Pelican Parts Forums

Pelican Parts Forums (
-   911 Engine Rebuilding Forum (
-   -   Ultimate Oil Pressure Relief Valve Thread (

304065 10-22-2012 11:39 AM

Ultimate Oil Pressure Relief Valve Thread
Many of us have had high, or low, oil pressure issues, so I thought I would combine all the information I could find into a single thread for reference.

First, a big thank-you to Walt Fricke, whose pioneering work in this area is the source of much of what appears below.

First, the factory diagram. Copyright Porsche KG.

Second, a note about the oil bypass modification. The factory started doing this in 1976, it had been discovered much earlier, supposedly by the Kremer Brothers, and was adopted by the factory. This entails drilling a diagonal hole between the oil pressure RELIEF passage and the inlet passage to the pressure side of the pump, then plugging the original outlet of the oil pressure RELIEF passage into the case. This way, when the 76.8 psi limit of the oil pressure RELIEF valve is achieved, excess oil pressure is routed back into the inlet to the pressure side of the pump rather than going into the bottom of the case.

As everyone knows by now, this modification reduced the oil level in the case, reducing losses due to windage, and allowing the scavenge section of the pump to be reduced in size, allowing for a larger pressure section in the same overall package.

Here is the original service bulletin describing this change.

Here is a better diagram from Bruce Anderson.

Here is a photo of the mod performed by Craig Garr:

Here is how the factory did it, this is my own photo of a 964 case:

NOTE: Both bulletins above do not say ANYTHING about changing the springs in connection with the oil bypass mod. Of course the piston is changed-- the old-style piston has holes through the top, and if used in an engine with the oil bypass mod, would bleed air from the engine case back into the pressure pump inlet, and the engine wouldn't make oil pressure.

These are the old and new pistons compared:

The system diagram changed accordingly. (Note that is is a later diagram, showing the external oil cooler and relocation of the pressure transducer to the pulley end of the engine.)

The HORIZONTAL spring is the oil SAFETY valve.

It rides in a bore like this:

The VERTICAL spring is the oil RELIEF valve. See the endoscopic photos below.

Here are some actual photos of the oil bypass passage, these are mine, taken with a USB endoscope. You can see the side passage that is uncovered by the piston when the pressure gets too high. More on that in a second. At the far end (top) of the passage, this is where the oil pressure sender is located on an early car, and the switch on a later one.

Now, well below the piston there is another drilling into the case, this allows oil that gets below the piston to drain into the case, and also allows the air behind the piston to vent, so the piston can move. If this were not present, the piston would probably hydrolock.

304065 10-22-2012 11:42 AM

Now, turning to the piston itself, these are the dimensions:

The bore threads are M18 x 1.5, same as an oxygen sensor.

Now, look carefully at the top of the piston and the bore from the endoscopic photos. See how the piston has a chamfer? In hydraulics, this is called a "pop-action" design-- you can see that a much smaller surface area is exposed to oil pressure when the piston is seated. As soon as the pressure on that smaller area becomes sufficient to lift the piston off its seat, then the full piston area is exposed to oil pressure, which "pops" the valve open immediately. This is a desirable property of pressure relief valves and the factory was smart to include that feature here.

304065 10-22-2012 11:51 AM

Now some comments about the springs.

The original springs were both 70mm long.

From the workshop manual:

Free length 70mm
Diameter of wire 1.4mm
External diameter of spring 12.2mm
Pressure exerted at length of 52mm 76.66 ft/lb
Pressure exerted at length of 46mm 101.98 ft/lb
Length fully compressed 33.3mm

Those are the specs for the original 70mm spring. Here's an illustration of the change.

In 1978, some SC 3,0 motors were having oil pressure issues, so the factory came out with a service bulletin, increasing the length of the PRESSURE spring to 86mm, including a spring guide inside the lower end so the spring didn't buckle, and changing the depth of the plug to 9mm.

Notice that this bulletin about the springs came out much later than the oil bypass mod-- there is some disinformation out there that the springs must go together with the bypass mod, but this isn't necessarily so. They are often done at the same time, but aren't required. (This becomes important when chasing oil pressure issues high or low).

The original SAFETY valve opens at 8 atu, or 113.7 psi. The original RELIEF valve opens at 5.4 atu, or 76.8 psi.

Here is some information from Walt about the spring rates:


2.7 (before the mod) had 70mm free length springs. It took 10.6kp/104newtons/77lbs/ft to compress this to 52mm, and 14.1/138.3/102 to compress to 46mm. Full compression length of spring is 33.3mm.

The 3.0 had a safety spring 70mm long, and a pressure spring 86.6mm long (this is the one with a guide). Safety spring had the same specs. Pressure spring at 50.5mm is 80N/8.4kp/59 lbs/ft.

For the 3.2 (which I think has the same stuff as the 3.0): Safety opening pressure is 8 bar/114 psi. The pressure spring for the 3.2 is 6.2 +/- 0.8 bar, or 88.2 +/- 11.3 psi. All of which fits how your engine used to work before something went wrong.
Here are some actual spring measurements from Aaron Burnham:


Walt, here is some data. If you need a specific let me know. I used a spring compressor tester to get the rates

Long spring(dims include piston and guide tube)
60mm=6 lbs
50mm=12 lbs
41mm=19 lbs (guide tube limits travel @ 40.5mm approx)

short(dims include piston)
60mm=6 lbs
55mm=12 lbs
50mm=20 lbs
Implied spring rates-- for the long spring
85mm-60mm = 25mm of compression for only 6 pounds of pressure-- so 6/25 =0.24 pounds/mm (6.154 pounds/inch) (first 29.4% of length)
85-50mm = 35mm for compression fo 12 pounds = .343 pounds/mm (8.709 pounds/inch) (first 41% of length)
85-41mm = 44mm for compression of 19 pounds = .432 pounds/mm (10.968 pounds/inch) (51% of length)

So this is a "progressive" spring that gets increasing resistance as it gets more compressed.

For the short spring:
70mm- 60mm = 10mm for 6 pounds = 0.6 pounds/mm (15.228 pounds/inch) (14.3%)
70mm-55mm = 15mm for 12 pounds = 0.8 pounds/mm (20.305 pounds/inch) (21.4%)
70mm- 50mm= 20mm for 20 pounds = 1 pound/mm (25.413 pounds/inch) (28.6%)

This is also a progressive spring, the rate increases as it gets shorter.

Spring rates tend to lag for the first 20 per cent of compression, then be relatively constant for the 60 in the middle, then increase for the last 20 per cent.

As Walt's photo shows, the original plug has a 12mm depth. The second plug has a groove around the head, and is 9mm deep, as is the final hex-head plug 9mm deep.

For the VERY early engines (up to 901282) there was a 2mm thick washer used between the first version plug and the case. The depth of the bore from the face to the piston seat was 57.5mm. Later, from 901283 onward, this was revised to be 59.5mm.

A standard crush washer is 1.5mm thick, it gets compressed down to about 1.3 on installation to 60NM.

* * *

An interesting aside: the original 944 used the same type of piston/spring arrangement, which was later changed to something like this:

Rather than relying on a potentially imperfect bore in the case, this is a complete insert that slides inside, with the o-ring intended to seal the end against the bore. When pressure builds up inside a machined steel piston inside this assembly, it overcomes a small spring, venting pressure into the original vent passage. I suppose they did this more expensive solution (this part is like $400) to have more precise tolerances and better control over oil pressure.

To continue the aside, this is the 944 diagram with the oil cooler and filter in the pressure circuit. They are a lot more effective this way, there is less air than on the scavenge side.

While we're on asides, the apple doesn't fall very far from the tree. . .

The 356 used a variety of springs. The A case lists:

547 07 113, 11 x 65mm for type 692/0 engine
999 521 016 00, 12,2 x 66mm for type 692/1/2/3 (same diameter as modern spring)
999 521 012 00, 12,5 x 61.5mm for type 692/1/2/3

* * *

Now, the 993 actually used a spacer between the plug and the spring, Walt, did you ever measure this one? See illustration below.

Also, I hear that the GT3 uses an even deeper plug, >12mm depth, along with a huge 102mm spring, but I haven't seen the specifications on this. Can anyone help?

Spenny_b 10-22-2012 11:55 AM

Nice write-up John, will have a good read later on when it's too late to be using air tools, lol ;) (I predict I've just been out-nerded!)

This is what you mentioned the other week, yes?


304065 10-22-2012 12:46 PM


Originally Posted by Spenny_b (Post 7045877)
Nice write-up John, will have a good read later on when it's too late to be using air tools, lol ;) (I predict I've just been out-nerded!)

This is what you mentioned the other week, yes?


Spencer, greetings! Yes, this is the compilation I was talking about-- some way to get all the references in a single spot for chasing these issues.

Out-nerded. . . well one doesn't get much dryer than the factory workshop manuals, coupled with the obscure dates of precisely WHEN the factory did WHAT. . . perfect for the armchair (or barstool ) pontificator! :)

Walt Fricke 10-22-2012 05:44 PM

I'll look to see if I wrote down the thickness of the spacer. It will soon be coming out of the motor, so I can measure again.

I did not run across any reference to a longer GT3 spring, or deeper plug. I can look up part numbers for those.

The GT3 still uses the original overpressure relief spring!


KTL 10-23-2012 02:43 PM

Nice compilation John. Very timely for my current thread going on, which I suspect you've seen?

Also note that your diagram here

is applicable to the 964 series 3.6 The PET indicates a spacer is used for all M64 N/A engines as well as the M30.69 (965 3.3L) and M64.50 (965 3.6L). The vertical spring installation is the same set of parts for all 964 engines. Same screw cap, seal ring, guide sleeve, spacer, pressure spring and piston. ALL the same, be it an engine with an oil filter console or w/out.

Spenny_b 10-24-2012 02:03 AM


Originally Posted by KTL
is applicable to the 964 series 3.6 The PET indicates a spacer is used for all M64 N/A engines as well as the M30.69 (965 3.3L) and M64.50 (965 3.6L). The vertical spring installation is the same set of parts for all 964 engines. Same screw cap, seal ring, guide sleeve, spacer, pressure spring and piston. ALL the same, be it an engine with an oil filter console or w/out.

Hi Kevin, also a timely comment (for me) about the 964 series...when I was stripping the engine (964 M30/69 3.3) this time last year, I meticulously bagged and labelled everything, yet when I came to reassemble the pressure relief screws, there was no spacer (item#5) in the vertical assembly.

Either a previous owner has been buggering about with this (and curiously, the 3 piece rigid-flexi-rigid oil feed pipe that goes onto the crankcase looks to have also been tampered with....related perhaps?), or maybe the Porsche info isn't quite correct?

I've got a new spacer in the box of bits - kind of "on standby" - when I posted a Q a few months ago, the advice from Walt was to put it back to how it came apart, then see what the oil pressure is, re-install the spacer if necessary. I'm typing this, I think I'm going to install the shim, then remove if necessary...


47silver 10-24-2012 05:41 AM

removing piston
how do you remove a stuck piston the one shown on the B section of the above diagram?

304065 10-24-2012 08:01 AM


Originally Posted by 47silver (Post 7049449)
how do you remove a stuck piston the one shown on the B section of the above diagram?

I would gently use a wooden dowel inserted inside the piston. You absolutely do not want to bugger the aluminum bore that it rides in.

The VW pistons are removed using a special tool, but I don't know if it's compatible with the 911 pistons, never tried it.

KTL 10-24-2012 08:31 AM

I agree that if the oil pressure was fine prior to disassembly, it would seem reasonable to reinstall the same set of parts in the way it came apart. However the factory oil pressure gauge is a bit hard to obtain detailed before & after information since it's gradations of bars make it difficult to notice a difference in incremental pressures.

A mechanical gauge would be very helpful to use in testing the before and after. A mechanical AutoMeter or VDO gauge, an adapter fitting for the case, and a length of AN-4 hose is all you need and the whole assembly is not very expensive at all. You could plumb the mech gauge in a test-only way by just laying/hanging it in the engine bay temporarily. I've got all the pieces you need if you want to copy it or even borrow it.

The stuck piston is easiest, of course, when the case is apart. You can gently push it into the center of the case, to avoid scratching the bore. If the case is assembled, you can try a very strong magnet to grab the piston and pull it toward the threaded/exterior end of the bore.

47silver 10-24-2012 08:34 AM

that tool looks like the ticket
i dont think i can push the piston to the inside of the case as it is not wide enough.
the tool you show looks perfect but it may be too wide above the spreaders.

304065 10-24-2012 10:00 AM

For a mechanical gauge, which I agree is key to diagnosis and testing, I used an Auto Meter M10x1.0 to 1/8" NPT adapter and an inexpensive WIKA gauge. Works great.

larrym 10-24-2012 08:26 PM



"when the 76.8 psi limit of the oil pressure RELIEF valve is achieved"

not long ago i posted a thread asking about this -

the 3.2 engine book says it is relieved at about 4.5 bar - 60 psi

why the difference ?????????

Walt Fricke 10-25-2012 11:05 AM

1) A stuck relief valve, because it is vertical, can be removed by pulling the idiot light sender on top of the engine (where the picture shows a mechanical pressure gauge), and using a suitable rod to persuade the piston to come out the bottom. This won't work for the safety valve, however, as the horizontal passage doesn't come out the other side of the case assembly. Don't recall if there is access to from the inside if the case is apart, but at worst one could apply air pressure in that instance.

2) It doesn't surprise me that Porsche gives one pressure for one model of motor, and another for another. Also note that the 3.2 oil pressure specs I got (from Bentley, if I recall) are higher. Don't know just what that might mean, other than they kept changing their engineering minds about how much pressure was needed. Lower pressure may equal better gas mileage, but not allow as high a redline RPM. The 3.2s have a lower rev limit than the 3.0s.

3) Here are some part numbers for the GT3 (and I assume, turbo) motors, from a PET parts catalog for the 99-05 996 GT3 (as of Aug 06):

Spring 996 107 127 80
Piston 911 107 512 00
Spacer ring 996 107 532 81
Plug 996 107 105 80

Spring 901 107 531 00
Piston 911 107 512 00
Plug 999 064 026 02

-So the safety spring is the same as it always has been.
-The pistons are the same as introduced in the later 2.7s, with a groove instead of the earlier hole. I was surprised to learn, from this discussion, that the earliest upgraded systems used the new piston, but not the new, longer, relief spring. One hears that keeping the old spring will lead to oil pressure which is too high, but I've not tested that. Maybe that was why Porsche had to change by '78?
-The GT3 relief spring is a part number unfamiliar to me, and I can find no information as to its length (maybe the longer spring referenced in earlier posts?). I'm going to order one. Perhaps that would be the key to my oil pressure problems with my GT3 pump in an earlier case.
-I didn't bother with the numbers for the hollow tube part for the relief spring, but think it has had the same part number all along.
-I don't know why the relief plug has a 996 part number. I guess I should buy one of those to see how, if at all, it differs. Previously, I believe both plugs (safety and relief) were the same.

304065 10-25-2012 12:19 PM


Thanks for posting the GT3 parts. Porsche Classic only publishes for 10+ year old cars, so the PET isn't available on their web site at the moment.

Yes, pistons same since 1976 and the safety spring same all years. That will make a good part number quiz!

The 996 relief plug is supposedly deeper. For $50 it's a bit steep for a test, I think I will send Wayne an email and ask him to measure one if they have it in stock.

KTL 10-25-2012 01:22 PM

Yeah I also find it strange about the 996 part number. Recall that the 996 Turbo uses an air-cooled case design as well. So maybe that's what made that part number become a 996 prefix? But also recall that the 996 Turbo pump is now the superseding part for the NLA tried and true 930 pump. So I don't know what would be different about that cap from a typical earlier model cap......

HKZ Bob 10-28-2012 10:44 AM

Is this the buypass valve on my 76 Block.

sorry for the bad pictures


304065 10-28-2012 10:58 AM

Bob, that is the bleed port to the case-- excess oil that makes it past the piston drains back to the case through that.

In order to determine whether your case has the oil bypass modification, it is necessary to probe inside the oil passage that is the inlet to the pressure side of the pump. This passage is under the oil cooler-- with the cooler installed, remove the "S" hose and probe with a wire or better yet, a USB endoscope, all the way into the passage-- on the top side you should feel or visualize the port. An uncut case wil not have the port.

KTL 10-29-2012 06:56 AM

Here's a video by John Edwards at Costa Mesa R&D. Shows you exactly where the bypass port is placed in the original passage.

Porsche 911 Oil By-Pass Block Modification - YouTube

All times are GMT -8. The time now is 11:14 AM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2021, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0
Copyright 2020 Pelican Parts, LLC - Posts may be archived for display on the Pelican Parts Website

DTO Garage Plus vBulletin Plugins by Drive Thru Online, Inc.