The rear seal installation is in 5 steps spanning 5 web pages.
Click green arrow at bottom to follow pages in sequence.
    A - Modify rear plate
    B - Machine slip ring
    C - Modify Viton seal
    D - Modify bearing cap
    E - Assembly
    F - Can I buy it?
    G - Video Installation
    H - Follow-Up Notes - (this page)

As much as I try to discourage people from doing this on a routine basis, substantial interest persists and the questions keep rising. So by popular demand, here is more information.

Gary Edwards in Kernersville, North Carolina, USA wrote:
"I have been thinking about polishing the ring. Barney says we need a RMS 16. At one time I found a chart that said you need a 240 grit surface to get to 16RMS, of course I could be mistaken. Now if you polish the ring with 1500 git paper and a buffing wheel is it too slick to allow the seal to work"?

I will take a shot at this in layman's terms. Since most of us know what a sine wave looks like, try this.
You might immediately recognize the 0.707 value as the sine (or cosine) of 45 degrees. RMS means "Root Mean Squared" which is a kind of mathematical average of the value of the variable equation, or in this case average height of the line on the graph (sort of). We know a rough surface is not a perfect sine wave, but this will serve as a decent illustration. In the graph, if V=1, then RMS=0.707V. The reciprocal is V=1.414RMS. Peak to peak on the sine wave is then 2.828RMS.

In the expression "16 RMS", the number is in micro inches (0.000016 inches). When RMS is 16, peak to peak on the sine wave would be 2.828x16=181 millionths of an inch. That would be 0.000181 inches (slightly less than 0.0002 inch peak to peak on the surface).

For a common reference, a surface carefully machined on a milling machine or lathe with slow feed and a sharp cutter can be 125RMS. For smoother surfaces you pretty much need a grinding process to get there, like grinding a crankshaft or camshaft for your car engine. 16RMS is about eight times smoother (or 1/8th as rough) as the finely cut machine surface. You definitely do not get there with 240 grit sand paper. Not even with 600 grit emery paper and oil. Well, maybe with 600 grit emery and oil with a light touch and high speed. However, crocus cloth and oil might make it too smooth. Yes, a surface can be too smooth for the rubber seal to work its best.

Surface roughness needs to be at least down to 16-RMS to have minimal wear on the rubber seal lip. Irregularities in the surface also hold oil which is good for lubricating the seal. If surface roughness (or smoothness) may be below about 4RMS it may not hold oil well enough for good lubrication of the seal. Think if this as a squeegee that wipes the surface perfectly clean so the rubber seal runs in direct contact with the metal with no lubrication. In that case friction may generate enough heat to destroy the seal. That would be similar to running your engine with no oil and melting the crankshaft bearings.

In reality, what you want is 16RMS maximum and 4RMS minimum surface roughness. In the engineering and machining world, a grinding process that produces the desired 16RMS (max) smoothness is very unlikely to accidentally make the surface smoother than 4RMS. Therefore it is not necessary to specify the minimum number on the specification drawing.

The primary problem with this device seems to be the care required to get it sealed around the edges without getting RTV sealant on the back of the lip seal. That is, the thing is not idiot proof. You cannot just throw the parts together recklessly and expect it to work every time. It requires some level of care (or finesse) during assembly. Jim Juhas' idea to use a card stock paper ring to guide initial application of RTV around the bolt heads was very good, and likely to improve the odds of success on first try considerably. YOU MUST NOT HAVE RTV SEALANT TOUCHING THE BACK OF THE SEAL LIP. The "School Of Hard Knocks" is a wonderful educational institution. Anyone who has to pull the engine out again will almost certainly get it right on the second try, but of course better to do it right the first time if you can.

I first did this in the mid 1990's as an emergency repair between weekends. I had just spent considerable money to repair a damaged engine (aftermath of a broken crankshaft). The engine was perfectly serviceable aside from a dramatic oil leak at the rear, and I didn't want to disassemble it again. The design was done to allow installation of the seal with minimal machine work, no modification to the crankshaft or flywheel or engine block or rear bearing cap (except for the two drilled drain holes as an after thought).

Most of the mods could be done by hand by a DIY mechanic in home garage. I had the rear plate center hole enlarged using a fly cutter in a milling machine (because it was handy). That might otherwise be done in a drill press if the spindle bearings were tight enough. That large hole could even be enlarged with a hand grinder if you had enough time and patience, or a grinding wheel in a hand held router. For the four #8 screws, I laid out the locations, center punched, drilled the holes with a hand drill, and tapped the threads by hand. I also cut the shell off of the standard seal with a hack saw, cleaned up the burrs with a grinding wheel, and drilled the four screw holes by hand. The only precision machining done at the time was for the slip ring.

If the seal wears out it can be replaced with another hand modified seal If the slip ring wears it may be repaired with a Speedy-Sleeve. If the crankshaft needs to be replaced, the slip ring might be removed and transferred to the next crankshaft (maybe). There is no welding or precision machining to be done on the crankshaft or engine block.

As this was originally a quick jury rigged patch job, I have been thinking for a long time that sometime in the ensuing 20+ years someone could have come up with a better idea (but apparently not). I haven't been thinking much about improving it, because I never needed to do it again. About the only part of this mod that cannot be done by hand at home is machining of the slip ring. I have some thoughts that this might be done using a 3-7/8" speedy sleeve in place of the machined slip ring, and a similar mod to a 3-7/8 shaft seal. The engineer in my thinks that may be somewhat fragile, but it might work if you could do it without damaging the Speedy-Sleeve during assembly.

On Jan 15, 2017, Jeff Sienkiewicz in New Milford, CT, USA wrote:
"Tom Young and I cut some slip rings, he on a lathe, me on a metal chop saw. The chop saw version required a few hours of hand work (maybe had I been more careful, less), but the moral of the story is that even the slip ring can be made from tube stock by a hack (me, not Tom) at home with limited tools".

When I first did this I had little idea that anyone else might ever do it again. Apparently there have been dozens of these installed by now (including one on the Wheeler Dealers TV show a year ago). People keep asking about this like it should maybe a standard retrofit to every MGA and early MGB, and perhaps every Austin 3-main bearing engine for all applications (as opposed to an occasional emergency fix). These engines were designed to drip some oil, just the nature of the cars of that era. I don't want to stop all oil leaks any more than I want to install Independent Rear Suspension. Stopping oil leaks is a luxury option, like air conditioning or a 5-speed gearbox, and none of it is cost effective. But apparently some people don't mind paying a lot of money for some minor improvements.

This can be a useful gadget if you have a gusher oil leak in your engine considering the standard alternative is the expensive process of line boring the engine block to restore original dimensions for the scroll seal (and maybe replacing the crankshaft at the same time). It might be nice if someone would consider making the parts up as a kit with slip ring and modified seal, and machined rear plates on a exchange basis. There seems to be enough demand to make up perhaps 50 rings and seals at once in which case the cost would be somewhat reduced and the convenience factor greatly enhanced. I have no intention of doing that, not my chosen profession, but I would have no objection if someone else wanted to do it.