Saar

ok, not quite thirdgen related unless someone wants to play with stuff that uses the 60V6 DIS....

but i'm having a hard time figuring out 1 thing mainly: from what i understand, the 6+1 reluctor signal is read by the ICM, then the ICM stores the pulses and sends out a 3X signal to the ECM(much like the ECM storing, dividing and sending VSS pulses for the instrument cluster?).

ok, now here's the confusing part: how does the ECM know which of the 3 pulses signifies the #1 piston being at TDC? with the later PCMs that use a cam sensor and a 24X crank sensor as well, i can see how, but using the 3X signal alone, it's confusing me. and since the ICM is commanded via the EST line as to how much advance thereshould be, i don't see a problem, it's when it comes to the injectors is what gets me.

does the ECM just fire the injectors on every 3rd reference pulse? it seems to me, that's the only way it can operate in batch fire since with a 3X pulse, any of the 3 could be the notch closest to the +1 notch.

i'm trying to play with a later PCM as well, but i'm having trouble figuring out a way to synchronize the 24X crank and 1X cam sensor with the 3X... unless there is some type of pulse divider i can make and just generate a 24X signal via a 555, and divide by 8 for a 3X and divide by 24 for the 1X?

otherwise, this is all drawn out on paper and in my head, just these details holding me back....

RBob

iTrader: (1)

On a non-sfi set up only the crank reference signal is required to the ECM. It doesn't care which cylinder is firing. That is up to the ICM to know. The ECM just uses the same four signals that are used in distributor set ups.

For example, the '93 f-body 3.4l engine is SFI and DIS. Has a crank sensor on a 7-tooth wheel, and a cam sensor. The ICM buffers the cam sensor and sends it to the ECM. But doesn't use it itself. The ECM uses for the SFI.

Can use the DIS system on the 3.4 with just the 7-tooth crank signal and interface it to any of the older 3rd gen ECMs. Need to change three timing set up parameters. But that's it.

For a test bench it is easier to use a micro controller such as a PIC. This way the proper signal phasing is assured. And it is much more stable then a 555.

RBob.

Saar

got an example of a good PIC for this kind of application? i'm clueless to them yet, and if "signal phasing" is what i think it is, that solves quite a few problems...

in case it matters, so far the only ECM/PCMs i'm planning on playing with for now are the 1227727 and 16196401, though i may move onto the 93-95 3100 and 93-95 3.4 f-body PCMs if/when i get an emulator and an EEPROM burner more capable than the BURN2.

RBob

iTrader: (1)

I've been using the PIC12F617-I/P. It is an 8-pin DIP, 6 I/O pins, four 10-bit ADC channels, internal 8MHz oscillator, and 3 timer/counters. And a decent amount of memory for look up tables.

Use one for each 'feature' of the bench. That is: VSS and VATS on one, O2 on another, DRP/CAM on another, and so on. At less then a buck apiece, makes development real easy.

RBob.

Saar

that one supports in-circuit serial programming according to a doc i ran across here.... at least it looks like i won't have to spring for another programmer yet. and it looks like 2K of EEPROM and 128 bytes of RAM. i'm not extremely fluent with the machine code yet, but that seems like overkill for something like this...

and let's see... a 24X signal at 6375RPM would be.... 153,000 pulses per minute, so 2550 per second. i don't see a specification for anything like that in the link, but judging by the rest of it, it definitely seems like it can hold it's own...

now, i just have to find me some schematics and the actual manual to it.

RBob

iTrader: (1)

Actually, it is 2Kwords of flash for program memory, 128 bytes of RAM for variables, and no EEPROM.

The data sheet is on Microchip's site. Try this link:

http://ww1.microchip.com/downloads/e...Doc/41302D.pdf

At 91-cents each in quantity of 10, they really can't be considered overkill. And with running the three timers for the three output signals, clock cycles start to count. Note that the chip can be run up to 20 MHz with an external crystal or oscillator.

Look at it this way. As you mentioned the 24x signal is 2.55 KHz. Need to double that for the number of interrupts per second. Or 5,100 times a second. It doesn't take long to eat up clock cycles.

Now, with 6 I/O pins:

three are used as outputs for the 3X, 24X, and cam signal.
use one as an ADC input from the POT for the engine speed input.
use a second ADC input from another POT for fine tuning of the engine speed (optional).
use an input as a 6/8 cylinder selection switch (optional, required for cam sync).

If need to go to a 20MHz clock speed. Skip either the 6/8 cylinder selection, or the fine tune POT for engine speed. To free up the CLK-IN pin for the oscillator.

RBob.

Saar

see, now if i knew how to do any of that, i'd be set, but i managed to find the manual for this PIC and read all 212 pages of it(the parts i could understand anyways), and i don't even know what to make of all of it.... hell, i'm still trying to understand all of the 68HC11 stuff, and while i feel that i've made progress recently, this is learning yet another language i shouldn't even be attempting.

i've looked at the company supplied schematics and saw things that made no sense to me at all. using the same pin for about 4 different operations? i even found the little tips/tricks document for it and read through things that seemed relevant, and it seemed to hurt more than help.

that being said, is there any other viable option out there that at least has a snowball's chance in hell that doesn't require messing with code? possibly a timer and/or pulse divider IC that's more reliable than a 555?

Saar

LONG time, no progress. for reference pulses, went a simpler route for now and am not emulating a cam signal, just the 3X reference(since as mentioned, i play with V6 a lot)

finally got to finishing the bench last night and was playing with it today. a lot less issues than i expected....

basically the only issues i'm having are with the 3 wire sensors... i threw (i think) 10K resistors in between the 5V input and the center terminal of a potentiometer, since if the 5V circuits aren't current limited, turning the pot all the way toward the grounded terminal would be a very bad situation.

and was playing with it again to get exact numbers for ideas here, and for some reason turning my A/C pressure sensor's pot effects TPS and MAP reported values.... a LOT. so now i guess i really need to take it apart again. i get the feeling i must have missed a resistor there? A/C pressure pot turned all the way in one direction made TPS and MAP go to their limits, turned the other way, back to the other limits...

otherwise, my only other issue is that if my 4K PPM VSS output circuit isn't connected to an instrument cluster(or i imagine a resistor dummy load could be substitued), it causes a lot of garbled data through the ALDL circuit. not sure if i need to wrap that circuit in foil or what, but it was odd to me.

oh, and my O2 sensor circuit isn't working at all, but considering i apparently goofed the A/C pressure, i should probably check the wiring of it first before assuming it's some other issue.

i had to disable DTC 42 (EST Monitor in A1), i'm not 100% certain on why, but i really don't care since it's not something i'll need to work around in a real vehicle.

Six_Shooter

iTrader: (2)

Just a clarification, the 3.4 DIS does not connect to the cam sensor.

Saar

yeah, i think RBob might have been thinking of the 3800s. those ran their cam sensors through the ICMs, but no 60V6 application did that i know of.

RBob

iTrader: (1)

Yes, the Buick stuff. Hard to keep track of all of the different systems.

For code 42, this is caused by not holding the EST line low when at low RPM (cranking, engine not running).

Use a PNP transistor (PN3906), with the collector to ground, emitter to the EST line, and the base to the BYPASS line via a 10K in series. Also place a 10K from the BYPASS line to ground.

This will cause the EST line to be clamped until the ECM releases the BYPASS to take over timing control. This transistor just gets added in, no opening of wires or such.

RBob.

Saar

as perfectly acceptable a solution as that is, i think i'll just disable the DTC. reason being it is already quite cramped inside the enclosure i'm using already (6" X 8" radioshack project box) and that will only make things even more painful when i need to open/close it. as far as i can tell, it doesn't effect code execution when disabled, and since i'm only running the EST circuit as a LED, it's not like i'll be potentially junking a motor or anything like that.

Saar

fixed the A/C pressure wiring, i had the 5V circuit with the 10K resistor connected to one of the outside terminals instead of the center one, now it will read a range of 22 to 213 A/D counts on that circuit and it doesn't effect the MAP or TPS readings.

however... MAP and TPS still seem to be quite limited in their range, TPS reads from 3.33 to 4.80 volts, which is a range of 170 to 245 A/D counts. MAP reads from 41.35 to 89.32 kPa, which is a range of 84 to 214 A/D counts...

any ideas as to why i can't achieve the full range or at least close to it? both pots have their 5V line limited by a 10K resistor and the pots themselves are 100K linear units. i get the feeling i have too much resistance? i think i have a spare TPS around that i can ohm test to see what kind of resistance the actual sensors are...

Saar

found the spare TPS... when throttle fully closed, i get roughly 5K between 2 combinations of terminals and 1.5K from another. throttle fully opened, same results, just with different pins.

i get the feeling i should have been using either a 10K or a 5K pot?

RBob

iTrader: (1)

For MAP & TPS I use a 10K linear pot. No other resistors, although a 1K in series with the wiper is OK.

For CTS & MAT I use a 100K with a 100 ohm in series. That is to prevent an instant DTC when turning the pot too far. Can use dual pots in series with the other being a 1K for fine tuning of the setting.

RBob.

Saar

yeah, i think i'm going to swap out the 10K resistors for some of the 1Ks i have...

it would seem my TPS test implies that there will be no less than 1K of resistance from any 1 terminal to the others in any situation, so it seems like the most proper course to me. if i need to afterwards, i'll scour up some 5K pots as well, since it seems that i'll never see above that in any situation either.

the dual pots for the non-linear sensors is interesting, but due to space constraints, i don't think it will be possible without finding a new enclosure. i may just have to live with the linear ones i have now or use logarithmic/reverse logarithmic to attempt to linearize them.

Saar

another update.... my pot wiring was all messed up... had the resistors between the sample circuits and the pots for some reason. fuel pump monitor is also working correctly, i had that one all messed up.

so now i have 1K resistors between the 5V reference wire and the pots for MAP and TPS, now TPS reads from 3.43 to 5.00 volts and MAP is 43.2 to 102.97kPa. A/C pressure also had a 1K installed, i think, and reads from 8 to 213 A/D counts.

so..... some improvement. i can at least hit(or come very close to) the 5 volt peaks for TPS and MAP now, but i still can't hit the lower ranges... in fact, my TPS and MAP minimums are even higher than they were with the borked wiring and different resistors.

and what really confuses me.... is why my A/C pressure sensor pot seems to be doing the relative opposite, in that it can almost hit the A/D minimum, but the upper range is off.

Six_Shooter

iTrader: (2)

Why are you using the 1K resistor at all?

Saar

based off of the schematic Ludis made for the 7749, it seems that the 5V reference circuit used for 3 wire sensors have no ECM-internal current limiting, so if i turned a pot all the way over to skew the sampled signal as low as possible, that would make a near 0 ohm connection to ground, which would likely be damaging to either the pot or the ECM or both.

or am i being overprotective?

RBob

iTrader: (1)

It sounds like the pots are wired wrong. One end goes to ground, other end goes to +5 volts. Center (wiper) goes to ECM. With a 10K pot, other then the ECM resistors, there will always be 10K between +5V and ground.

MAP input has a 51K to ground and the TPS input has a 220K to ground ('7730), nothing to worry about.

RBob.

Saar

the way they're currently wired:

one outside terminal to ground
center terminal(wiper?) is the +5V reference (with a 1K)
the other outside terminal to the ECM pin for A/D sampling

so from the sounds of it, i can remove the 1K resistors and move the wires around.

don't suppose anyone could explain a question that's been running through my mind for a while: the different resistor values to ground for various sensors? i get the feeling that it's somewhat for signal filtering, since there is also a cap in some of the A/D sampling circuits.

anyways, i ask since i converted a few of the unnecessary a/d circuits that feed 5V into the circuit for use with temp sensors into a grounded type circuit by moving the 1K/2K/4K resistors from being connected to a 5V rail and the sample circuit to ground and sample circuit. none of those have caps in their sample circuits, something i'm not sure if i should add or not and what kind of capacity i should be messing with.

RBob

iTrader: (1)

The ECM ADC inputs that have resistors to ground, they, from what I can tell, are for various reasons. On the TPS input it is to make sure that if the TPS or TPS wiring is open the input goes to ground and the ECM knows something is wrong.

Wouldn't be able to drive the car if the TPS was open and the ADC input was floating all over the place.

The 51K on the MAP input to ground is likely for impedance matching. That resistance is spec'd in the MAP data sheet as "the load impedance seen by the sensor shall be 51K ohm."

Filtering, mostly on the MAP inputs, is the series resistance and capacitor to ground. This creates a low pass filter to eliminate high frequency noise.

Most of the other inputs have a 10K in series and a capacitor. The cap is likely a .01 uf or a .1 uf for some filtering. The 10K also provides over voltage protection to teh ADC unit if 12 volts is placed on that input.

RBob.

The_Punisher454

Looking back at the 2010 posts here regarding the talk of using microcontrollers on your test bench. I totally agree that uC's are the way to go. Saar mentioned the hurdle of figuring out programming these things.
the aproach I take with microcontrollers is to just program in a high level language instead of assembly. Pic's and AVR's(my choice) both have some real good Basic compilers (some bad ones too). The huge amount of time saved programming this way more than offsets the minor cost increase of using a slightly faster version chip to compensate for the less slower running compiled code.
I'm currently using 32Mhz X-Mega AVR's which average 1 instruction per clock (vs typical 4 clocks per instruction on pics). These cost under 10 bucks each and have an accurate internal clock(no crystal needed). These are of course SMT, but I make my own boards. For easy to fab through hole stuff you can get 20Mhz AVR's for a couple bucks each. Even when programmed in basic a 20mhz AVR will easily handle your timing needs. I use Bascom basic which has over the years evolved into a pretty good compiler.

For Pics I have seen the Swordfish compiler for the pic18x series which I believe is free.

Saar

got around to messing with the pot wiring tonight, just did the MAP, now it has full range. amazing. swap two wires, go from non-functional to more or less perfect function. the pot seems to be a bit touchy near high MAP values compared to low ones, but this is acceptable to me.

and as much as i'd liked to have used a more sophisticated solution for the pulse inputs, i don't actually know any high-level languages either. i do ok with 6811 and similar assembly, but since i've never needed to, i haven't looked into other stuff. BASIC, i did program a few calculators back in middle school without any real prior instructions. eventually, i'll be messing with Arduinos for what would be best described as a homebrew BCM, so that learning curve will probably be fun. i basically need it due to lacking enough I/O ports on my target PCM to run things like allowing some time after key-off to roll up windows/etc, control variable power steering assist solenoid, re-integrate an intelligent CCP solenoid control scheme(need the output for TCC PWM solenoid) all kinds of fun stuff.

so, basically this thread is almost wrapped up, so a few final questions and thoughts: for anyone looking to build their own bench, two things that will help immensely are use a LARGE enclosure and use very generous lengths of wire. without these two things, it is painful to open/correct the internals. otherwise, research, research, research... and if you have a sensor to emulate that needs to generate an AC signal (magnetic VSS for example), simplest solution is to use a small cap(.01uF will do) in series with what would have been a square wave from a 555 or other signal gen.

and now the question : when it comes to logarithmic and reverse-logarithmic pots, which ones are which? for example, the CTS circuit has a gigantic "dead spot" from ~100K to like ~15K, so turning the pot 80% of it's travel results in a 10* difference, then in that last 20% you'll jump from -30*F to 305*F really, really quickly. which of the two would linearize this a bit?

EagleMark

The more I look are Arduino the more it looks like an entry level programming lesson along with an interesting way to do things including some I/O for a test bench. Or for things like motor drivers for IAC and sensor tests.

But since you need more I/O things assocoated with a BCM maybe a BCM would be better? I've never looked into one to see if chip could be programmed but have always been interested in one for all the options your looking for and a way to program them. Look at ALDL A4.ds down to BCM and you'll see it has everything your thinking of and then some.

As for your test bench I had the same issues with touchy pots on JimStim but they got me by and I lost intrest in building a more stable pot.

HTH!

Saar

i'd considered it before, but most BCMs, or at least the more full featured ones, seem to want to communicate with other modules, something i don't want to deal with. i'm sure that could be programmed out but if it's unnecessary, i'd like to avoid it.

Saar

....

fire up the bench this morning, now TPS reads perfectly from 0-5 volts, but MAP and A/C pressure don't allow going anywhere near their lower limits...

guess i'm taking it apart again....

Saar

and i confirmed was the problem was: the pots themselves are flaky. removed the MAP from the box, i can heat up the end terminals and throw a small amount of solder on them and the resistance from end to end will vary anywhere from 100K to 26K... when turning the dial it does odd things to the end to end resistance as well.

MAP worked last night, but i decided since i already resoldered two of it's three terminals, i might as well do the third. more proof that when something is working well, it's best just to leave it alone.

so i hooked up the pot again and watched the A/D results in the datastream. if i swing the pot around really quickly, i could get some wildly different values than if i turned it slowly. slow turn, minimum MAP came out to maybe 40kPa, if i took it from ~100kPa and turned it down fast, it would drop down to 10kPa, then turn it back up, stuck on 10K until half-travel, then it shoots up to 75kPa or so...

now i have to see if i have any more spare pots...