Best way to find vacuum leaks? argon? oxygen? propane? ether?

[r_p_ryan]

Updated:Full write-up on DIY: http://www.peachparts.com/Wikka/M103VacuumLeak

Original posting
I'm trying to survey the various ways to find vacuum leaks in the intake. I've heard of everything from brake cleaner (interrupts combustion and causes idle speed to go down) to ether (increases combustion and should cause idle to go up). The problem with these approaches is that most of our cars have a computer to keep the idle at a set range. So adding something to speed up the idle would be countered by the computer via the idle air valve. What about argon? For those not familiar with argon, it is used in welding and other processes because it is totally inert and should theoretically interrupt combustion.

And while we're on the subject of finding vacuum leaks, what should we do to the engine to make it easier to identify? I could imagine disabling the computer by removing the OVP so it doesn't try to adjust the idle speed.

Thanks,
Robert

[Kestas]

I remember seeing something on "Shadetree Mechanic" where they introduced a product that injects smoke into the vacuum system (engine off). With the right lighting you can see the smoke seep out at the leak points.

The most popular way is pretty much the way you descibed, by spraying solvent at suspect areas. In addition to brake cleaner and ether, you can use starter fluid or flow propane over suspect areas.

I've always used a small hose. By putting it in my ear and then moving the other end around possible leak areas you can easily here the vacuum leak.

I watched my Dad do this in the fifties when I was 5 or 6 years old.

Good luck,

[r_p_ryan]

Listening makes alot of sense - I hadn't thought of that. I have a stethoscope and I'll be sure to give it a shot next time.

Thanks,
Robert

[stevebfl]

Smoke testing is probably the professional standard. Stethescopes work fine, ultrasonic ear is cool. Carb or brake clean can be a quick way and more sensitive if watching the O2 sensor or exhaust gas.

[Roncallo]

Originally Posted by r_p_ryan

And while we're on the subject of finding vacuum leaks, what should we do to the engine to make it easier to identify? I could imagine disabling the computer by removing the OVP so it doesn't try to adjust the idle speed.

Thanks,
Robert

I cant say for sure but I dont think the engine controls will react fast enough to a quick spray. When you find the leak it drops the engine speed pretty abuuptly. Last time I did this was on engines with carburators.

J. Roncallo

[Lee Roy]

There is a very simple way to check for intake leaks. Take a spray bottle with water in it and spray around the enginge compartment. If there is a leak the idle speed will change.
The water will not damage anything. I have been using this
technique for 25 years, works like a charm.

[r_p_ryan]

The stethoscope and water sound like great ideas. I was thinking about how spraying a fluid would effect the idle. I wouldn't be surprised if when the fluid finds the vacuum leak it temporarily "seals' the leak, thus briefly returning the engine to it's normal idle. Water and a stethoscope sound pretty safe too.

Smoke and ultrasonics sound cool, but out the budget range of this amateur

Thanks guys for all the great tips. When I'm done with my ball joints I'm going to inspect both my E's for vacuum leaks. I'll first try the stethoscope and then the water. Once I've found any leaks I'll try an assortment of chemicals/gasses to see if there's any difference from water. I also will tap into my O2 sensor to see if an amateur like me can notice any difference. I do know that my O2 sensor (and my duty cycles on X11) are all across the board. Would it help to disable the computer or plug any of the ancillary vacuum connections? Has anybody found any readily-available plugs to cap-off the vacuum lines while diagnosing?

Thanks!
Robert

[r_p_ryan]

Steve you are the Master!

Full write-up:
http://www.peachparts.com/Wikka/M103VacuumLeak

I tried everything mentioned on the thread and the only thing that really worked (and it worked very well) was to monitor the O2 sensor and use some carb cleaner. It was amazing how easy it was. If I sprayed it for just an instant at the leak the O2 sensor responded. If I sprayed too much in any direction the vapor would get sucked into the throttle body and alter the sensor reading. The trick was to shoot very small amounts and be a little patient. I started by unplugging the EHA and ICV, removing all the vacuum hoses and taping the nipples, and unplugging the throttle switch. Then I took a 3mm Allen wrench and changed the mixture so the O2 reading was about .1 volts. A shot of carb cleaner at the ICV and the O2 would hop up to .8 for a couple seconds and then return to .1. You couldn't hear any change in idle speed. Cracked bypass hose on the head side.

Also found the injector seal on #5 leaking.

Argon, propane, oxygen and water didn't do much at all. the stethoscope might work well for a professional but I just couldn't discern all the noises coming through it. After I found the leak I tried all these methods to see if there's difference and I couldn't find one.

After I repaired the vacuum leaks (temporarily until the parts come) I reset the mixture to be rich at .8 volts. Then I sequentially reconnected each of the vacuum hoses while the engine was running. When I removed the tape the sensor would drop to .1 volts, then I'd replace the hose and watch the sensor. Each time it went right back to .8. This gave me confidence that there were no leaks in the vacuum system or brake booster.

I reset all the fault codes and then set the mixture control using a duty cycle meter.

The car starts up much much smoother. When it was warm it could be a bit cranky at times. Now it fires up instantly.

Thanks for all the tips.

Robert

[danwatt]

I usually just use starter fluid. If the computer is able to immediately compensate for a change in the idle speed, try spraying some in the air intake. It will still rev up.

[emmydotnet]

Robert:
excellent vacuum diagnosis procedure. I have been searching for something like this. thx.

by the way, what was the problem on your car that you were trying to solve?
was it rough idle or hard cold start?


thx,
emmy
91 560SEL

[r_p_ryan]

the problems were difficult start when warm/cold and a slightly rough idle. Starting on cold mornings was easier than starting on warm afternoon.

[LeaUK]

An excellent post regarding the detection of leaks - thanks.

But I'm wondering if the Lambda control signal can also be used, or is it's reaction too slow? This signal is great for identifying a leak in the first place - in my very limited experience.

Of course the O2 sensor will react in around 100ms which is pretty much real-time but I note on my M119 (LH injection) that a leak take 10 seconds to finally reach stabilization.

Any thoughts about using the Lambda control signal?

Lea

[emmydotnet]

Robert:
so whether cold mornings or warm afternoon, was your hard start problem only when the car has been sitting (parked) for a few hours or so? just wanted to understand whether your problem was hard start on a "cold" engine or on a "warm/hot" engine.....

thx again,
emmy

[r_p_ryan]

Basically if the engine was stone cold on a morning where the outside temp was below about 60 deg.f the car start fairly easily (the cold start valve was kicking in).

On those rare weekend days when I'd start the car in the afternoon and it was 75 deg. f and the car had been in the sun then it would take a bit to get it started as well.

If the engine had been driven and then sat for an hour it would take a few seconds of cranking and giving it gas. It would typically catch and then instantaneously quit.

Now it fires up with no gas pedal and 2 seconds of cranking.