O2 and misfire?
 

Howdy all,
I've been dealing with an occasional misfires (or at least think so) of my Vortec. Things fixed so far: replaced distributor cap, plug wires. I've ordered a new distributor set and a coil (they're under way). That issue really keeps me awake.

Today I have replaced the wires (I would have some "friendly" words to the designer for a wire's harness underneath brake booster).

As soon as I had it done, I wanted to observe what is going on: I've switched AT to park and start the engine. Starts lovely, from a 1st kick, on idle I had a steady 500rpms (or about). Just felt a tiny shakes (from an occasional misfires?). O2 readings were normal. Engine running in a closed loop. I've set the throttle to get about 3000..3500 rpms. The engine speed up roughly. Noticed, that 2nd bank O2 sensor readingwent down to zero. A few seconds after O2 went to zero, I had a several loud blasts. I've quickly closed the throttle to avoid any further damage.
I haven't found literally any error codes.
I suspect, there may be a problem with either:
- miss-firing and engine knocking (already done some work), or
- a fuel injection unit or a fuel pump

The roughness goes away, when the engine is running for some time (like 15-20 mins). Or at least was going, as for the moment I don't go on a long runs, just doing the garage works in a free time.
At a 1st glance I thought of an ignition coil, that it gets heated and evaporates the moisture out. But then I've quit the idea as the ride stays in garage, that isn't wet. Besides I have the relation - the higher RPMs are, the more rough operation.
I thought of a distributor timing, but yet, on idle it's more or less OK, and the problem persist with getting a higher rpms. The timing is done by a mechanical coupling, cannot change in functions of RPMs...

I don't have any info on the spider assembly. I thought of doing a:
- hydraulic circuit that would flow a cleaner
- electronic control unit to drive the injector
I would like to hear Your opinions. Thing is, on our polish market we have some basic things for Blazers available. I am getting my supplies directly from US, but it takes time.... too much time for an eager person like me.

Cheers, Mike

\
Sounds like a wiring issue or possibly a bad sensor. If the O2 sensor is intermittently not sending a signal the ECM probably momentarily richens the mixture. Which could cause mis fires and when the unburned fuel in the exhaust does ignite it could make anything from a pop to blowing the muffler off.

Thanks for the reply.
I believe the O2 readings are accurate. On idle, the signal is steady, I have nice trace, w/o missing data.

I wonder if that could be related to a distributor. I had been removing the old cap for cleaning (my first thought) while I noticed, that one screw isn't torx20, but a Philips one. Then I found, that one of the cap mounting bracket have had been broken and glued to a cap. In a crappy way of course, as the bond was broken. I have ordered and applied a new one, which seats firmly, but things aren't OK yet.
I wonder, if the distributor rotor could spin off the center (on acceleration the engine swings) and miss one side of the cap pins. That would explain one bank misfire. Though, I am not very familiar with this distributor mechanical assembly.

Anxiously looking forward a new distributor set that should arrive around a new year's day.

Thanks for advice :)
Cheers, Mike

What do the misfire counters look like on live data?

George

Hm...
Good question George.
I went to a car with a OBD Dongle and Torque Lite. Got connected, started the engine to heat up and waited till it goes into a closed loop. Meanwhile I set the misfire plots for each cylinder. Since the fuel/O2 loop got closed, I've started increasing the rpms to see when the misfire counters start increasing. Strangely, but they don't. Just a flat lines. Neither the knock counter. Moreover, the engine did not run harsh as it was yesterday (wasn't smooth thou, I have had few quiet blasts and a little uneven operation). I could wind up to 6k rpms easily.

Your advice lead me to a conclusion it is not a misfire, or at least, it is not recognized by ECU as a misfire. I am not aware as a misfire detection mechanism works. I could swear it is, as it feels exactly like one.

Now it starts driving me crazy, as not only I have no idea what would that be, but I have no answer what conditions have changed it was +- ok for today, but not ok a day before. Need to observe that for a few days more.

Another thing is, that I found a record in a car documents, that this ride was running with an LPG installation for a while, but it was dismantled. That was a 1st generation of the LPG, that has an evaporator and delivers the mixture of oxygen and gas directly through the intake, not by the fuel supply system. I guess these problems I have on a petrol significantly increased with LPG installation that most probably blew up a plastic intake several times, so the owner decided to dismantle that... That would suggest it is not a fuel system. I am getting so confused now...

Cheers Mike

P.S. I suspect that some of the indicators available on Torque Lite app are not real, but a calculated value. Are they? If so, how could I distinguish the physical sensors from a calculated ones? Should I dig for a sensor map for my engine...?

EDIT: OH, yes, another thing You might ask me - my mileage; for the moment I have 112k. No overhauling I guess except some basic maintenance.

Misfires are counted when the crankshaft has a slight change in rotation speed near the firing of that cylinder so it is an implied fault and there is no indication of the cause. Either there are no misfires or the ECM is not detecting them properly.

What are your fuel trims at idle and again at 2500 rpm? Do they change abruptly when the truck starts to run like crap?

An O2 sensor voltage going to zero and staying there indicates an extremely lean mixture or a failure.

Have you done the fuel pressure/leak down tests? If so, have you left the gauge hooked up to see what is happening when the truck runs bad?

George

I don't believe that Torque Lite will read misfire counters. You have to use Torque Pro and also enable GM Enhanced PIDs in the the software. Before you are done with installing a new distributor you will want to check and adjust cam sensor retard. Torque is not going to do this for you. If you have 1998+ Blazer you can use Car Gauge Pro or Dash Command. If 1996/1997 then only Dash Command will work. Dash Command requires you to buy enhanced PIDs at $10 extra per vehicle. Car Gauge Pro does not require any additional purchase.

@GeorgeLG
Yet another good question. My respects! I am going to check the trims and if they're not zero I'll look for a possible leaks. I am truly grateful for pointing me into a right direction.
EDIT: I took a ride for a short trip with my mobile in a cradle and a Torque connected. Luckily, today it was running rough.
Here my observations:
On idle, the FT trims (dunno if they were LT or ST, that is unspecified in Torque, just a "FT/1" and "FT/2" plots) for both banks are zero.
On acceleration, up to 2k rpms, both trims are going positive up to 20%. I didn't want the mixture to go too lean and to avoid blasts I've did not go beyond 2k rpms, but I suspect it would go even further with triggering a MIL indicator.
On deceleration the trims are going negative with up to -20%.
Now I wonder if it is possible that I have a vacuum leak, as on idle the air leak would play a significant role and the trim stays around zero. I assume, the lower intake gaskets were never replaced, and as Christine stated lately, they are prone to break. I guess I should check the vacuum sensors now? MAF/MAT, etc or You'd have another suggestions?
EDIT 2: I have disconnected MAF sensor so the ECU would run on its map. The engine runs smoothly, as it never was before. I already gave a call to my supplier here, but he doesn't have MAFs on stock. Kindly asking what should be the pinout resistance, so that I could check if the resistors are OK. If they are, the cleaning should do the job. I also got an oscilloscope probe from a friend, so that I could check the traces of a high voltage that goes to the plugs and see if there are any problems with a distributor or an ignition system in general.
Thanks George, I owe You for Your help.

By the way, I have found an excellent video that explains the trims:
@Lesmyer
Well, my dongle is a WiFi type. I went through a Car Gauge Pro settings, but seems to me that only a blue tooth dongles are supported. So, is that correct I should get another dongle to run a Car Gauge?
To be honest, I have tried to search the difference in between Torque Lite and Torque Pro, but haven't found much of an info, neither on the developer's website. As I understand, the Lite version is somewhat limited to a Pro with the amount of a different indicators + the ads.

I am going to change the distributor anyway due to its condition. The prev. owner did a really "quality job" around with its cap mounting brackets.

Yes, Car Gauge Pro is Bluetooth only. Dash Command will work with WiFi but cost will be $10 + $10 more for enhanced PIDs for your vehicle. if you decide to buy bluetooth dongle, I recommend BAFX OBD2 as sold on Amazon for about $23. It is very fast refresh rate in comparison to most (~30 readings/second in Blazers). I once had a cheap Wifi adapter that did only 3 readings/second. This can be important when monitoring things that change rapidly like multiple misfire counters, O2 sensors, fuel trim readings, etc. Torque Pro does have a counter that displays scans/sec.

If you end up not using Car Gauge Pro (and you have purchased it), you can email and they will probably refund. They have done this for other people if Car Gauge Pro did not work for their particular application. Biggest argument for Car Gauge Pro is (as I understand it) is it will read and clear codes (not display PIDs or control) in ABS, SIR, BCM, TCM, other vehicle systems and you do not have to buy further licenses for each vehicle to get additional functionality. I do know that I can read and clear BCM codes on my 2004 Avalanche with Car Gauge Pro. Torque Pro will do Enhanced GM Powertrain PIDs as well as Car Gauge Pro, but Cam Sensor Retard is not supported - nor is reading/clearing of other system codes. Downside to Car Gauge Pro, is that it does not support Enhanced GM PIDs after CAN communication protocol was implemented (2006+) nor does it support Enhanced GM PIDs for 1996/1997 either - however it does support CAN protocol enhanced PIDs for the more modern Ford and Nissan. On 2006+ and 1996/1997 GM, Car Gauge Pro will only do the same things as Torque Lite. It's kind of hit and miss - but it works quite well with most of the Blazers.

All Torque Lite gives you is the ability to read/clear powertrain codes and monitor generic SAE PIDs which does include Fuel Trims and MAF and O2 sensors and coolant temp, etc but not misfire counters.

Even with a new distributor you still should check that cam sensor retard is 0° ± 2°. This value is an indication of proper alignment of the rotor with the cap as spark passes through. Ignition timing is actually determined by crankshaft sensor and PCM on this system, but the rotor still needs to point exactly to the right terminal or misfire can occur. This is a common cause of misfire in these Blazers. As long as factory hold down clamp has not been modified, the only possibilities for this being off include stretched/worn timing chain/gears, worn distributor gear, or excessive end-play for distributor shaft. If you get a P1345 after changing the distributor, it is in a tooth off.

It's your choice on what software to buy and use. I have Torque Pro, Car Gauge Pro, and Dash Command on my cell phone. I have HP Tuners VCM Scanner and Scan XL Pro on my Windows notebook. I have found a use for each and every one, at one time or another. I find the Scan XL Pro to be the least value for the money, but it does do Enhanced GM PIDs using CAN protocol on my 2006 Pontiac - so even it has it's purpose. Only cheap software that could do the same was Dash Command.

Finally - Positive fuel trims mean that system is adding fuel over the factory calibration in an attempt to correct fuel mixture. A vacuum leak would tend to affect fuel trim at idle far more than at RPM. +20% would mean that pulse width of injectors has been increased 20% over factory calibration in order to make mixture give correct O2 sensor readings. FT of zero (no alteration of factory calibration) would tend to mean no vacuum leak (unless system is going to open loop and not controlling the fuel mixture at all). O2 sensor readings would tell.

Better I think to get your distributor in and some good software besides Torque Lite before coming to any conclusions.

I still don't remember you telling us what year Blazer you have.......

Good luck and best wishes

Thanks Lesmyer for a detailed info on the differences in between Torque and Car Gauge. It is more clear and detailed than the info I found on a developer's website :)
I find it a bit downing the Car Gauge does not support WiFi standard. I've chosen a Wificause I expected it to be more handy - it might be also supported by some laptop software, as mine does not have a BT. And could possibly offer a higher data transfer.

Nevertheless, I don't like half measures and these dongles cost next to nothing. I have already ordered another one with a BT and I am going to go for a Car Gauge Pro.

As for checking the timing settings, I am going to check it with Car Gauge as You have mentioned, and record the value just in case. Next, I am going to replace the plugs for an AC Delco iridium set I have recently received. Removing a plug that goes behind a steering column is a pain in ass, so I think I'll combine that with a distributor swap (and a compression test). I'll bring the 1st piston to DTC and see what I find on a markers under the cap. For the moment my distributor does not seem to be a problem, as with a MAF disconnected, when the engine runs in an open loop, it is all perfect.

I have borrowed an oscilloscope probe that taps to a plug wire and allows You to observe the ignition pulses. It is a very similar to this one:
Automobile spark plug oscilloscope probe
I expect that this probe is going to be more handy than info delivered by ECM, as You see the analog readings of a pulses, so their shape and width. Digital is digital - it counts the pulses but does not refer to their shape. I may drop some photos here if anybody likes.

For the moment I'd need an info of the MAF resistor values, so that I could check if mines is OK. If it's busted I'll get a new one, but I want to make sure 1st. Maybe they are covered with a thin film of some crap that prevents a good heat dissipation and cleaning with an alcohol would be enough.

Cheers, Mike

Oh, and regarding to my ride, all the details are in my signature: Chevy Blazer '00 4X4 4D LT. I am not gonna need the extra features of OBD from 2006+ :)

Here are some readings from my 2001 Blazer at idle (in drive IIRC) that I had previously taken using Dash Command (different software displays different units). Maybe it will be of some use to you. I consider these to be quite normal. You can convert the units from kPa and g/s if necessary:

RPM = 578 rpm
MAP = 46 kPa
Baro = 98 kPa
ECT = 91°C
IAT = 63°C
IAC Position = 109 counts
EGR Position = 0%
CMP Retard = -0.8°
MAF = 6.46 g/s

Now I see that your vehicle is in your signature!

You can definitely see misfires with an oscilloscope, but you need to know what to look for and where to look in the traces. The leading part of each trace when displaying secondary traces in raster format is where I always used to look (when I had access to an oscilloscope).

In Car Gauge Pro, you will find two different PIDs named CMP Retard. Use the one with the single * for your Blazer. Somewhere on this site I have written some basic instructions for doing some things using Car Gauge Pro. You can search for that if you like. Tells you how to set up multiple gauges on same screen among other things.

Here's readings at 55mph steady, taken using Torque Pro

RPM = 1701 rpm
Speed = 56 mph
MAP = 10.3 psi
MAF = 35.0 g/s
Intake Temp = 100.4°F
Coolant Temp = 194.0°F
Throttle Position = 22.7%
Timing Advance = 27.5°

:) Let's say I'm an old school guy in these matters and I prefer oscilloscope or any other device that measures a real, physical value rather than the dongle gadgets, where You are not sure exactly what is measured and what is just calculated.

I have two channel DSO, but only one probe, so not much of a choice. I may:
- watch one trace at a time; quick math here: 500rpms/60seconds = 8 1/3 turn per second; 8.33/6 cylinders= 1,39 firing pulses per second or a pulse every 720msec
- quickly make 2nd probe from a transformer and a diode set of two Zener's (in antiparallel) to limit the voltage on the output. Then I could set the coil pulses as a base for the trigger and use 2nd channel to check out the pulses on the selected cylinder. I guess I'd be too lazy for that.

Kindly asking once more, has anyone have any specification to the MAF circuit please? Or at least remember the values?

And sometimes the only thing that is important is what the vehicle computer thinks. You don't get answer because factory manual specifies a scanner reading under certain conditions, not a simple resistance. I don't have my manual here at work to give you the spec, but I have given you scanner readings of MAF sensor at two different sets of conditions. You should be able to use Torque lite to read the MAF and tell if it is OK.

Don't get me wrong, I did not say the dongles aren't useful. They are. By a dongle I could do trim lookup and do accurate troubleshooting today. Personally (and that is just only my personal preference) I prefer to stick to the physical values. Lots of indicators, where many amongst them are just a result of math equation (that are not explicitly given) within ECU are somewhat confusing to me. As an example, if my ride is in an ambient temperature (so does intake) of 40F rather than 100F as in Your data, what changes (and where) should I expect? What if that is just only a wiring problem..? ECU cannot neither divide the problem between wirings and sensors.

I have reverse engineered the flow sensors used in an oil mist detectors on marine equipment. Don't remember the brand, but the idea was yet simple. And a same idea goes for a MAF sensors. There's no rocket science :) There were two PT100 temperature sensors located in a same air stream. One of them just measures the temperature of a flown gas, which is used as a reference (IAT). The other one does the same, except it is driven by a much higher current. The high current results in significant heating up of a sensor. Not to involve into thermodynamic rules, the air flow cools down the heated element proportionally to the air flow. By reading and comparing these two values into an equation formula You get the calculated flow amount.

Easy. Except, that if You lookup the PT100 conversion table, a 5 Ohms difference (that could easily be a result of a corroded contacts in connectors), for PT100 sensor would result in a 10K (or Celsius) temperature difference. That is pretty much, and if it is on a line to the heated element, that plays more significant factor in the equation You may have everything messed up.

Sorry if I went too much into theoretical digression. I am just trying to keep a common sense - the shipping from US takes time, and a parcel handling is a bit pricey. And I love troubleshooting. Thought somebody might have done it before me and could drop some feedback. Thank You for the data, I promise I'll do the MAF sensor tests tomorrow.

The MAF sensor outputs a changing frequency to the ECM. You check power and ground then watch the frequency of the output at different flow rates. Since you like direct measurements, break out your DVM and sharp probes. This may help:

Part 1 -How to Test the GM MAF Sensor (3.1L, 3.4L, 4.3L, 5.0L and 5.7L)

On fuel trims: you always add the short and long term together for a given bank. That tells you how much the mixture is different than the look up table for those conditions. A positive value is adding fuel. You usually look at idle and 2500 rpm. 10% absolute value is generally considered OK. I think 25% sets a code but many things can be wrong on your way to 25%. It is highly unlikely that you actually have zero fuel trim, especially if an engine is running that crappy and straightens completely out with the MAF unplugged.

George (Grzegorz)

Yesterday I drag my multimeter to a garage to check out the resistance at the MAF terminals. To be honest, I did not expect hundreds of KOhms or MOhms, but then I have realized, that probably the designer was smart enough to design a signal converter, as for a simple PT100 connectors resistance would play a role.
Today morning I have a really prompting reply from You George, some more fun on the way :D
I have read the article up to the pinout description and went for a kynar wire (30AWG in a teflon coating), removed the insulation, stick stripped end into a female connector of MAF sensor and put the plug in. I did not like the idea with punching probes. Here You may see how it looked like:
https://cimg5.ibsrv.net/gimg/blazerf...623e83bd8a.jpg

Then i start the ride and check the traces I got on my DSO:
Here's the one for idle:
https://cimg0.ibsrv.net/gimg/blazerf...212789d66d.png
And here for something about 2k rpms. Couldn't know exact value as I've had m oscilloscope under the hood and been operating throttle manually.
https://cimg1.ibsrv.net/gimg/blazerf...0d61f71226.png

The units did not record (probably didn't set something while USB data storing, but it were 2V/div vertically and ... horizontally 100uSec :) )

If You wanna look how it looks live, here's a short video I made:

Now for the fuel trims. I haven't bought Torque Pro yet (waiting for BT dongle and Car Gauge) so my Torque Lite version gives me a chance to look at the STFT bank1 only and LTFT both banks. On idle, with around 500rpms I have readings in range of 0 up to -2%. No big deal. Cannot do idle test in a CL on 2k5 rpms, as mixture gets too lean. But I guess, my observations on LTFT that goes over 20% on acceleration and -20% on deceleration are sufficient.

EDIT: I think I'll take the MAF out and give it a bath in an isopropylene alcohol. I'll keep You updated.

Cheers,
Mike

So, to keep You updated.
Although the article gave an info on how to hook up to a MAF, there was no information how the proper traces should look like (no reference). Therefore, I had nothing to compare my traces with. At least, I've seen MAF seemed to be transmitting data.
So, I have decided to give it an alcohol bath:
https://cimg1.ibsrv.net/gimg/blazerf...b00c216afd.jpg
I've been shaking the MAF in a bath to get at least some cavity. I didn't see a reason not to wash also the electronic circuit unit, as IPA fluid is quite often used to clean PCBs after manufacturing process.
I haven't find a lot of crap at the bottom on the end:
https://cimg9.ibsrv.net/gimg/blazerf...14c2f754c6.jpg
I left the MAF for a bit to dry and mounted back on the intake.
Luckily, I could increase on idle up to 6-7k without any significant problems. I hooked up with Torque Pro to see the fuel trims on 2k5. This time the conditions let me to do so. On 2k5 my STFT/LTFT are about -8...-10%. Now the question - what's an acceptable range? I know, in ideal they all should be zero, but, hey, there is no ideal conditions, so they used to deviate with some values. What is Your experience?

On fuel trims, take another look at post #16.

George

So, if I have my trim up to -8% on 2k5, then I should assume my engine is running OK? Great :) So the problem is solved! Thanks to You all for a great support!
Cheers, Mike