At 11:08 PM 1/20/02 -0800, Malcolm Cox wrote:
>"On rebuilding my gearbox .... I noticed that the detent depth in the gear selector rods do not conform to the latest (lemme see, that's about 40 years ago!) factory recommendations. The rods have clearly been replaced at some time and show little wear, so they predate 40 years ago."

Got any use for some more rods? I have a few sets of them left over, as they never seem to wear out. The 40 year old parts usually do look nearly new.

>"I have been vintage racing the same gearbox for about 5 years ( perhaps 6,000 track miles) and have never had it "automatically" jump out of gear, except due to occasional driver incompetence. So before I spend hard loot on new selector rods, or ponder myself into a hole in the ground, I wondered what has been your opinion on the deep detent controversy!"

I have personally put nearly 200,000 miles on MGA 1500s, with lots of autocross time and a little open track time, and never had one jump out of 3rd gear. Then 18 months ago [that would be June 2001] I had the good fortune to discover the circumstances under which the thing would pop out of 3rd gear. I had run my MGA for several hot laps around Blackhawk Farms track with no problem. Then I switched places with a friend so he could get some seat time. The first time he came bearing down on the last turn it popped out of 3rd gear, even though it had never happened to me. And it continued to do so consistently for him.

It was closely related to his driving technique (not so great). As he was approaching a turn at high engine speed in 3rd gear (~6000 rpm) he would depress the clutch and release the throttle while applying the brakes. With the engine vibration and zero torque on the gearbox internal parts, the 3/4 shift rod would vibrate out of position and end up in neutral as the engine was slowing down. If you just took your foot off the throttle without depressing the clutch it wouldn't do that, as the reverse torque in the gearbox would keep enough friction on the 3-4 sliding hub so it wouldn't slide. Now that I know when it will do it and when it won't I have quit worrying about it and have never bothered to do the suggested modification to the detent. It has still never popped out of gear for me.

If you want to do this modification, you do not need to buy a new shift rod, just grind the detent notch a little deeper in the original one. Complete instructions for the modification are in the MG Series MGA Workshop Manual, Section FF.

Addendum January 2009:

There are two separate problems at play here. One "fix" is the update to later design specification that is well documented in Section FF of the Workshop Manual. The other possible issue is a factory error that crept in for short time during mid 1500 production when incorrect springs (too short) were installed for the shift detents. This problem and the fix for it is detailed in Confidential Service Memorandum MG/241 dated 18 February 1959. The correct longer spring is and always was factory part number 2K4909, 1-3/16-inches (30.16 mm) long. The incorrect spring that was installed in some units is factory part number 22A75, 31/32-inch (24.61 mm) long.

The kicker to this is that in more recent years there has been widespread cases of replacement part vendors (maybe most of the current vendors) supplying the incorrect shorter spring for this application. If you buy new springs in the interest of "fixing" this problem, you might inadvertently end up removing the correct long springs and installing the incorrect short springs, thereby creating the problem that did not previously exist. Any time you have these springs out of the gearbox for service you should measure the free length (and compressed length force if possible) and service the unit accordingly. There is a page FT-038 in the Faulty Parts section detailing the problems with current vendors supplying the wrong springs.

If you do not have and cannot get the correct longer springs for the detents, there is a close fix using the shorter springs. You can add shims to jack up the height of the short spring to equal the specified final load of the correct spring at the specified load height. That is, 18 to 20-pounds at 3/4-inch test height. Since the spring wire diameters and/or spring rates are not specified, this requires measurement of the spring height under a known load in order to determine the compressed length and amount of shims needed. It is the same procedure used to shim valve springs to a desired force and specified working height. Once you determine what shim height is needed, you can use a 5/16-inch diameter metal rod (or straight shank of a 5/16-inch bolt) cut to the length needed for boosting the spring to correct height. Drop that shim into the hole before installing the shorter spring.

A word of caution here. When fully depressed the spring may go to solid height at a "too tall" height for adding the amount of shims you thing are required. You need to measure the solid height of the spring with a detent ball on top, and also measure the depth of the spring pocket hole below the shift rod to know how much space is available for stacking shims. If you install too much shim you will not be able to depress the detent ball below the shift rod, and you will not be able to install the shift rod. If there is not enough space to add the needed shims, then the only fix is to find a longer spring.

You will also find that the various spring pocket holes are different depths below the shift rods. This is how the factory applies an intended different pressure to each of the three detents without having to use three different springs.