Car cranks but doesn't start

[jrallye]

Welcome, just before Christmas I bought an old non runner JDM 1994/5 Gen 2 Subaru Legacy Twin Turbo (EJ20H engine code) with the intentions of having something to mess with and fix up to it's former glory

To begin I'll walk you through what I've done so far (rightly or wrongly)

1. Washed the car
2. Cleaned the plugs + checked all fluids + little bit of oil in each cylinder
3. Tried to start the car cranks won't start realise the fuel pump is dead
4. Chucked in a working second hand fuel pump, pump turns on and pumps fuel to the rail again car cranks but won't start
5. Suspected the third party cobra immobilizer so carefully ripped it out and connected the OEM wiring all back together where other wires had been spliced in again cranks but won't start doesn't even splutter just cranks

Aaand that's it so far, I've done a ton of googling and think my next steps should be

1. Properly test for spark
2. Test the injectors?
3. Double check timing

Does this sound about right? Is there anything else I should do before hand? Anything to take into account because it's a boxer engine? Ideally I want to be spending as little as possible on getting it running before throwing big money at it

Thanks in advance

[jrallye]

E-bmw said:

Get in touch with the previous owner & find out the history of why it was laid up.

Do a compression test, it could be because the cam belt/chain (not sure what that will be) has gone south.

Have tried this but no luck the car was recovered from a collection that had been sat for years, the current owner basically left the cars after a marriage went south and asked someone to remove the cars from the property I bought the car off that someone so have no contact with the owner

[jrallye]

tapkaJohnD said:

jrallye,
You bought an old car so as to have "something to mess with and fix up to it's former glory" Which implies that you have experience and expertise in this. Yet you give up and appeal for help before you have even tested for a spark! Your diagnostic skills may not be up to this task.

Think logically, TEST for a spark! Two ways, take off an HT lead connect to a spare plug on the block, and see a spark when you turn over. Or use "Startyabarstard" spray into the air intakes. Even a brief cough of ignition indicates you have a spark.

Goodluck!
John

No experience or expertise in getting a non runner starter other than a 50cc mini moto so this is all new to me

See this is what I came here for, a little bit of abuse but a sensible answer, I knew I had to test for spark but I didn't want to get it wrong and cook myself

These engine are coil over spark so I'm assuming it's the same process albeit with a coil connected to the spark plug?

Thanks mate

[jrallye]

Oldandslow said:

Just to clarify spark check. Absolutely use a spare plug and leave the one in the engine. From my recollection Subaru plugs are a bugger to get at. Fit the coil on to the plug and touch the metal base of the plug onto the engine block so that it is earthed through the block, needs that electrical connection. Don't have too much light under the bonnet. Get some one to crank it for you and keep fingers clear of any spinning bits.

I've had the plugs out and whilst it wasn't fun it wasn't hard have since bought some better tools that will definitely help me out

Not a bad shout using a spare plug though, it'll show me whether or not the spark plugs are even getting anything from the coils

When testing for spark is it wise to unplug the fuel pump so as not to flood the combustion chambers?

[jrallye]

tapkaJohnD said:

You''re a brave man! These are not simple cars! You had good advice above, just come back for more! Good luck!
John

Brave or stupid? I haven't decided yet

I've got the entire weekend to play with this so going to get started and see where I end up after that

Edited by jrallye on Wednesday 1st May 09:31

[jrallye]

Oldandslow said:

If you're at it for a while I would. If you get spark the first couple of cranks then job done but if instead you go through a lot of fault finding iterations I'd not want fuel needlessly fed to the cylinders. Should be a fuse for the pump somewhere.

Edit just as another thought occurred, how fresh is the fuel? It can go bad and cause starting issues.

Edited by Oldandslow on Tuesday 30th April 17:28

I'll disconnect it to be on the safe side only takes seconds anyways

Had some old fuel in the tank so I drained it and put in 5 litres of fresh VPOWER before trying to start it the first time, will have to run and get some more though as that's a couple months old at this point

[jrallye]

UPDATE

So got busy today and found out the following...

• I have spark both from the coil through the test lead to the block and again with the coil straight to the spark plug (I have not tested every plug because it's a subaru and I really couldn't be bothered, but I now know the car is at least sending the signal to spark which is good)
• The resistance of the injectors are 11.2 across the board
• The injectors have 12v constant (when ignition is 'ON') from one wire and have 0.4v from the signal wire when cranking (I realise my cheap multi metre won't be quick enough to see the pulses but the fact it's showing voltage gives me a bit of hope)
• Despite the above there doesn't seem to be any fuel getting into the cylinder, the plug I looked at was wet but after filming down the spark plug well I can't see any fuel entering the cylinder
• There are no codes held in the ECU the EML flashes in the 'OK' sequence

Next steps in my mind are

1. Listen to the injectors to double check whether they're actually firing or not. If they're firing I'll move onto step 2
2. Figure out a way to test for fuel pressure, not done this before so will have to read up on it/ buy a pressure tester
3. The next logical step after this would be to check the timing, again not something I've ever done before so will have to read up/ acquire tools
4. Order new plugs as I dropped one under the intake manifold and can't get it back (worth buying regardless)

Does the above sound like a good plan or is there anything else I should think about?

Thanks in advance, feels like I'm getting somewhere now ish

[jrallye]

GreenV8S said:

Based on your earlier comment I think you're saying that as far as you know the car was a runner which has been laid up for a long time (how long?) and now won't start.

That suggests you aren't looking for any major mechanical problem but may have stale fuel, clogged injector, electrical failure or something along those lines.

To rule out the whole fuel system I suggest you disable the fuel pump, put clean dry plugs in (gapped correctly, preferably warmed first) and crank it over with a freshly charged battery / jump start while somebody sprays easy start / cold start spray into the intake. If you get any sign of life at all, that proves the ignition system is doing something so you know to focus on the fuel supply side. With any luck, it will actually fire up and run as long as you keep spraying.

This definitely sounds like a sensible next step if it coughs and splutters at least I know its working properly then can get to work making sure it's getting fuel

The only issue is when I carefully removed the cobra alarm system I could've messed up some wiring somewhere but hopefully not

I was told it's been sat up for 15+ years so once I know the thing runs I'm going to do a huge service including belts all fluids flushed and changed just to lower the chances of it self harming

[jrallye]

UPDATE

The car runs... on starter fluid but it runs

Fuel comes out the filter so that's not clogged

After listening for the injectors with a big screwdriver and there's nothing, so pretty certain they're not firing

Any ideas on what to look at next? I'm not going to rule out fuel pressure but it seems strong and comes out the filter

Had an check engine light come on after running the engine and when reading the code the check engine light flashes *long *long *short *short *short or code 23 - "Air flow meter or circuit (exc. Justy); Pressure sensor (Justy)" but I cant see that stopping the injectors for firing but will be something to look at in the future

EDIT: Looking at another source for the engine codes I found the following
23 | Air flow sensor | Abnormal voltage input entered from air flow sensor | Controls the amount of fuel (injected) in relation to engine speed and throttle sensor position
This makes me think it might be the airflow metre that's stopping them from firing but I'll do some more digging

Edited by jrallye on Tuesday 7th May 21:25

Edited by jrallye on Tuesday 7th May 21:31

[jrallye]

GreenV8S said:

That's an encouraging start.

I suggest you check the injectors electrically and check the fuel pressure. The with have a steady +12V on one terminal and the other terminal is grounded by the ECU to open the injector. So verify you have +12V on either terminal while the engine is cranking/running. You should also see the other side being pulled down to 0V ish once or twice per engine cycle. A NOID light connected across the injector makes that easy to see. It's also clearly audible if you know what to listen for. You can disconnect an injector temporarily and dab 12V across it briefly if you want to hear it. If you find it's being driven electrically but you don't hear it opening, perhaps the injectors have dried up and glued themselves shut. In that case they need servicing.

Another way to verify that the injectors are actually flowing fuel is to pressurise the fuel rail, confirm it's holding pressure and then briefly open an injector. If it's flowing, the pressure will drop sharply each time to do that.

If the injector is getting 12V but not being opened by the ECU, that might mean the ECU is missing a signal. You mention of the air flow signal suggests there's something like that going on.

I've checked the injectors so I know they have 12V constant when the ignition is in the ON position but not when cranking so I'll check that next! The other wire sat at 0.4V when cranking but that doesn't tell me a lot as the multi meter probably isn't very accurate

I'm assuming a noid light will show the pulses being sent to the injector and if that looks good I can assume the injectors are blocked?

The error code only appeared after we sprayed the starter fluid in, I didn't want to spray the fluid at the start on the intake because of how far it had to travel so we took the intake pipe off and sprayed it in after the airbox/ maf so it could just be that the ECU got confused that it was running whilst reading almost no airflow?

Is there any chance of damaging the injector by sending 12V through it?

Thanks for the help man!

[jrallye]

GreenV8S said:

Disconnecting the MAF would exactly explain the error code.

The noid lamp would light while the injector is driven electrically. At low low these pulses can be too short to see with a meter. The noid lamp will produce a flash which is easy to see as long as you are not somewhere bright. Alternatively you can use an oscilloscope.

If the injector is being driven electrically you have proved the electrics are OK and the ECU is trying to run the engine. That means you're looking for a mechanical problem with the injector, or it isn't getting fuel pressure.

Brilliant thanks mate!

We didn't disconnect the MAF just pulled the intake pipe off the manifold but I imagine even that would be enough to throw a code (it has so it must've been enough)

Might have a look at this tonight if I've got time if not I'll update whenever I get round to it

I just really hope the wiring's all intact/ working not warm to the idea of fixing wiring issues

Edited by jrallye on Wednesday 8th May 14:55

[jrallye]

Back again with another update!

Not really got any further although I have made a noid light from a brake light bulb, 2 pieces of wires and 2 hair clips

The injectors are getting a pulse so my guess is not enough fuel pressure or clogged injectors although we took one out and it looked like a clean injector

But before we delve too far into the injectors I bought a second subaru that runs (not a twin turbo) for parts and chucked the pump from that in the twin turbo, the pumps are a vastly different size but I'm hoping the pump from a 2.5L will at least get the 2L TT idling so we'll end up with a couple outcomes



And hopefully after that we'll be good and I can start doing some more exciting stuff to the car

[jrallye]

GreenV8S said:

I wouldn't recommend that - the brake light bulb will be drawing too much current and will also be very slow to respond. Do something similar with a 12V LED and you'll be fine.

It works fine once the bulb warms up after a full cycle or two and you could clearly see the pulse

Is there any worry with it drawing too much power or just that it won't work as well as an LED?

[jrallye]

Penelope Stopit said:

GreenV8S make a good point here

The brake light bulb (21 Watt) will draw close to 1.8 Amps

You need to know the injector resistance before using big bulbs to test

See this



Taken from here (a good read) http://fuelinjectorclinic.com/faqs/

I'll have a read through this and see if i can get my head round it all

So are we saying that the brake light bulb even though it pulses on and off might not mean that the injector wiring isn't faulty?

[jrallye]

tapkaJohnD said:

jrallye,

Your determined and logical diagnostic process is admirable! I take back any implied previous criticism!

As you already know, if you want to see injector pulses, a multimeter won't do it. You don't say how much you paid for this clunker, but you might consider buying a "Picoscope". This clever bit of kit will turn your laptop into an oscilloscope, well able to display signals into the microsecond range. The simplest in the 2000 range costs £120 and would do this job, a full Picoscope Automotive 2-channel £700, or a Hantek 6022 equivalent (Chinese knockoff) is half that price.

JOhn
PS Oh! You've made a signal test lamp! Well done, but see comments above. These scopes don't have that problem. J.

At the start I think the comment was warranted I had missed some pretty big steps out, I'm starting to map out my problems and try and figure stuff out in a way that makes sense on paper first before just going for it!

I'm trying to get this started on a low budget and the amount of times I plan on starting a non runner are fairly low so I'll see if I can borrow a better tool from my dad or brother

[jrallye]

GreenV8S said:

Both. Either a 'grain of wheat' bulb or LED would be safer and work better.

I'll have a gander and see what I can do! Thanks for the help!

[jrallye]

GreenV8S said:

If they're being driven electrically, they should be opening. That makes quite a loud tap which you can easily hear if the engine is stopped, but is quite hard to pick out from a running engine unless you are familiar with the sound.

If you are still feeling brave enough to experiment, you could try unplugging an injector from the loom and briefly touching a fused 12V supply across the injector. If the injector is mechanically working you will hear a loud tap as it opens when you apply power and closes when you remove it.

The last test needs a fuel pressure gauge to measure the fuel rail pressure. It should build up when the pump runs and more or less hold steady when the pump stops. If you briefly operate an injector now, the pressure should drop quite sharply. If so, it proves the injector is both opening and flowing fuel.

Brilliant mate, thanks again!

I'll try and get the injector to open myself and listen for a tap

I should probably buy a pressure tester regardless as they're not crazy expensive and it'll come in handy if any of my other cars start running poorly I can rule out the pump right away

[jrallye]

Have been messing with this a little more and we've got somewhere but I don't know how or why

I'm not sure if I said before but I put the pump and filter from a working 2.5 na car onto the twin turbo just after we burnt out the starter motor

I replaced the started motor and noticed there was a hose clamp missing from the smaller turbo when I went to re install the top mount inter cooler, so found one I had lying around and chucked it on and put it all back together and tried to start the car to make sure the starter was working and now the car coughs and splutters like it wants to start

I've opened 3 of the 4 injectors with a 9V battery and they all open nice audible click (can't get to one of them without moving ac pipes so thought i'd leave it for now)

I've retested the injectors with a lower watt bulb I had lying around still not an LED but it lights up much quicker and again 3 of the 4 pulse correctly (I haven't tested the forth again but I read these injectors shares a ground so if 3 of the 4 are getting a signal they all should be unless the wiring is physically broken)

The tank is pretty grim inside lots of rust and gunk so will be replacing it with the tank from the other Subaru I have I drained the tank completely and put it some fresh fuel 95ron not sure if they run very well on 95 or at all as they were designed to run on 97-99 over in japan could be wrong there again just what I've read/ been told

I've tried the old original pump and the second hand 2.5 na pump and it coughs in the same way

If I unplug the pump it goes back to cranking with no coughs or splutters so fuel is there but maybe not enough to start the car

I started it with cold start again and stop spraying once the car gets going and nothing so I'm leaning towards the fuel pump at this point

Here's a video of the old girl trying her best Legacy Twin Turbo No Start

I personally think I just need to grab a new pump at this point and stop messing around with second hand units

I know the injectors are working to some extent they get the correct signal and the open and close as they should

The car runs on cold start so it has spark and all that but it sound to me like it's not getting enough fuel?

Any more ideas or anyone who can hear why it's not starting will be greatly appreciated, we're close now!

[jrallye]

I've gone ahead and ordered a pressure tester!

It arrives Wednesday so I'll update then

I didn't really want to buy one but I'm sure it'll come in handy given the average age of my cars

I'll test the fuel pressure first and see what's going on there, and if it's low I'll look at the pressure regulator before going straight for a new pump

Thanks for all the help again I don't think I'd have got this far doing research alone!

[jrallye]

Almost a month late but the pressure tester came and the fittings leak so I bodged it together with some silicone and got a reading of 43-45psi when primed so there's pressure there

I still need to test the pressure when cranking but as a start this isn't horrendous there's pressure it'll very slowly loose pressure when the pump switches itself off after priming but that could always be the gauge leaking and not the fuel system

Would it be worth trying to force the fuel pump always on and see if it'll start that way? Obviously not what I plan on doing in the long run but would definitely rule out the fuel issue for testing?