other data it is getting. Try it with no G.

TPS very critical, and much underrated in all literature I've seen; it is possibly the most basic control parameter, since it tells the ECU what the driver wants. False/erratic info is just like trying to please somebody who can't make up their mind. Said Taurus started out with a strange high idle at part warm (sometimes), then acted like you had your foot on the gas going downhill (sometimes). Fixed the TPS and mileage went from 27 to 32 on same usage, power went up about the same, and the miserable shift points on the autobox all straightened out - and there was no direct & obvious connection between the engine and trans controllers (pre electronic trans control)
The ECU effectively reads both a voltage and a rate of change of voltage off the TPS, so dropouts cause low voltage with relatively much higher V on either side, leading to incomprehensible rate of change, sometimes with reversed sense of direction of change. So if the throttle is moving such that voltage is dropping, but there is a high resistance or open section, voltage drops precipitously, but then increases equally quickly once it makes contact again. What's a poor ECU to think? And the big point is that in an electromechanical device, greatest wear is where the thing spends the most time in a small area = idle & cruise.

Never took an AFM apart. What is a "parable resister"??

In my on line/CDR WSM, the fuel cut test procedure is completely based on a SST "Engine Signal Monitor and Adapter", so is useless. But, it shows two lamps that flash alternately, red and green but of course my copy hasn't got colour. One is labeled 0-6V, the other 0-20V, so I think it just shows the injectors switching ON/OFF. You switch it between the terminals for each injector pair in the test.
It seems that it shuts the injectors OFF on decel above 1900rpm, and limits DOHC to 7300 and SOHC to 6300.

FRM

Announcement

BPT(GTR) ECU injector signal (crazycanadian/gofanu)

BPT(GTR) ECU injector signal (crazycanadian/gofanu)

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment