P0455 and some other questions about fuel tank

[dhanger]

Hi folks-
Got the common p0455 and have done a lot of research so I have a pretty good handle on troubleshooting this but something odd about the fuel tank that I wanted to pester you about. Of course I replaced the fuel cap, tested the purge and vent solenoids and both worked on bench testing but I don't have a scan tool to see if they can be activated (wiring check) but for the moment I am assuming they are good and looking at other possibilities.

Amongst all the other bits of info I picked up while researching, it seemed that my car (04 TB LS SWB) did not have a replaceable rollover valve, it was integral with the tank and if it failed the whole tank had to be replaced, do I have this right? Well as I was poking around underneath I noticed that my fuel tank had what appeared to be a large nut right at the filler nipple on the tank, which surprised me as I didn't think I was supposed to be seeing that as it appeared to be replaceable. I then recalled many years ago (bought the car in its first year used with an extended warranty) that I had to bring the car in to the dealer because of fuel smell, so I looked through my records and sure enough they had replaced the tank, this was Feb. 2007, part #15229311. I looked up the part # from the records here and that number was listed as "2nd design" whereas the 1st design was part #15231628. So getting curious I looked up the part numbers and found this bulletin, https://static.nhtsa.gov/odi/tsbs/2015/SB-10089488-2280.pdf dated December 2015, long after my replacement. My dealer invoice noted that the filler neck on the tank was cracked which is why they replaced it, but the replacement had nothing to do with the bulletin so if I hadn't had the extended warranty I would have had to pay the whole price but I only paid a $100 deductible. Can someone confirm that the tank I have should have the replaceable rollover valve?

I bring that up in relation to the P0455 because I wonder if the rollover valve may be causing the code and maybe that's what I need to replace? I bought a cheap smoke machine and was trying it out today but the dang thing kept overheating and a thermal switch kept shutting it down, so I never really got a good smoke test on the system (I ordered a better one so I will be trying the smoke test again in a few days). While I was running the test at one point I removed the gas cap and I could hear the smoke machine pressure coming from the fuel tank, which surprised me as I was assuming that the rollover valve should be preventing air/smoke from escaping from inside the tank. Is this correct? Another symptom I've had for some time is a small amount of vomit at the gas pump, right at the end of filling. Not the problem of constantly shutting off the pump but just the little vomit and no problem filling otherwise. This is making me wonder if indeed the rollover valve is bad. Whether that solves my evap leak or not remains to be seen but should I look at replacing the rollover valve anyway?

 

[TJBaker57]

don't have a scan tool

What DO you have? Handheld vacuum pump maybe? OBDII adapter dongle for use with phone apps etc.?

 

[dhanger]

What DO you have? Handheld vacuum pump maybe? OBDII adapter dongle for use with phone apps etc.?

I have both Torque Pro and Car Scanner. I do have a vacuum pump (compressed air venturi type)--what would that be useful for in this context?

 

[dhanger]

Oh yeah, just remembered I have a vacuum pump I got for A/C service.

 

[Mooseman]

I was typing this response while the above two were posted.

That is a new-to-me bulletin because the original one only covered the EXT/XL tanks. I know because I had mine replaced on my '02 EXT in 2011 paid by the extended coverage. When they issued this later 2015 bulletin, it was almost pointless because it would have covered, at the latest, 2004 trucks up to 2019 (15 years). And this also explains the rash of these SWB vehicles lately with leaky tanks.

Amongst all the other bits of info I picked up while researching, it seemed that my car (04 TB LS SWB) did not have a replaceable rollover valve, it was integral with the tank and if it failed the whole tank had to be replaced, do I have this right?

That is correct.

Can someone confirm that the tank I have should have the replaceable rollover valve?

It should not. All OEM tanks have non-replaceable ROVs. What may look like a nut is probably a fake molded into the tank.

Another symptom I've had for some time is a small amount of vomit at the gas pump, right at the end of filling. Not the problem of constantly shutting off the pump but just the little vomit and no problem filling otherwise. This is making me wonder if indeed the rollover valve is bad.

That could be indictive of a bad ROV. It's secondary job it to also close the inlet when the tank is full, preventing that "vomit" or splashback. When it fails, the tank gets overfilled, which can damage the charcoal canister, and will cause gas to puke out of the filler. I've been there. However, this would not be the cause of the P0455 unless it is leaking. Check again and see if you see streaks of fuel down the tank coming from around the filler.

If you do have a leaking ROV/neck, there is a kit available from Dorman to fix this without having to replace the whole tank, which may be NLA. It was originally marketed for the EXT/XL but should work on yours.

Dorman Products - Aftermarket Auto Parts & OE Solutions - 577-106

Fuel Filler Neck Inlet Check Valve

www.dormanproducts.com

 

[dhanger]

Can someone confirm that the tank I have should have the replaceable rollover valve?

It should not. All OEM tanks have non-replaceable ROVs. What may look like a nut is probably a fake molded into the tank.

It did look molded in as there was no clear delineation between the "nut" and the tank body.

If you do have a leaking ROV/neck, there is a kit available from Dorman to fix this without having to replace the whole tank, which may be NLA. It was originally marketed for the EXT/XL but should work on yours.

I did not see any leakage evidence around the filler neck. I've seen that Dorman product in my research as well so I was aware of it. If the filler neck is not leaking/cracked then my only concerns there are splashback and possible canister fouling, I take it. I don't think the canister is fouled yet, according to the description in the service manual (https://tinyurl.com/yc45ehmw) I should be getting P0446 (canister restriction) and I read that one of the symptoms of a fouled canister is the constant gas pump shutting off, but that has not been a problem. But I wonder if I should replace it anyway based on its age (still factory original)? If there is an accident/rollover, as long as the filler tube is not affected then the gas cap should prevent any spills, so the chances are pretty slim of this being an issue. Just trying to justify whether it's worth pulling the tank and cutting out the old ROV to replace it.

So right now the next step is waiting for the new smoke machine to try to track down the leak.

 

[TJBaker57]

I have both Torque Pro and Car Scanner. I do have a vacuum pump (compressed air venturi type)--what would that be useful for in this context?

Hand vacuum pump is used for testing the sealing abilities of the purge and vent solenoid valves.

Using the either the Torque Pro or Car Scanner apps if you don't already have it I can show how to enable viewing/recording fuel tank (evap) pressure and maybe vent and purge valve operational status.

I may even be able to show you how to temporarily activate the vent and/or purge valve solenoids. Maybe.

 

[dhanger]

Hand vacuum pump is used for testing the sealing abilities of the purge and vent solenoid valves.

Using the either the Torque Pro or Car Scanner apps if you don't already have it I can show how to enable viewing/recording fuel tank (evap) pressure and maybe vent and purge valve operational status.

I may even be able to show you how to temporarily activate the vent and/or purge valve solenoids. Maybe.

If there's a possibility of activating the purge and vent solenoids using one of the apps without the need for a pro scan tool (at least it sounds like that is what you are proposing) that would be great as it would give me confidence that the system as a whole is either working or not. Bench testing the solenoids only tells me that they are working passively, not whether the PCM and associated wiring/relays are actually doing their job as well. It's logical to me to assume that if the vent valve for example was not actually activating (closing) during the large leak test then of course there would be a large leak through the vent but a smoke test would not reveal that even if the valve was manually closed or the vent otherwise plugged off. Whether or not that's even a viable scenario I don't know as I've never worked on an evap system before and I don't recall seeing that specific case anywhere in my research on the subject.

But I'm kinda getting ahead of myself anyway, until I get a good working smoker I'm just poking at it. That will be later next week, but if you want to feed me some information about using the apps I could play with that in the meantime. Sounds like a fun and educational project and hopefully useful to others following along.

 

[AmpOverload]

If there's a possibility of activating the purge and vent solenoids using one of the apps without the need for a pro scan tool (at least it sounds like that is what you are proposing) that would be great as it would give me confidence that the system as a whole is either working or not.

I don't want to intrude too much on this thread since you're already being helped by @TJBaker57, who as all the forum regulars know, is supremely capable and helpful in these matters.

I only want to suggest that you read this post of mine where I discuss some related issues, including the possibility of monitoring the EVAP performance during the automatic self-test. (This is, in part, a more verbose discussion of part of what @TJBaker57 has already suggested, I believe.)

@TJBaker57: Don't hesitate to let me know if I can be of any assistance in finding or confirming any EVAP-related PIDs or valve-control commands on the OP's vehicle, using my vehicle simulator.

 

[TJBaker57]

Bench testing the solenoids only tells me that they are working passively

And if you are only checking for the mechanical movement this does not tell you they are sealing 100% as required. This is where the hand powered vacuum tool comes in.

The first time I was fiddling with my own I removed the vent valve/solenoid and connected my little hand pump to the vent line. I pumped for a good while and then heard a 'bump' sound. I had applied a vacuum strong enough to slightly flex the tank inward! Waay too much vacuum but it proved my purge valve was sealing and most likely I had no leaks as if there were significant leaks I never would have been able to achieve a decent vacuum with a little hand pump.

I did find my vent valve filter and housing were full of dirt from driving on desert dirt roads. This may have been enough to prevent my vent valve from sealing well.





I will say this much... I know the messages to transmit on the vehicle data bus to temporarily activate my own 2002 4.2 EVAP vent and separately the purge solenoids. I suspect your vehicle would use the same message(s). Sending said message will activate the solenoid for about 5 seconds. This is enough to confirm the circuit but not enough for any leak testing.

 

[dhanger]

I will say this much... I know the messages to transmit on the vehicle data bus to temporarily activate my own 2002 4.2 EVAP vent and separately the purge solenoids. I suspect your vehicle would use the same message(s). Sending said message will activate the solenoid for about 5 seconds. This is enough to confirm the circuit but not enough for any leak testing.

As you say even 5 seconds activation is adequate for confirming all circuitry is working as expected. When bench testing I confirmed by blowing and sucking that both valves were clear when open and would not pass air when closed.

I only want to suggest that you read this post of mine where I discuss some related issues, including the possibility of monitoring the EVAP performance during the automatic self-test. (This is, in part, a more verbose discussion of part of what @TJBaker57 has already suggested, I believe.)

Yeah that covers a lot of what I'm trying to do, interesting reading. I too was under the impression these little dongle apps were 'read only' and not bi-directional

At the moment I'm not really expecting that either valve (purge and vent) is the cause of my large leak but confirming their operation would be nice in order to eliminate that possibility. Smoke machine is coming tomorrow, sooner than I expected at first, so maybe Monday I'll be able to find out more about the leak.

 

[AmpOverload]

Sending said message will activate the solenoid for about 5 seconds. This is enough to confirm the circuit but not enough for any leak testing.

With the caveat that my testing was done on a P04 PCM (2004 Buick Century, specifically), the valves will retain their commanded positions as long as periodic Mode $3F ("Test Device Present") commands are sent. My Android app sends 1 every 4 seconds. (I don't know if they could be sent less frequently, but I know that it works at that rate.) I suspect that a 2004 Trailblazer (with a P10 PCM, IIRC) would work similarly.

 

[TJBaker57]

valves will retain their commanded positions as long as periodic Mode $3F ("Test Device Present") commands are sent.

True that. I was aiming to keep things simple, at least as a start. Especially when dealing with 'Device Control' messaging.

 

[dhanger]

Progress report--hooked up the new smoke machine this morning, better quality smoker that isn't shutting off on me repeatedly. I disconnected the hose between the vent valve and canister at the vent valve and plugged the smoker in that end of the hose going to the canister. Bottom line is I had no smoke escaping anywhere after at least 20 minutes of constant smoking. I looked everywhere--especially around the filler neck at the gas cap as I know that's a weak area, but nothing there. Took the gas cap off and smoke poured out so I know I was getting smoke at least to the tank. I went to the front and removed the purge hose at the throttle body but nothing there so I think it's safe to say that the purge valve is not leaking. Unless nothing is getting to the purge valve yet--I removed the other 2 hoses at the canister and had smoke out of them so I didn't think the canister was plugged but I removed the canister anyway to check it. It was pretty heavy so I weighed it to compare with a new one, a Dorman at Amazon which was listed at 78 oz. while the old one weighed 95 oz., also when I shook the canister vigorously I heard nothing so I'm thinking the canister is fouled.

I'm going to replace the canister before I proceed any further. Can't say for sure if the canister is plugged to the point that smoke couldn't get further than the gas tank so I don't really know if I'm even getting smoke to where the large leak is yet. Unless--the signal to close the vent valve is not working as discussed earlier and that's where the large leak is. I removed the valve once more and bench tested it, clear passage when not activated and no air passage when activated. If you guys can help me with activating it through the scanner app that would settle that question. Meantime I will replace the canister when it arrives Wednesday and re-smoke it then.

 

[Mooseman]

It`s also possible that the tank vacuum sensor is plugged or bad. If it doesn`t detect a vacuum when it does its test when started, it would throw a code. It closes the system and applies a vacuum to test. It has to hold the vacuum for a predetermined amount of time to pass. A large leak detected would be that it doesn`t hold vacuum at all. Since there are no leaks according to the smoke test, either the vent valve isn`t closing or the purge valve isn`t opening to apply a vacuum.

Did you check the other end of the hose where the vent valve vents? Since it`s venting through the canister, the smoke might not make it through the canister and out the vent hose. You should test that the vent valve actually seals when power is applied. You could take it off and put the smoke machine on it, put 12V to it and see if smoke gets through. Or you could just suck and blow through it.

 

[dhanger]

It`s also possible that the tank vacuum sensor is plugged or bad. If it doesn`t detect a vacuum when it does its test when started, it would throw a code. It closes the system and applies a vacuum to test. It has to hold the vacuum for a predetermined amount of time to pass. A large leak detected would be that it doesn`t hold vacuum at all. Since there are no leaks according to the smoke test, either the vent valve isn`t closing or the purge valve isn`t opening to apply a vacuum.

Good thought--although without a proper scan tool to simulate a large leak test I don't know of a way to monitor the test when the PCM commands it, unless I happened to catch it at the right time which is unlikely I'm thinking. If I reset the code and then waited for it to reset would a freeze frame show the data for that? Or maybe I already have a freeze frame since I haven't reset the code yet. I'll have to check that.

Did you check the other end of the hose where the vent valve vents? Since it`s venting through the canister, the smoke might not make it through the canister and out the vent hose. You should test that the vent valve actually seals when power is applied. You could take it off and put the smoke machine on it, put 12V to it and see if smoke gets through. Or you could just suck and blow through it.

I covered all that in the previous post, either I didn't explain well or it was TLDR for you, LOL. If you need it clarified let me know and I'll re-word it.

 

[Mooseman]

I did read it, just misunderstood it. I still think there's something there despite no smoke escaping the vent. Like you said, maybe the canister is plugged.

Doesn't Car Scanner show the tank pressure sensor? Torque Pro might. Even though you can't close the vent via a scanner, you could close it manually via a direct 12V to it and watch the pressure rise (or not) in the tank. This would work where you are (not in the cold).

 

[dhanger]

Doesn't Car Scanner show the tank pressure sensor? Torque Pro might. Even though you can't close the vent via a scanner, you could close it manually via a direct 12V to it and watch the pressure rise (or not) in the tank. This would work where you are (not in the cold).

They probably do but honestly I'm pretty clueless how to use these apps. I've searched a couple of times for some good tutorials or documentation on how to set them up but just come up empty handed for the most part. I can see basic stuff like RPM, MPG, even set up O2 sensor readings, but not much beyond that. They seem to assume a basic level of understanding that I just don't possess. If anyone can point me to some articles or such on how to get in depth with them it would be much appreciated.

 

[TJBaker57]

anyone can point me to some articles or such on how to get in depth with them it would be much appreciated.

I'll see what I can do in the morning over coffee...

 

[TJBaker57]

They seem to assume a basic level of understanding that I just don't possess.

I see these apps and such as being in the realm of the hobbyist and not so much the mainstream. It took me years to get any traction.

It seems like none of the apps are fully equipped right out of the box. They can supply the generic legislated data points but not the specialized configurations for particular manufacturers.They are getting better though.

That's where the community comes in. Here at this site we deal with General Motors vehicles for the most part. As such we have gathered a good deal of informations on using apps like Car Scanner ELM and Torque Pro as well as a few others.

There are some long running threads here dealing with PIDs (Parameter IDs). We chatter about this data point or that and what we did to try and get it figured out. Some find it interesting while others may see it as just another road block that leaves their head spinning.

Let's just start by getting you access to the Fuel Tank Pressure sensor. I will assume you have already added your vehicle in the app. When you added your vehicle you had to select a "connection profile". This can be either "OBD-II/EOBD" or "OBD-II + SAE J1850". Either will work. This can be altered if desired in the "Settings" menu item.

Also in the "settings" menu item you will see "Sensors". Select that and you will see at the top "User Created (Custom) Sensors". Select that and the next screen will almost certainly be blank as it is not likely you have added anything in there yet. This is the area where we define data points discovered by the community.

Near the top right of the screen there is a plus sign, tap that and an entry labelled "New Sensor" will display. Edit this entry to look like the following screenshot. There is no "Save" button the edits are live as you make them. The "Test" button will not work unless the app is currently connected to the vehicle but it is not required to test the sensor anyway. Only when things are not working as they should do I use the test button. Exit the editor with the back arrow at top left of screen.

This should let you view the voltage returned by the fuel tank pressure sensor. It's a starting point.

 

[dhanger]

Okay, off to a great start. Added to my dashboard just for this purpose, not planning on keeping it there but just for now. Here's what I got after connecting:



Based on this document (bottom of page):

Evaporative Emission (EVAP) Control System — 2004 Chevrolet TrailBlazer 2WD L6-4.2L VIN S Service Manual | Operation CHARM

Detailed repair manual for the 2004 Chevrolet TrailBlazer 2WD L6-4.2L VIN S.

tinyurl.com

lower voltages indicate higher pressure, so 1.5 is in the ballpark of what I would expect to see. I'll check it again later this afternoon as it warms up (expecting 82F today, sorry all you northerners LOL). I would expect the voltage to drop at least a little by then if the sensor is working properly. Edit: it may not change at all unless I plug the vent which I can't do today, so I'm probably wrong about this.

Don't have much time to play with this today, I'm expecting the new canister tomorrow so I may have a chance to work on the car Thursday at which time I will plug off the vent and watch the voltage while putting a vacuum on the system. You mentioned before using a manual vacuum pump and hearing the tank sucking in, do you recall how much vacuum you applied at that point? Don't want to overdo it.

In the sensor editor you have a tilde (~) prepended, does this have any particular meaning or is that just for sorting purposes--I noticed that it went to the very end of the sensor list which makes it easier to find.

Thanks a bunch--I am doing some more homework on this stuff so I'm not just relying on being spoon fed, I'm really interested in getting better with understanding.

 

[Mooseman]

The pressure or vacuum won't change unless you close the vent valve. Since you don't have bidirectional control of the PCM, you'll have to do it at the valve with 12V applied to it directly. If still no pressure change, plug the vent hose. If there is now a change, then the vent valve is bad.

 

[TJBaker57]

doing some more homework on this stuff so I'm not just relying on being spoon fed

It is a very big can of worms. I took this slow approach to just get your feet wet.

Correct, I use the tilde for the purpose of keeping all my personal custom sensors together at the end of the listing. The name can be anything the user wants.

This evening I fiddled around with placing items on a Car Scanner Dashboard that can activate my 2002 4.2 vent valve solenoid and also for the purge valve. For the purge valve I chose 100% duty cycle.

I have the ability to make these instructions persist more than the 5 seconds I previously mentioned but I hesitate to do so for safety sake.

I once used my Tech2 for these same purposes and the tool terminated the test when the vacuum on the tank exceeded 10 inches of water column. Thats like 0.36 psi. The tool reported this was beyond the limits imposed by the software.

I recorded tank pressure today during a drive to a neighboring town.

Here is a short snippet showing the EVAP purge valve duty cycle and the resulting fuel tank pressure sensor readings.

 

[dhanger]

A little bit of progress today but still no answers for the large leak. Now having the FTP sensor configured has given me at least a little more insight--using your methods above would be even more helpful. I spent a few hours yesterday scanning your sticky More PIDS for Torque App, lots of detail there so I tried to absorb what I could but not get bogged down in the details. Best thing I got from there was the spreadsheet that @azswiss compiled (thanks, neighbor, waving at you from across town in Buckeye), tried adding evap vent valve status from that but haven't got it right yet. More on that later...

So today the new canister came but before installing I wanted to play a bit with the FTP sensor, if only to check if it's giving me valid numbers. I attached vacuum to the service port under the hood and then observed the pressure sensor while applying vacuum. I performed this test twice with the following conditions:

1st test: Detached the canister hose at the vent valve and plugged the hose end. This took the vent valve completely out of the equation.

2nd test: re-attached hose to vent valve, then energized the valve with 12V to close it.

Here's the graph from both tests:



I vacuumed down to about 4.5V (didn't have a vacuum gauge so don't know what that equates to and didn't want to exceed 5V) and then observed the decay for 5 minutes on both tests. Both ended about the same voltage so I think that's a good indicator that the vent valve is sealing good mechanically. I haven't seen any specs on what decay rate the PCM is basing the large leak on so nothing conclusive in that regard.

So at this moment I have the following conclusions: FTP sensor is working as expected and the vent valve is working well when mechanically activated but still don't know if it's actually being activated by the system and sealing the system for the large leak test. Have not done any testing on the purge valve yet other than bench testing with 12V power supply to confirm solenoid actuation and valve opening and closing.

At the end of the testing I installed the new canister and then performed a smoke test once again with the vent valve closed off, still no indication where the leak is. I observed the tank sensor while smoking it and saw it go down to 0.2V and stay there the whole test. This means that the system is getting pressurized nearly to max (lower voltage=higher pressure) from the smoke machine with a 1 PSI regulator built in.

I have more questions and comments but I'm done for the day and will come back to it later.

 

[TJBaker57]

still don't know if it's actually being activated by the system

I wanted to do an in vehicle test of my setup for activating vent and purge valves yesterday but I had to trap 4 of my 7 shop cats for vaccines. That turned into a bigger ordeal than I had anticipated.

So I hope to get to the in vehicle tests today. I have it setup in Car Scanner and it works on the test bench so I believe it should work in vehicle.

 

[TJBaker57]

tried adding evap vent valve status from that but haven't got it right yet. More on that later.

The formula/equation syntax is quite different from Torque Pro to Car Scanner ELM.

Try this...

edit: Car Scanner formula are case sensitive...

 

[dhanger]

I wanted to do an in vehicle test of my setup for activating vent and purge valves yesterday but I had to trap 4 of my 7 shop cats for vaccines. That turned into a bigger ordeal than I had anticipated.

Hope you don't come down with Cat Scratch Fever...

So I hope to get to the in vehicle tests today. I have it setup in Car Scanner and it works on the test bench so I believe it should work in vehicle.

I'm in no rush, I have until May of '27 to re-register when I'll need to smog test. As a practical matter I would like to have it fixed before May this year when the Arizona heat engine starts stoking up.

I could have taken this in for repairs but like most people here I'd rather do it myself and I'm having an enjoyable time learning this stuff. I'm a long time shade tree mechanic starting in the 1960's repairing my own bicycles as a boy, so I've had to live through the huge transition between purely mechanical/electro-mechanical to what we have now and I haven't really kept up with the transition that well, especially working a full time job. Now I'm retired I have more time to try to catch up, well, if you can call working on a 20+ year old car 'catching up', LOL.

 

[dhanger]

The formula/equation syntax is quite different from Torque Pro to Car Scanner ELM.

Try this...

edit: Car Scanner formula are case sensitive...

Okay, that explains why I kept getting a formula error using the one from the spreadsheet even after checking my typing a zillion times LOL. So that means that everything in that spreadsheet only works in Torque?

 

[TJBaker57]

Okay, that explains why I kept getting a formula error using the one from the spreadsheet even after checking my typing a zillion times LOL. So that means that everything in that spreadsheet only works in Torque?

Well maybe somewhat. (hows that for an evasive answer)

I actually have not looked at that spreadsheet.

Torque Pro has functions which Car Scanner lacks or implements differently.

Torque Pro grabbed the lions share of the market by being possibly the first app to come on the scene. However they fell quite short in updating and documentation.

Car Scanner is a more recent entry but is frequently updated and has decent support by the developer. Plus Car Scanner far outshines Torque Pro in the ability to record and display live data.

 

[AmpOverload]

Here's the graph from both tests:

View attachment 118504

I vacuumed down to about 4.5V (didn't have a vacuum gauge so don't know what that equates to and didn't want to exceed 5V) and then observed the decay for 5 minutes on both tests. Both ended about the same voltage so I think that's a good indicator that the vent valve is sealing good mechanically. I haven't seen any specs on what decay rate the PCM is basing the large leak on so nothing conclusive in that regard.

Aside: What's up with those highly irregular time values on the X axis of that Car Scanner graph?

Using your graph, my data, and math I won't bother to show here, your vacuum decay seems about 60% faster than the one I saw during careful monitoring of the automatic EVAP test on a 2004 Buick Century, so it seems like you really do have a leak somewhere. Of course, this would be better if the FTP PID in pressure units (not the sensor reading in volts) was used here, but the observation might still be relevant.

My "smoker" is a DIY (paint can) thing, so I don't know what the instructions for a typical commercial automotive smoker say, but I saw this a few years ago in a GM service manual, in case it helps:

Introducing smoke in 15-second intervals may allow smaller leak areas to be more noticeable. When the system is less pressurized, the smoke will sometimes escape in a more condensed manner.

I also once saw a suggestion that a laser pointer can help locate hard-to-see smoke leaks. I've never tested that idea myself, though.

 

[dhanger]

Aside: What's up with those highly irregular time values on the X axis of that Car Scanner graph?

That was just my first attempt at graphing while testing. Only valid areas are the 2 obvious descending graphs, the rest is just messing around before and between the 2 tests. I didn't use a separate graph for them.

Using your graph, my data, and math I won't bother to show here, your vacuum decay seems about 60% faster than the one I saw during careful monitoring of the automatic EVAP test on a 2004 Buick Century, so it seems like you really do have a leak somewhere. Of course, this would be better if the FTP PID in pressure units (not the sensor reading in volts) was used here, but the observation might still be relevant.

Would be nice to know what values GM uses to get a better idea, but my gut feeling was that it seemed a pretty rapid drawdown to me. Converting to pressures would be nice in the formula but so far in all my research I haven't seen the actual range of values that GM assigns to corresponding voltages. I did read in the manual (and other sources for other vehicles) that 1.5V represents atmospheric pressure equalized internally, i.e. the vehicle at rest. So <1.5 is positive pressure and >1.5 is vacuum (negative pressure).

One of the takeaways of this test for me was confirming that the vent valve was holding as good as manually sealing off the vent hose. Again, no confirmation that the PCM is actually changing the vent valve state.

My "smoker" is a DIY (paint can) thing, so I don't know what the instructions for a typical commercial automotive smoker say, but I saw this a few years ago in a GM service manual, in case it helps:

I also once saw a suggestion that a laser pointer can help locate hard-to-see smoke leaks. I've never tested that idea myself, though.

I'll try those suggestions next time I attempt it. This is the smoker I used:

https://www.amazon.com/dp/B07JVT1NG2/?tag=gmtnation-20

Plus accessories:

https://www.amazon.com/dp/B0FN2NQZFG/?tag=gmtnation-20

 

[dhanger]

Thinking about it some more it seems to me that the decay rate would be somewhat dependent on the volume of fuel in the tank, but maybe not enough to matter. Manual says that one of the conditions for the large leak test is the tank being between 15 and 85% filled:

P0455 — 2004 Chevrolet TrailBlazer 2WD L6-4.2L VIN S Service Manual | Operation CHARM

Detailed repair manual for the 2004 Chevrolet TrailBlazer 2WD L6-4.2L VIN S.

tinyurl.com

 

[AmpOverload]

That was just my first attempt at graphing while testing.

Sorry if it seemed otherwise, but I wasn't questioning your Car Scanner actions or testing in any way. Your graph was very useful. I was highlighting the app's poor, inconsistent labeling of the X-axis time values. I consider that labeling to be a bug. I tested a beta version of that app about a year ago and found some glaring bugs. I just grabbed a recent beta (in part based on this thread's use of that app) and found that many of the old bugs are still there along with this (new-to-me) axis-labeling bug.

That thread that I linked to earlier (in my first post in this thread) showed a nice graph of FTP in both volts and pressure. But since that's on a 2004 Century, I have no way to know if the mapping of volts to pressure is the same on that vehicle as it is on your vehicle.

I don't want to distract you from the path you're on, but if you can predict when the automatic EVAP test will run, it's nice to be recording all the right PIDs while that happens. On a typical Buick of that era, the EVAP test runs a few minutes after a cold engine start. But the TrailBlazers might behave differently.

 

[dhanger]

Sorry if it seemed otherwise, but I wasn't questioning your Car Scanner actions or testing in any way. Your graph was very useful. I was highlighting the app's poor, inconsistent labeling of the X-axis time values.

I hadn't even noticed that before, it is a bit wonky. I do know that I ran both tests for exactly 5 minutes so at least they're consistent with each other test time wise.

I don't want to distract you from the path you're on, but if you can predict when the automatic EVAP test will run, it's nice to be recording all the right PIDs while that happens. On a typical Buick of that era, the EVAP test runs a few minutes after a cold engine start. But the TrailBlazers might behave differently.

Since TJBaker posted the correct formula for the evap valve state I went out this morning and ran a cold engine test. The manual page I just linked earlier shows the test conditions which do imply that the test runs while cold but there's an inconsistency that doesn't make sense so maybe I'm misinterpreting that part, which is:

- The engine is running.
- The ignition voltage is between 10-18 volts.
- The barometric pressure (BARO) is more than 75 kPa.
- The fuel level is between 15-85 percent.
- The engine coolant temperature (ECT) is between 4-30°C (39-86°F).
- The intake air temperature (IAT) is between 4-30°C (39-86°F).
- The start-up ECT and IAT are within 9°C (16°F) of each other.
- The vehicle speed sensor (VSS) is less than 121 km/h (75 mph).

So the coolant temp and the intake air temp have to be within 16F of each other, but since the engine warms up past that range within a minute or two I don't see how the test has an opportunity before that happens. I must be misunderstanding that part. Anyway I ran the engine until over 190F coolant temp and the vent valve did not activate during that time.

 

[dhanger]

- The engine coolant temperature (ECT) is between 4-30°C (39-86°F).

More I think about it this must be a typo/error in the manual. Here in Arizona in the summer you can wake up to a cold engine that's already out of range. How can this possibly be correct? Or I'm really, really misunderstanding...

 

[AmpOverload]

More I think about it this must be a typo/error in the manual. Here in Arizona in the summer you can wake up to a cold engine that's already out of range. How can this possibly be correct? Or I'm really, really misunderstanding...

Actually, I think it's correct. Or, if it's wrong, it's widely duplicated incorrectly in various service manuals! For comparison, Ford/Lincoln/Mercury vehicles of that era use 40-100 degF as the limits (but they don't run the EVAP test while idling after cold startup). IIRC, several states with periodic inspection requirements allow passing with one failed vehicle monitor. Not sure about AZ, though.

Can you please report the 5th character ("platform") in your 2004 TrailBlazer's VIN? It should be "S" or "T", I think. And is the 8th character ("engine type") "S"? (When I run Tech2Win with my vehicle simulator, it helps to know this.)

 

[dhanger]

Actually, I think it's correct. Or, if it's wrong, it's widely duplicated incorrectly in various service manuals! For comparison, Ford/Lincoln/Mercury vehicles of that era use 40-100 degF as the limits (but they don't run the EVAP test while idling after cold startup). IIRC, several states with periodic inspection requirements allow passing with one failed vehicle monitor. Not sure about AZ, though.

I can only figure I just don't understand but I guess it's really not important. I just thought it would be a simple quick test to see if the vent valve was being activated or not--start the engine cold and watch for valve actuation and watch the pressure change (or not). It still isn't logical for me--how could the test ever take place in the summer under those parameters? Oh well, I'm moving on from that.

Can you please report the 5th character ("platform") in your 2004 TrailBlazer's VIN? It should be "S" or "T", I think. And is the 8th character ("engine type") "S"? (When I run Tech2Win with my vehicle simulator, it helps to know this.)

5th character=S
8th character=S

 

[TJBaker57]

Aside: What's up with those highly irregular time values on the X axis of that Car Scanner graph?

Car Scanner rounds to the minute sometimes depending on the level of zoom applied.

 

[TJBaker57]

So I fiddled quite a bit here today and am confident that I have the particulars correct for activating the EVAP vent (closing) and Purge solenoids for the 4.2 LL8 engine from 2002 to 2005.

I have tested on my running 2002 4.2 and also a 2003 PCM and a 2005 PCM at the bench. If I had a 2004 I would test that as well but 2002 to 2005 use the same P10 PCM so I think it highly unlikely 2004 would be any different.

2006 and 2007 use a P12 PCM and those commands are entirely different.

1st image I closed the vent.
2nd image I activated Purge at 100%.
3rd image I activated purge and closed the vent.

The tank was practically full so the purge active and vent closed shot up rapidly and the engine labored under all that extra fuel vapor. Like pulling a choke partially closed on an old vehicle.

 

[dhanger]

By Jove, I think you've got it! That will be great to try out and be able to confirm one way or the other if the valves are cooperating with the PCM. Based on what I know at this point I see that as quite an accomplishment to get bi-directional control from a simple bluetooth dongle and app! What's even greater is having this kind of information on a public forum for the benefit of others as well.