Subaru WRX Forum banner
1 - 20 of 157 Posts

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #1 ·
PossumLink / Walbro GSS342 incompatibility thoughts.

I've been running a Walbro GSS342 without problems with my Unichip for many months. Immediately upon installing the PossumLink standalone ECU, I began experiencing fuel supply issues.

The problem is manifested as the vehicle stalling entirely while in operation, especially after a long (30-45 minute) period of cruising. It feels as though the fuel supply ceases, and examination of PossumLink logs shows this to be exactly the case. There are also audible cues as the pump can be heard to erratically switch between duty cycles and finally cut out.

A key-off, key-on cycle is needed to restore fuel pump operation.

On my way to the drag strip this evening; the fuel pump cut out no less than 10 times within a 5 minute period, stranding me in heavy traffic twice, and once in the middle of a busy intersection. Normally the vehicle will run uninterrupted for between 10 and 30 minutes before the problem happens again, but on this occasion I could only travel 100 yards before fuel flow would once again cease. I was running the tank near empty (light not on).

I pulled into a gas station and added two gallons of (nice cool) 91 octane. This allowed me to travel the remaining 10 minutes to the drag strip before experiencing another cut. At the strip I added 2 gallons of 104 octane, taking me to just over a quarter of a tank of gasoline. I managed to complete two quarter runs, but experienced another fuel cut half way down my third run, while spraying Nitrous. :eek: This time, I was able to keep the car running and cruise the remaining distance.

I left the track immediately, and filled up with 91 octane outside the drag strip. The car then ran the remaining 30 minutes to my home with no problems.

Phast and davidm_sh are experiencing a similar issue. I'm not sure if it is exactly the same problem or not. I have a long commute to work, so I seem to be more susceptible to this problem.

My variation on their theory: The pump is getting hot, possibly causing increased current draw. The PossumLink has some kind of failsafe (or unintentional trip) that is triggered when current draw for the fuel pump relay reaches a certain level. The full tank of gas helps keep the temperature of the pump down, and prevents or lowers frequency of the problem.

Link-USA are not being much help right now. I'm hesitant to get a different pump (such as an expensive PE) as there is no guarantee it will fix the issue. i.e Is this the responsibility of the Link, or is the Walbro at fault?

If anyone is interested, I have attached a log showing the portion of my quarter run where the pump temporarily ceased operation. The fuel flow appears to have ceased between count 533 and 535. Ignore the O2 values; I still have the MIL eliminator in place and its clamping the voltage.

Thoughts? Comments? Similar experiences? I commute 100 miles per day, so either the Link or the Walbro has to go! The car is rapidly attaining the point of being unsafe to drive in traffic. :eek:

-Pace
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #2 ·
Also; anyone know the meaning of the Flags1 and Flags2 fields in the log?

-Pace

p.s Experiencing a disruption in fuel flow while spraying a wet shot of Nitrous is something I hope none of you guys ever have to experience. My head whipped forward onto the steering wheel and I immediately glanced over to my laptop expecting to see a huge knock event. My motor got lucky. Underwear not so fortunate. ;)
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #3 ·
Interesting; if the logs are to be believed, my 565cc injectors went static shortly before the pump cut. :confused: :eek: :confused:

I wonder if the logs are showing some bogus data. Jay was experiencing that due to a Link bug triggered by electrical interference.

-Pace
 

·
Registered
Joined
·
295 Posts
I'm wondering if all this isnt due to interference. I'll let you know in a couple weeks after I get my new chip flashed. Oddly enough, nobody else has come forward with this problem on a Link except users in CO. I know for a fact that there are plenty of other folks running this same pump/ecu combo with no issues to speak of.
 

·
Registered
Joined
·
192 Posts
It's not a problem with your fuel pump at all. It's a problem with the LINK. There's a TPS 'bug', which causes TPS value to drop to a near zero or zero reading. I'm not a link user, but I believe there are new versions/revisions available which correct your problem.

If you look at your datalog (I've cut/pasted a bit below), you'll see that in frame 533 your Map is 180kpa and TPS is 101%.

In frame 534, your MAP is 91kpa and TPS is 10%.

There is no way that you were able to remove your foot from the gas pedal that fast... It's the TPS value that is causing your problems. TPS goes down to near zero, injectors shut off, blow off/bypass valve vents pressure from manifold.

You are actually hearing the fuel pump slow down... That's because the fuel pump is controlled by the fuel pump controller. It's a little box mounted in your trunk. It gets a TPS input from the ecu. Basically it gives the fuel pump a duty cycle to run, based solely on TPS value. When tps is 0-33%, fuel pump runs at 33%. When TPS is 33-66%, fuel pump runs at 66%. Anything over 66% and the fuel pump is at 100% duty cycle.

They do that for emissions - so there is less fuel flowing back into the tank through the return line, less purge gasses..

Bottom line, your fuel pump is fine... It's the link that's troubling you. At least that's my guess.

------------------------------------------


Count RPM MAP TPS OXY ETEMP VOLTS ZONE %FF ADV %WG KNK IAT IDLE KPH ACTTPS LAMBST IDLEST KNKCNT CURRFL CURRIG FLAGS1 FLAGS2 Markers RadFan NeutPk AirCon MAPLim RPMLim TGVClsd TPSCls Knock

532 5600 182 101 67 86 13.7 755 75 15 96 84 32 34 86 0 T 0 108 15 12 0 I Off On Off Off Off Off Off Off

533 6200 180 101 67 86 13.8 760 80 16 97 82 32 34 89 0 T 0 108 16 12 0 I Off On Off Off Off Off Off Off ;

534 5300 91 10 67 86 13.8 450 22 25 115 93 32 34 84 0 T 0 95 19 12 64 I Off On Off Off Off Off On Off ;

535 5800 27 10 68 86 13.8 155 7 31 117 40 32 34 87 0 S 0 131 31 12 64 I Off On Off Off Off Off On Off ;

536 5500 20 10 66 86 13.9 155 3 31 116 26 32 34 85 0 S 0 131 31 12 64 I Off On Off Off Off Off On Off
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #6 · (Edited)
So why is the car entirely stalling when this happens? That log shows the first and only time that the car has remained running... all the previous times (20+) the car will stumble for a few seconds at idle, die if you touch the gas, then eventually die anyway. My fuel pump is audible at idle at any duty cycle. When this occurs you can hear it isn't running at all.

Not sure if the comment regarding being able to lift my foot that fast holds up. As soon as the car stumbled I did whip my foot clean off the throttle...

Not arguing with you BTW. It's just that I have observed this carefully now for two weeks and your answers raise more questions than they address. I am in communication with Link-USA and this problem appears to have them baffled. Surely if it was the known TPS issue they would have asked me to send in for a re-flash.

Also, why does this happen only when I am low on fuel. The less fuel I have, the more frequently it will occur once I pass the 30-minute threshold.

Thanks much for the feedback...this is good discussion.
-Pace
 

·
Registered
Joined
·
160 Posts
If you remember you pull your foot off the gas after you notice the stumble AND you see from the datalog that the TPS signal drop occurs at the same time the stumble occurs, then you clearly have an ECU glitch.

That is, if you agree. I would definitely adjust the TPS signal regardless.

I have the GSS342 and used to run it with the Link. Never had that particular problem, but I bought one of the more recent boards.
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #8 ·
WizardBlackWRX said:
If you remember you pull your foot off the gas after you notice the stumble AND you see from the datalog that the TPS signal drop occurs at the same time the stumble occurs, then you clearly have an ECU glitch.

I have the GSS342 and used to run it with the Link. Never had that problem, but I bought one of the more recent boards.
I'm not disputing that it's a Link glitch, but the fuel pump is evidently playing a role in the problem. With all of us that experience this, there is a pattern to 'when' the problem occurs. At that point, exactly how frequently it occurs depends repeatably upon how much gas is in my tank, so I can only deduce that the current draw of the pump is triggering the problem. Maybe it is an interference problem as Jay said that is causing the TPS problem, and maybe that interference is caused by the pump or the relay!?!

It's worth noting that nobody is yet reporting this problem with any other fuel pump... just the GSS342.

-Pace
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #9 ·
WizardBlackWRX said:
If you remember you pull your foot off the gas after you notice the stumble AND you see from the datalog that the TPS signal drop occurs at the same time the stumble occurs, then you clearly have an ECU glitch.

That is, if you agree. I would definitely adjust the TPS signal regardless.
..
Agree on that first assertion, but I don't understand that last sentence...?

I will run the TPS bug by the Link guy and show him this latest log.

-Pace
 

·
Registered
Joined
·
47 Posts
The Link can set flags on up to 10 fields in the digital display. I assume that this is a way to track these flags and values.
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #11 ·
bcblues said:
The Link can set flags on up to 10 fields in the digital display. I assume that this is a way to track these flags and values.
Doh! That's what it is...

Gonna call Link today. Will report back with the verdict.

Thanks,
-Pace
 

·
Registered
Joined
·
3,501 Posts
Someone, probably at (phast) mentioned that the Walbro maybe so high an impendance for the link ECU, but I'm sure that the circuit is relayed, so I don't see that as the issue.

The Link, while powerful, does have it's share of bugs. With so much fuel pressure on tap and a TPS signal error, you could run into a fail safe condition from the Link which kills the fuel pump.

I'm certain that the Possumlink guys will sort it out with a firmware update.

Where's my UTEC? I want to ask firmware programming questions on an ECU that I don't understand too. ;)

The price for high performance is often an engine management struggle. You rocky mountain guys are hard-core link fans and if you work with Possumlink, you'll fix all the bugs and have very fast, reliable cars!

I'm trying both the TXS's UTEC and Godspeed's stock ECU reprogrammer. I'd rather let an all-pro-programming-geek solve my ECU issues while I work on getting more flow in and out of my engine.

-Jim
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #13 · (Edited)
Just got off the phone with Tariq @ Link-USA. Here's the verdict:

I am experiencing two distinct problems. One is not well understood and is related to the current draw of the Walbro GSS342 pump causing the Link to shut down that circuit. They are investigating. This is what I experienced on the way to the drag strip last night. When it occurs, the car does not continue to run and the Link needs an ignition reset. This is consistent with my findings that a full tank of gas helps alleviate the problem (cool pump).

The second problem which I experienced during the quarter run is thought to be Jay's interference problem. The TPS does not cause the problem, rather the TPS is just a symptom of the bug and occurs *after* fuel flow has been disrupted, although the logs misrepresent that the TPS caused the problem. This bug just temporarily disrupts fuel flow and the car does not stall. For this there is a new chip available. He is going to examine my logs to confirm that is what I experienced, and if so I will be sending my board in.

I'll give them a week to work on the Walbro problem, otherwise I may try a different pump.

-Pace
 

·
Registered
Joined
·
176 Posts
WizardBlackWRX said:
If you remember you pull your foot off the gas after you notice the stumble AND you see from the datalog that the TPS signal drop occurs at the same time the stumble occurs, then you clearly have an ECU glitch.

That is, if you agree. I would definitely adjust the TPS signal regardless.

I have the GSS342 and used to run it with the Link. Never had that particular problem, but I bought one of the more recent boards.
Gotcha... Hey Mark I am the guy who bought your Link and I have also experienced the EXACT same problem; interstate cruising/lot of light throttle over long periods of time and then BAM car just dies and throttle becomes unresponsive, pull over push the clutch in and the rpms go straight to 0. After a restart everything seems fine.

Which REALLy has me concerned that there are 1-3 different board versions (phast, myself, and pace) and we are all experiencing this problem.

Personally I think it's a "new" bug that has never been discovered because honestly how many people ACTUALLY drive their Link ECU'd cars on road trips (100+ miles) at "normal" cruising conditions?

There are about 4 of us with this problem now and it seems, like pace said, that Link is just scratching their heads so far. It seems like they want to get it working but, as of right now, don't know where to start.
 

·
Registered
Joined
·
63 Posts
Thanks for this

Pace,

Thanks for this - the first problem is most definitely what I am experiencing. I am going to try one or all of the following fixes:

1. Supra fuel pump (plenty of people run these and Links together with no issues.)
2. Hardwiring the fuel pump with a relay and tie into the ignition switch. This is not my first choice, but I think it would solve the problem.
3. Some kind of electrical "booster" like a capacitor in the line - this was suggested to me by an RX7 guy who had similar problems with a Haltech ECU and had to alleviate it. He installed a capacitor in the circuit and rewired the main power and grounds to his fuel pump with larger gauge cable. I dont know much about this, but he assures me it would make a difference.

Let us know if Link tells you anything else about this.

Cheers!!

JaMa
 

·
Moderator
Joined
·
1,263 Posts
what sort of diagnostic tools do your guys have access to?
CROs, data loggers etc. It seems there are a number of signals that are not logged by the link that could be crucial to finding the crux here.

From what you are all saying it would be interesting to monitor the supply voltages on the pump and the link to see if the supply rails are being pulled low by the pump load.

If this was the case a capacitive power supply filter could work.
It would also help to determine what is triggering the link 'bug(s)'.

What data rate is the link table that was provided?
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #18 ·
I'll report back when I hear anything. Part of me is hoping that the new chip for #2 (if they let me have it) might have a side effect of fixing #1 if the problems are related, which they seem to be at least superficially.

Failing that, my next approach was to go with a Supra pump. Until then, I'll just try to keep the gas level high. The cold weather should help.

I have no doubt that I am experiencing the known interference/TPS problem, so I WANT that new chip. My logs show my injectors as going static for two counts. No way. There's nothing in my fuel map that could cause that on 565cc injectors. Jay was seeing similar screwiness in his logs and they confirmed it was the interference problem.

I think David hit the nail on the head. I probably commute more miles in a single hit than most people that have this setup...

-Pace
 

·
Registered User
Joined
·
2,210 Posts
Discussion Starter · #19 ·
aussieinstlouis said:

..

What data rate is the link table that was provided?
I asked Jay the same question last week. He wasn't sure. My guess would be around 4Hz.

I think Jay and David had talked about trying to monitor current flow to, and/or potential across the pump. Don't know if they ever pursued it.

-Pace
 

·
Registered
Joined
·
63 Posts
<10 miles

I go less than 10 miles and I can get this problem pretty much on demand as long as the car gets parked hot and restarted.

???


JaMa
 
1 - 20 of 157 Posts
Top