'84 GTI w/CIS and Starting Problems

[turbobaja]

I've been daily driving my '84 GTI for over 12 years and it's slowly developed issues starting both hot and cold, very much like a vapor lock problem, but possibly more?

The car had sat mostly for a couple years before I bought it, water leaks, electrical issues, broken engine mounts, blown strut, etc. But from the original owner at least and totally unmodified, all original w/92K on the clock (that didn't work...). Battery and fuel and it ran and drove home. I gave the entire rig a run through and got it fully functional and back to a healthy condition. It ran out of power one day and I found the 3 week old fuel filter was plugged with debris, rust and crud. Over the next few months I replace the fuel filter 2 more times due to crud plugging them up and restricting flow. I replaced the fuel pump then as well and over the next 8 or 9 years of daily use I had virtually zero issues with it. New filter every 2-3 years as preventative maintenance only.

Slowly, It started to need 2 or 3 tries to start it when very cold out. It would fire right up initially, then starve out, like it ran out of fuel. But it would always start, and once running would still run great. Eventually, It started to have hot starting issues too. After a brief hot soak, it would have an extended crank, then briefly fire, only to stall out. Then another extended crank, start and stall. When it finally starts after 3 - 5 tries, it stumbles to life and acts like it's running on 2 cylinders for a while. I've tried cranking with no throttle input, and with full throttle, as if trying to clear a flood. No matter what, on a hot soak start it really struggles to fire on all cylinders, even after some light free revving up to 2,000 rpm. It always ends up clearing out and running great again.

I replaced the fuel pump and filter again, new cap/rotor/plugs/wires, and clean air filter over the last year with no affect on my starting issues. I also inspected the fuel tank through the sending unit hole only to find a sparkling clean tank with zero signs of rust. I've pulled the injectors out and checked for leaky tips and all were dry with the pump running (injectors were replaced back when I first got it on the road). I'd always wanted to know what the actual fuel pressure was like, so I built a little DIY gauge and valve setup and did some testing. I found my cold regulated pressure on a 32F day was about 35psi. The warmed up regulated pressure was almost 68psi and dead headed it was over 85psi. Every time, the system pressure bleeds down to zero within minutes after shut down. I checked the check valve on the fuel pump outlet (new with recent pump) and it tested good. The pressures all seemed a bit high, so I bought a different brand gauge (both 0-100psi) and got virtually the exact same readings. So I pulled the system pressure regulator out, found one thick and one thin washer. Pulling the thick washer dropped all the pressures 3-4 psi, but system pressure still bleeds off within minutes and hot soak starts were still a huge struggle, even with system pressure reading almost 60 psi the instant the engine starts cranking. So I finally checked the fuel accumulator and found it was leaking internally and it got replaced. I have not checked system pressure since replacing the accumulator, but I have to assume it's still bleeding off because it struggles even worse to start, both hot and cold now. I still daily drive this thing and I didn't like having the gauge plumbed in 24/7 so I pulled it out, it could go back in in minutes. I feel like it must be running lean ( evident when I checked the plugs recently also) which I'm sure isn't great for this poor old engine. As I understand it, high pressure means lean fuel trims with these units.

I've heard of issues with the hall sensor heat soaking and causing hot start problems and have NOT verified yet if I've got a spark problem affecting things or not. I've clearly got a fuel pressure problem, and I'm mostly looking for any advice or experience with these types of pressure control problems and starting problems. I'd like to pull the injectors back out as well as the cold start injector and recheck for leaks. But I suspect the loss of residual pressure and excessive running pressure is due to something fouled up inside the main fuel metering unit, possibly a restriction on the return side?? Is there any reason I shouldn't take the fuel control unit apart and see whats inside? Should just replace the system pressure regulator? Should I be checking/replacing the O2 sensor? Any input or advice would be greatly appreciated.

[Marc]

Give this a shot before buying any (more) parts... https://youtu.be/yBwVfpPqSRM?t=142

[turbobaja]

Give this a shot before buying any (more) parts... https://youtu.be/yBwVfpPqSRM?t=142

I've had that unit off and apart, but I'll do it again for good measure. Do you think crud in the WUR could be causing system pressure to bleed off as well?

[Marc]

...Do you think crud in the WUR could be causing system pressure to bleed off as well?

I don't think so, but the checkvalve sure could.

[falcor]

nevermind, missed that in your post.....

[turbobaja]

...Do you think crud in the WUR could be causing system pressure to bleed off as well?

I don't think so, but the checkvalve sure could.

I'm curious what component is responsible for keeping fuel from just flowing down the return line back to the tank once the pump shuts off. Is it the seat of the System Pressure Regulator, or ??. I've checked the check valve on the fuel pump twice and it seems to seat and seal correctly to not let fuel flow back into the pump. I wish I could safely pinch off these fuel lines in a few places to isolate the problem better, but I'm afraid of damaging the lines by crimping it enough to seal off flow completely. I've had a few of the banjo fittings off the metering unit and they've all had signs of rusty sediment collecting in the lower part of the banjo, as if water was in the fuel at some time. I've added half a cup of Marvel Mystery Oil to the tank, as well as a can of SeaFoam, neither seems to improve the situation.

I'll pull the WUR off and have a closer look at it, probably pull the system pressure regulator back out and see what I can see down it's bore as well. Then reinstall the pressure gauge assembly to see if anything has changed.

[Piledriver]

The control pressure from the WUR is the driver of everything.
It is also the weakest link in the system.

From memory, CIS+Lambda control pressure should be about 25-30 PSI cold and no more than 55 hot.
Don't mess with anything else unless control pressure is in spec or a tiny bit lowered.

WURs are reasonably easy to make "adjustable": small adjustments make a big difference.

Another common issue is the lines tend to seep at the injectors end, the fuel evaporates but any air in the system and pressure loss effects startups hot or cold. Cutting the lines back a 1/4" and recrimping solves this for another 30 years.

I ran a later model meter (CIS-E) with no WUR for ages on my 914, easy electrical remote fuel mix control.. The DPR can be adjusted via a screw on the back so that its baseline mixture is ~ideal, and it will run well with no external help/purely mechanical system other than a cold start squirt. AFAICT the CIS-E and CIS-M DPRs only differed in color and the baseline adjustment of that screw.

[Jadewombat]

As Pile alluded to, get a set of CIS fuel pressure gauges. It's well worth the $90 (at least when I bought mine ~14 years ago) and takes a lot of the guesswork out.

Pull the cold start injector and put it in a cup and watch it. Make sure it's working correctly. Probably not a huge factor at all, but if it's leaking, etc.

Follow the Bentley manual trouble shooting steps and graphs. You'll find the culprit.

[Piledriver]

I cheated on my "gauges" and just modified a couple spare banjo bolts for 10/32 Clippard adapter fittings to std gauges.
pop in the original bolts or cap screws in with thread sealant when done.

The computerless CIS-E worked better than any other CIS system I ever messed with, ditching the WUE was a huge win.

There's even a Megasquirt 2 Extra code mod floating around that can directly control a CIS+Lambda or CIS-E or -M setup for full programmability.

[ps2375]

I found it much easier, more powerful and economical after I converted my CIS car to MS. I never had a hot start issue again. And with a digi fuel rail and later a digi dist, it was an easy conversion to fully programmable fuel and spark. CIS is a good and reliable system, but when WUR's get to the cost that they are, if you can even find them now, other options become more feasible and realistic.

And one of the main reasons I made the switch was the issues with hot start and at that time, WUR's were only available as "rebuilt" and cost almost as much as the MS unit. And one of the benefits of the MS over the CIS is I no longer needed to worry about cam overlap and still be able to achieve a decent idle.

[66brm]

How many miles on the car now? Have you done compression and leak down tests? The reason I say this is it could be something mechanical rather than an issue with the fuel system.....?

[turbobaja]

How many miles on the car now? Have you done compression and leak down tests? The reason I say this is it could be something mechanical rather than an issue with the fuel system.....?

It's got about 160K miles on it now. Haven't done a compression/leak down test in years. Not a bad idea to check it out while I'm chasing the fuel problems. I did check the valve clearance when I resealed the valve cover last month and they were still within spec for cold lash. I've always run synthetic 5W-40 oil, never had a temp problem and keep a close eye on it in general. It's been a great little car.

[turbobaja]

I found it much easier, more powerful and economical after I converted my CIS car to MS. I never had a hot start issue again. And with a digi fuel rail and later a digi dist, it was an easy conversion to fully programmable fuel and spark. CIS is a good and reliable system, but when WUR's get to the cost that they are, if you can even find them now, other options become more feasible and realistic.

And one of the main reasons I made the switch was the issues with hot start and at that time, WUR's were only available as "rebuilt" and cost almost as much as the MS unit. And one of the benefits of the MS over the CIS is I no longer needed to worry about cam overlap and still be able to achieve a decent idle.

This makes a lot of sense, I love what MS has done for my wife's bug and my turbo Baja. The CIS has worked so well for so many years, and it's simple, and already there , I've got too many irons in the fire to do an MS swap on this thing right now. I always had a feeling this fuel system was going to give me trouble eventually.

[turbobaja]

The control pressure from the WUR is the driver of everything.
It is also the weakest link in the system.

From memory, CIS+Lambda control pressure should be about 25-30 PSI cold and no more than 55 hot.
Don't mess with anything else unless control pressure is in spec or a tiny bit lowered.

WURs are reasonably easy to make "adjustable": small adjustments make a big difference.

I'm gonna focus on this for now and see if I can get the system pressures to make sense.

Thank you everyone for all the input and suggestions!

[Piledriver]

One other little tidbit that may or many not be well known:
CIS injectors are "dumb" and unless ~totally plugged up will flow whatever the meter sends them.
You can switch them around and see that, if a lower reading follows an injector, its unusual, and junk.
They do have little removable filter screens in them, as do the meters.

A 4 hole CIS-E or -M meter will flow every bit of flow a Bosch 044 pump can throw at it pulled wide open.
That's ~like 600HP of fuel.

If you have a flow mismatch, many meters have some adjustment for individual injector flow, but if seriously off you are probably better off finding another meter, as its a very fussy adjustment and may make things worse.

Note this applies to all versions of CIS, so early injectors work fine on late meters as well as the reverse for the most part.(The breaking pressure is somewhat higher on later oring sealed injectors, but usually lost in the noise/overcome via CO screw adjustment)
The later injectors have finer threads so this implies the hose gets swapped too: The meter end never changed.

I picked up a couple sets of the inexpensive brass "mercedes" injectors from Autohaus AZ and they worked peachy.