The MGA With An Attitude
Bleeding the OIL PRESSURE GAUGE signal line -- DT-104

At 09:59 AM 10/10/04 -0600, John Keller wrote:
>"I am about to replace my oil/temp gauge. .... I figure some oil will come out of the oil line when I disconnect the line. Is this a problem?"

Only if dribbles on the carpet.

>"If I connect the oil line with a little bit of air in it, how will the air be displaced?

I have (until now) never worried about air in the line, and have never bothered to bleed it. It seems that the gauge works fine with air in the line. Oil will simply flow up the line to compress whatever air is there to the same pressure as the oil, and the gauge works anyway.

If you were real particular about it, you could bleed the air out of the line. Disconnect the tube fitting at the back of the gauge. Be prepared for dribble, so have a small container handy to catch the oil. Start the engine and run it until oil comes out of the tube. Stop engine and reconnect the fitting.

>"Might be as simple as the gauge isnít air tight,

Oh, the gauge is definitely air tight. It has to keep the oil from leaking out at up to 80 psi pressure.

>"but I donít want to screw up.

Not to worry. It's just about idiot proof. There is supposed to be a tiny leather washer for a seal between the end of the tube and the nose of the fitting on the back of the gauge. Fiber or nylon would likely do as well. It has a part number, and you can buy it from Moss. I almost never bother to mention it, because it's probably not necessary to install it. For all the times I have R&R that gauge, I have never even looked at the tube connection to see if the seal washer was there, and of course it isn't. To the best of my knowledge it has never leaked in 340,000 miles of driving. I suppose the flare nut fitting makes a pressure tight seal with metal to metal contact, pretty much the same as a brake hydraulic line does. Just that it's 180 degree flare here (flat face) rather than a 90 degree cone flare. But if you every have one that leaks, the prescribed seal washer would definitely fix it.

I seem to recall that the brass connector fitting on the frame (between the flex hose and the steel line) has a very small restrictor orifice in the center, intended to slow the flow of oil to dampen out any rapid fluctuations of the gauge needle. If the line was filled with non-compressible oil and no air, and there was no restriction in the line, then small rapid pulsations from the oil pump might make the needle vibrate. Having air in the line would actually serve as a shock absorber. Occasionally we get a report of an oil pressure relief valve acting up, and the pressure reading may fluctuate by about 10 psi at about one second cycle time. That doesn't have anything to do with the gauge, but some air in the signal line might reduce the amplitude of the fluctuation. Without air in the line it might well indicate a truer fluctuation of double that amount.

The question then is whether the volume of air being compressed slows the reaction of the gauge, due to the time it takes oil to flow in and out of the pipe with pressure changes. You may or may not notice this when the engine is running, as the engine speed and oil pressure don't change all that fast. But it can have something to do with how long it takes the gauge to come up to full pressure when you start the engine (a full 50 psi sweep). Dammit. Wait a minute there.
Ah-ha! Now you just caused me to go spend 15 minutes in the garage to run the test. I started the engine and ran it for a minute at fast idle to stabilize the oil pressure. Then shut it off and waited for the pressure to drop to zero. I was a bit surprised at how long it took to drop. It went from 50 to 10 in 3 or 4 seconds, then took another 5 or 6 seconds to get to zero. I'm sure the engine itself doesn't hold the pressure that long. The slowness of the gauge motion could be related to cold 20W50 oil in the small signal line.

I started it again (running at fast idle where I knew the pressure would be at least 50 psi) and watched to see how fast the pressure indication would rise. The needle came off zero almost immediately when it started, meaning there is no pressure lag at the engine, but it took about 5 seconds to get up to 50 psi. I repeated the test a couple more times just to be sure it was consistent.

Then I disconnected the line from the gauge (not a drop of oil on the fitting), held a little tub under the line, started the engine, and waited for the oil to come out. Again I was a little surprised, as it took about 10 seconds before anything at all came out of the pipe. Then it blew a few bubbles, and then it began to dribble very slowly. I continued to run it for at least a full minute, and I'm sure less than a half teaspoon of oil came out in that time. Then I reconnected the line and did the test again.

After bleeding the air out of the line, the pressure reading would go from 0 to 50 in about 2 seconds after start up (noticeably quicker). When shut down, the pressure went from 50 to 10 in about 2 seconds, and then down to 0 in another 4 seconds or so. Repeating the test produced consistent results.

It was an interesting experiment. After seeing how slow the oil flows through the line I'm not surprised at the difference in reaction time of the gauge with and without air in the line. I can now heartily recommend that folks should bleed the air out of the oil gauge signal line. Maybe sometimes the slow rise of pressure indication after startup (due to air in the signal line) may be mistaken for oil having drained down out of the oil filter while the engine was parked. Anyone who has that problem might try bleeding the gauge line before they jump to conclusions about the oil filter.

Thank you for your comments -- Send e-mail to <Barney Gaylord>
© 2004 Barney Gaylord -- Copyright and reprint information