Jump to content
Ultimate Subaru Message Board

2000 Legacy Difficulty Starting - Fuel Pump?


Recommended Posts

Hello there,

This is a 2000 Legacy L Wagon MT.  I'm experiencing three separate categories of symptoms.    I suspect a bad fuel pump and possibly a starter motor starting to get worn. 

The first is the biggest problem: difficulty starting, especially when the engine is cold.  The starter motor turns over the engine, but I don't hear ignition until a seemingly random time later when I try it the third or fourth or fifth time.  Doesn't seem to happen when the car has been run recently, just after it's been sitting.  The engine will also often give a false start, sputtering and dying.  When I try it again right after that, it usually starts up properly.  This makes me wonder if the starter motor is not performing well enough to start the engine properly.

Second symptom is sudden loss of power.  This happens very infrequently, but it's happened often enough that I'm convinced I'm not imagining it.  I will suddenly lose power, usuallly in the lower RPM range at mid speeds.  This is part of why I think it could be the fuel pump.  A failing fuel pump could also explain the hard starting issues (but why only when the engine is cold?).

Third symptom is a woo-woo-woo sound under the hood (or possibly under the car) at random times, like an owl calling out on a loop.  It happens at random times, perhaps once or twice a day that I notice.  it isn't super loud but it's there. I haven't tried to pinpoint the sound yet but it doesn't seem to be coming from the fuel pump., more under the hood closer to the passenger side. This is the most mysterious issue.  It doesn't seem to cause a problem so I haven't tried to fix it.

The pump is a Delphi.  I replaced it preemptively due to the defective retaining cap issue, just shy of a couple years ago.  I've also replaced plugs and wires, as well as all filters and normal maintenance items.  I already checked the fuel filter for clogging; there was none.  I replaced a friend's pump in a Ram 2500 with a Delphi, and it gave out after 6 months.  I don't know if that's because the gas tank wasn't cleaned as was recommended, or because these pumps have a reputation for failing.

I also suspect the starter motor.  It's old and had a solder connection inside the solenoid come loose, causing it to not work until I fixed it.  The solenoid plunger and contacts were also heavily worn.  I was considering just replacing those, but perhaps a new starter would rule out the motor itself going weak, causing starting difficulty. 

There is also always the possibility of the ECM going bad.  Do you guys have any clues as to what might be causing these issues?

Edited by dirty_mech
Link to comment
Share on other sites

overfilled refrigerant can cause a hooting noise. This time of year, a/c is still used for defrost. Was it serviced recently? you could try turning the comp. off next time you hear the noise.

If the intake filter 'box' is not seated at the bottom where the tabs fit into slots, it can make a whooshing sound.

 

any codes stored?

Link to comment
Share on other sites

Hi Texan,

No codes stored, other than a P0420 cat code that's been a perennial issue for my busted cat.  Going to replace that later.

It could be refrigerant.  I tried adding refrigerant at the start of the summer and it squealed pretty hard.  Even when the PSI to the recommended level, it would steal squeak occasionally.  I haven't used it often, but I found that the compressor sometimes seems to run when the fan is on, even if the button for AC is off.  I'd have to fix that issue first before I could appropriately diagnose the issue.  But I think that's a strong possibility.  If you're familiar with the AC-staying-on issue, let me know.

Regarding the main issue, it's finally stopped working for me at all. I tested for fuel pressure and there is nothing there.  Definitely a fuel delivery issue.

I swapped out the fuel pump for a known good one, and still nothing.  Tested at the connector with the key in the ON position, and I got sub-12 volt voltages at two pins.  The connector is a six-pin one, which includes for the fuel level sending unit and perhaps a low fuel sensor.  A diagram of my readings is below with some photos.  Can anyone help me figure out why it would be sending low voltages to these pins?  Going to try and figure out if there are any relays that might have gone bad.

Anyone have experience with the problem of low power going to the fuel pump?

34828345_IMG-0253-Copy.jpg.3198fb48db41ae0c2d7eee53162ccc5c.jpg
IMG-0254.jpg.907df2492c2a553e0811217d160add8d.jpg
IMG-0256.jpg.39b5ad970a40649ad108200fe632c619.jpgIMG-0255.jpg.045e07227c5623b45494ee256c2fcab0.jpg

Link to comment
Share on other sites

not sure I can help much but, to be clear, it now cranks but won't start?

fuel pump is odd in that, it runs only for a very short time immediately after key on, to fully pressurize the rails. You should be able to hear it. If you don't, there could be a bad relay somewhere.....then I guess it starts-up again sometime after starting.

tank has gas in it right? no problems with fuel gauge reporting accurately?

Edited by 1 Lucky Texan
Link to comment
Share on other sites

Hi all,

Ding ding ding, you were right Imdew.  Fuel pump relay.  I changed the fuel pump preemptively to avoid problems, but not the relay, and this happened. What a pain.

Long story short, my car is working again.  I'll update you guys on what I found out about the pump and relay wiring later.  I've got a couple other posts I need to update as well.

Thanks for all the help and support!

Link to comment
Share on other sites

Here is what I found with a multimeter as memory serves.  These suckers do go bad.  I tested mine straight at my battery using some extra multimeter probes as test leads.  However it wouldn't be a bad idea to just test it by replacing it.  If I had realized these go bad, I would have replaced it preemptively.  Makes me wonder what other relays I should replace. 

IMG-0253 - Copy.jpg

IMG-0262.jpg

IMG-0264.jpg

IMG-0267.jpg

Link to comment
Share on other sites

On 3/1/2021 at 2:16 AM, dirty_mech said:

Makes me wonder what other relays I should replace. 

I had to replace the Main Relay in our 97. It is a dual contact relay. (2 in one) It was mounted on the bracket beside the fuel pump relay.

Our 95 has over 473 thousand miles on it and still has the original fuel pump relay and main relay.

Don't know if this is what you are hearing, but the air intake system can sometimes make a humming sound. Our 97 was terrible, so I replaced it with a 95 system (from the air filter box to the throttle body) and it was much quieter.

Link to comment
Share on other sites

  • 1 month later...

 

On 3/2/2021 at 6:37 AM, Rampage said:

I had to replace the Main Relay in our 97. It is a dual contact relay. (2 in one) It was mounted on the bracket beside the fuel pump relay.

Interesting!  I didn't even realize cars had a main relay.  Just a couple weeks ago my starter relay (also near the fuel pump relay) went out.  I was ready to replace the starter whose solenoid I had just rebuilt, but it turned out to once again be the relay.  Took apart the old relay for shits and giggles, and the only thing wrong with it was a little bit of black arcing deposits on the two contacts.  I don't know if 20 years of regular use is just a typical lifespan or what, but I might get a head start on replacing the main relay and any other critical relays that could leave me stranded if they go bad.  I now keep an extra fuel pump relay in my ashtray, and an extra fuel pump in my cargo area, just in case either goes bad on me in the future.

Never had my air intake system make noise to my knowledge, but it's possible.  I'll check to see if any of the air housing is loose, and check the filter too.

Edited by dirty_mech
Link to comment
Share on other sites

On 3/1/2021 at 7:07 AM, lmdew said:

Your pictures show you in the circuit when measuring voltage.  Never a good idea to be touching the pins with skin/fingers.

Well, let's not be silly.  Not always a bad idea either.  At 12V, in relatively low-current and low-impedance circuits like this one, you're at no risk of either shock or affecting the readings.  In 12V circuits at high current, there's a burn risk.  But here it's not going to matter.

Edited by jonathan909
Link to comment
Share on other sites

On 3/1/2021 at 7:07 AM, lmdew said:

Check out Load Pro on youtube.   Great tool, to add to your meter.

Can you give me the Cole's Notes version?  I started looking at the youtube vids on this probe and the first thing I got was some knucklehead complaining that there are always gonna be haters, whaddayagonnado?  Lots of "users" blabbering on about it without actually getting to the point.  Manufacturer's web site isn't much more informative.  I mean, I don't need a half hour video.  If you can't describe it in two minutes or less, you don't understand it.

Is this just about adding a pushbutton and a resistor to the probe to (effectively) lower the meter's input impedance?

Edited by jonathan909
Link to comment
Share on other sites

3 hours ago, jonathan909 said:

Well, let's not be silly.  Not always a bad idea either.  At 12V, in relatively low-current and low-impedance circuits like this one, you're at no risk of either shock or affecting the readings.  In 12V circuits at high current, there's a burn risk.  But here it's not going to matter.

When using a volt meter to test one end of a wire that is open, or has a very high resistance at the other end, touching the probe with your finger will cause voltage from your body to show on the meter. How much voltage depends on how much static your body has built up and how sensitive the meter is.

Also, extra long leads on the meter can pick up stray RF signals from numerous sources and display voltage that is not really there.

Link to comment
Share on other sites

That's not really how things work in these situations.  Nobody is going to confuse a little noise with a legitimate DC reading.  Also, this has nothing to do with "static" charge or stray RF.  You're grounded quite enough for this not to be the case at all, and your skin resistance lowers the impedance of the circuit in question enough to make any of the tiny effects you're describing disappear.  Besides, the measurements of interest are DC volts, not the AC microvolts or millivolts that may be riding on it.  What you're describing are non-issues when taking readings in a car.

If you want to talk about an RF test bench, that's another matter altogether.

Edited by jonathan909
Link to comment
Share on other sites

13 hours ago, jonathan909 said:

Can you give me the Cole's Notes version?  I started looking at the youtube vids on this probe and the first thing I got was some knucklehead complaining that there are always gonna be haters, whaddayagonnado?  Lots of "users" blabbering on about it without actually getting to the point.  Manufacturer's web site isn't much more informative.  I mean, I don't need a half hour video.  If you can't describe it in two minutes or less, you don't understand it.

Is this just about adding a pushbutton and a resistor to the probe to (effectively) lower the meter's input impedance?

basically - it puts a simulated load on the circuit to make sure the circuit is working as it is supposed to.

kind of like when testing alternators, you first test with all accessories off to get a baseline, then turn things on to make sure it is working properly when loaded.

Link to comment
Share on other sites

Yeah, that's pretty much what I expect - except that it's not a "simulated" load.  With a live circuit either it's a load or it's not.  If you're doing a computer simulation, you can simulate a load along with the rest of the circuit's behaviour.  But if you've got a real system and are trying to measure it with a real meter, it's going to have to be a real load.

The problem I have with the lack of information on this thing is that the details matter - without knowing the switchable load's resistance or power rating, it's difficult (impossible, really) to know how the system is going to behave - or how easily the load itself can be burned out.  So without some further data, I'd file this under "sketchy".

Link to comment
Share on other sites

On 4/26/2021 at 10:14 PM, jonathan909 said:

You're grounded quite enough for this not to be the case at all,

If you have clothes and shoes on you are not grounded to the vehicle.

Basically, AC voltage is a combination of increasing and decreasing, plus and minus DC voltages at a certain frequency. So is RF, but at a higher frequency.

Link to comment
Share on other sites

7 hours ago, Rampage said:

If you have clothes and shoes on you are not grounded to the vehicle.

Basically, AC voltage is a combination of increasing and decreasing, plus and minus DC voltages at a certain frequency. So is RF, but at a higher frequency.

I don't want to beat this to death, but you should understand that you're trying to give basic electricity lessons to an EE who's been in the field for more than 40 years.  You might get a little sense of my background if you search some of my old postings, particularly the thread about figuring out why it's so hard to get the Alpine remote lock fobs to sync with old cars (I went a little crazy with that one, to the amusement of some here).

To your points:  You don't have to be naked to not be carrying any static charge.  As soon as you touch the car you're bleeding it off.  And you'd be amazed at how little dielectric strength (that's "insulating power" to you) most shoes offer, whether you're handling a defective power tool or getting struck by lightning.

As for your second lesson, please donate your DVM to someone in need, because it's just confusing you unnecessarily.  You should instead be using an old-school analog-movement VOM, because it'll give you nice, solid, accurate readings for all the things that matter in the car (short of digital communication channels e.g. between the ECU and TCU, for which you need oscilloscopes, logic/protocol analyzers, etc.), and you won't have the DVM's dancing digits making you believe that things you saw once in a youtube instructional video matter when they don't.

In closing, I don't like to brag, but you're kind of needling me into it.  This picture is of the discharge electrode of a Tesla coil (built by a good friend of mine in the Bay Area).  The electrode is 2 meters in diameter, and on top of the coil's secondary, 10 meters in the air.  I am standing inside the electrode (not just for fun, either, though there was a lot of that - I was up there taking measurements), and together we are at a potential of approximately 1.5 megavolts.  If anyone ever asks whether Faraday cages really work, this is what you show them.

Now, since I've shown you that I'm reasonably well acquainted with both big volts at high frequencies and little volts at DC, was there anything else you think I need a little brushing up on?

tesla23pwr.jpg

Edited by jonathan909
Link to comment
Share on other sites

Oh - and after all that, if you STILL have doubts about my knowing what I'm talking about, next time you're in the neighborhood of New Kensington - about 3 hours from you, according to google maps - you can pay a visit to a working computer museum another friend of mine runs there (and I'm talking about big vintage iron - mostly DEC - not little crap like VIC-20s), and I'm sure he can straighten you out.  PM me for the location details.

Edited by jonathan909
Link to comment
Share on other sites

Woah this old thread blew up outta nowhere!   And on a comment I totally missed.  Props to Jonathan for rocking a little of the electrician version of the marine copypasta on Rampage.  You guys crack me up!  My first though was "Theoretically, yes, but I've never noticed that kind of interference on a meter, and if you touch the vehicle metal then there would be no static charge."  And that sounds like what it boils down to.  I'm not sure if I grounded myself but I probably was because of the doorframe or body metal at the pump. 

But I also work on AC and as far as forming habits go, I'd say you want to avoid carelessly touching the probe or any piece of metal that might be energized, and even when it is DC just to reinforce the habit.  But it's hard to take a picture with one hand and hold a probe with the other hand holding two probes to show a reading on the meter.  That's why that picture looks awkward; I'm trying to hold those two probes like a pair of crossed chopsticks and keep pressure on them.

I'm sure most here already know but I forgot to mention the 12v at pump connector will only show 12v for a moment when you first turn the key on then go off after the pump pressurizes the fuel line.  You have to test it for 12v right when you turn the key then you can eliminate a wiring issue as a cause before deciding to replace the pump. 

Edited by dirty_mech
Link to comment
Share on other sites

On 4/26/2021 at 7:02 PM, Rampage said:

Also, extra long leads on the meter can pick up stray RF signals from numerous sources and display voltage that is not really there.

I've seen this in Romex or other conductor with more than one wire on 120v AC circuits.  Wire A will be energized at 120v, and Wire B which is neither grounded nor energized will show some amount of voltage (usually within 5 - 50 volts) when tested, but only because the energized Wire A is adjacent to it in the same run of line.  I don't necessarily think it has the amps to actually shock you if you were to touch it, and I'm guessing the voltage would probably disappear as soon as you grounded the wire.  I'm not sure what this phenomenon is called or about, but I might have seen it referred to as sympathetic voltage.  Maybe Jonathan has a clue.

Edited by dirty_mech
Link to comment
Share on other sites

Sure.  I've never noticed it myself, but I guess you can get some inductive (like a transformer) and capacitive coupling between the two wires, and how much would depend on how long the run is.  But you probably wouldn't be able to measure much between the two because there'd be little phase difference.

Link to comment
Share on other sites

15 hours ago, dirty_mech said:

But I also work on AC and as far as forming habits go, I'd say you want to avoid carelessly touching the probe or any piece of metal that might be energized, and even when it is DC just to reinforce the habit.  But it's hard to take a picture with one hand and hold a probe with the other hand holding two probes to show a reading on the meter.  That's why that picture looks awkward; I'm trying to hold those two probes like a pair of crossed chopsticks and keep pressure on them.

Yeah, probes are a whole thing.  For many years my preferred probes have been the extendable needle point probes made by Huntron:

https://www.jensentools.com/huntron-98-0078-mp10-microprobes/p/447me109

Fluke used to sell them as well, but not anymore, afaIk, which is a shame.  They may have replaced them with a version of their own.  Huntron's stuff (at least, their early products like the "Tracker") was bogus, attempts to bring troubleshooting to people who don't understand what they're doing.  Yet another friend a long time back called it "the new charlatanism".  But those probes were and are brilliant, only one of their features being that because they're so slim, you can easily handle them exactly like chopsticks.  The only thing that's not perfect about them is that they're relatively fragile - they're not made of a flexible/resilient plastic like most probes, and if you step on one it's a goner.  I've never actually fired anyone for allowing one to sit on the ground, but I have issued "the only warning you will get for this" a few times.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...