Fuel injection rebuild

[Simmy]

Thanks for the pointers - I will call around today for pricing. Not sure if there are any recommendations, but I think Powerkote are the guys to go to up in JHB?
I am going to get the crank, fly, clutch and pulley balanced again as a unit, but the pistons and rods already went for balancing after the first regrind - not going to have them done again. I will have to check again that the coating won't make any difference to the mass or deck height.

But back to the build - I managed to check the geometry, but decided not to cut the pushrods until I had barrel shims on the 3/4 side. I know I should only have to measure one valve, but I am paranoid and seem to like to triple check myself to avoid costly screw-ups later.

Found that the stolen clothes peg method works well enough for the lifters.


Triple checked that my cam measurements for peak and half lift were still good... landed up finding that (at least on this rocker) that the ratio comes out to 1.14.


With one 1.5mm shim (using 1/2 height relative to the 0.05 opening point). Leaning over slightly to the rocker shaft.


With one1.5mm shim (using 1/2 height relative to zero lift) - as expected the adjuster leans toward the rocker shaft slightly more.



With two 1.5mm shims (using 1/2 height relative to the 0.05 opening point). Very close to parallel. Relative to 0 lift, I would think it would be close to dead on.

[fourier]

With one 1.5mm shim (using 1/2 height relative to the 0.05 opening point). Leaning over slightly to the rocker shaft.


With one1.5mm shim (using 1/2 height relative to zero lift) - as expected the adjuster leans toward the rocker shaft slightly more.



With two 1.5mm shims (using 1/2 height relative to the 0.05 opening point). Very close to parallel. Relative to 0 lift, I would think it would be close to dead on.

Setting up valve geometry is a tedious but necessary task, if you want parts to last. Just checking one valve is risky. When you place a straight edge across all four valve stems, you will see variations, even on new heads. I learnt this the hard way, by setting up geometry and push rod length on one valve, thinking one size fits all... The result was a set of ruined 1.4:1 rockers, due to coil bind on three valves.

It may be an optical illusion (top two pics), but your dial gauge seems to be sitting at an angle in relation to the valve stem. In order to take accurate valve lift readings, the dial gauge travel should be parallel to the valve stem movement. How close are you to advertised cam lift? If the geometry and resultant push rod length is off, you will sacrifice total lift. You want to end as close to advertised lift as possible.

When you blue the valve tips, what does the wipe pattern look like, and where is it located in relation to the valve centreline? The wipe pattern should be fairly even on either side of centre, after a full lift cycle. From the bottom pic, it appears that the contact point is above the centre of the valve stem at half lift (as seen from the rocker shaft side). If so, you probably have too many shims under the rock blocks. The valves sit at an approximate 10 degree angle to the head, slanted towards the rocker shaft. If you add shims under the rocker blocks, the valve/rocker contact point moves above the centre line of the valve.

When using the type of adjuster that you are using (or the "elephant foot" type), it sticks past the bottom of the rocker a lot further than a stock adjuster. These swivel feet adjusters often require grinding away a section of the rocker, just above the adjuster head, to allow it to screw into the rocker further, in order to correct geometry.

There are so many variables, and often requires lots of time to get right. Good luck!

[Simmy]

Thanks for the really detailed reply! The internets would lead you to believe that shimming the end play and getting it parallel with the valve stem is sufficient! But you make very good points, so I will see what I can do to rectify what I have.

I did check the levels of the valves some time back and I recall that the straight-edge touched all of them, so I will see again tonight.

I will need to check my notes on the lift values in order to confirm. But point noted on the dial gauge. I suspect there is some double optical illusion going on, because it looks skew to me as well, but didn't at the time. Like I said - I will do this a couple more times to make sure I get repeatable measurements... I have learned that my first attempts at this usually result in some error or omission.

When setting up the rockers, I followed a build thread to which I have now lost the link. But basically, I tried to get them slightly off-centre, but still over the stem throughout the range of motion (which I guessed at the time). The rationale for this was to allow the valves to rotate a little every open/close event in order to even out the wear. However, I can appreciate the notion of keeping it as close to the centreline to keep eccentric load to a minimum... so somewhat of a trade off but I realise that this must be applied to the "vertical" and "horizontal" alignment. That said, I will see what the wipe pattern looks like, because I think that will offer a more objective view.

However, I am a little unclear on the point made about grinding away at the arm. The adjusters thread in almost fully, but I think it binds on the thread, rather than on the arm itself. I suspect a little bit of counter-sinking on the arm would deal with this. I must also just check how much space is left between the cup on the adjuster and the valve. I know they come close, but I don't want them to contact.
Also... is it right to assume the best way to find the pushrod length is to check for each and cut to its unique length?

[fourier]

Let's clarify...

When viewing the valve tip from the top, draw a vertical line and a horizontal line through the centre. The horizontal line will be parallel to the rocker shaft.

You want to set up the rocker position so that it contacts the valve tip to the left or the right of the vertical line. This assists with valve rotation, as you have mentioned. This offset, is created by adding or re moving shims between the rockers and the rocker blocks. Do not get too close to the edge of the valve tip. A slight offset is all you need.

The geometry determines where the wipe pattern sits in relation to the horizontal line. If you blue the valve tip, and run the valve through a full lift cycle, you ideally want the pattern to be equal amounts above and below the horizontal line. If the pattern is mostly above the horizontal line (far side of the horizontal line in relation to the rocker shaft), you may have too many shims under the rocker blocks.

I usually set valve geometry to have the adjusting screw in line with the valve stem at 50% lift. Some prefer to have this alignment at 60 or 70% of valve lift, due to spring pressures increasing as they compress, causing greater side load on the valve guide towards full lift. This approach may be beneficial if you are running seriously heavy springs, but I always run the lightest springs possible, that are still capable of controlling the valve motion without float. In order to run light springs, I use light weight valve train components. Items like titanium valve spring retainers, heavy duty aluminium push rods (certainly not chromoly), and light weight valve lifters, come at a cost, but save significant reciprocating weight, requiring less spring pressure to control. Lighter springs will also extend the life of valve guides, seats, camshafts, and rob less power to drive the valve train. That is why I stick to adjuster/valve alignment at 50% lift.

About grinding away the rocker arms... I have found this necessary, every time I use swivel feet adjusters.



When you compare swivel foot valve adjusting screws, to a stock VW item, you will notice that the swivel end sticks out of the rocker a lot further on the valve end, than a stock VW adjusting screw. This has quite a detrimental effect on geometry, as you need to space the rockers away from the head quite a lot, to get the adjuster to align with the valve stem at 50% lift. Then you will end up with a wipe pattern that is mostly above the horizontal line (like what would appear to be happening in your last pic). Remember, the valve sits at a +/- 10 degree angle to the head, so shimming the rockers away from the head will move the wipe pattern as described above.

In order to correct this, you need grind the bottom of the rocker in the area where the adjuster screws into it. This will allow the adjuster to screw deeper into the rocker. You can then remove a shim from under the rocker block, and the wipe pattern will move closer to where it should be in relation to the horizontal line.





Or an even better, but more difficult/expensive way...






There is one added complication with the specific type of adjuster you are using. It has an oil hole, which needs to align with the oil feed hole in the rocker arm +/- 1 or 2 turns (allowed by the thinner area around the oil hole). If the oil hole area is not aligned to the oil feed, the adjuster will fail. These adjusters cannot really move much in relation to the oil hole, so this modification is probably going to misalign the oil holes, and cause failure.



If you decide to grind the rockers to correct the geometry, I would suggest going to the "elephant foot" or "Porsche" type adjuster, which is not reliant on an oil feed from the rocker, but uses splash-oiling. It can then screw further into the ground down rocker, to create the needed clearance to correct the geometry.



As far as making unique push rods lengths for each valve... Well, that would be ideal but probably not worth the time it would take. If all valves are at the same level, you would probably be OK with measuring one-size-fits-all, and going with that. You will not get close to coil bind with your cam, so no risk of suffering a similar failure to mine.

[Simmy]

I have gone back and reassessed what I have given your feedback. Thanks for taking the time to explain!

What I seem to have is the following:

• the wipe pattern is definitely on the upper half of the stem

• you can see that the foot rides rather close to the edge of the valve and needs to move further down.

• adding shims fixes this, like you said, but it I can see that the rockers need to be ground back, which should allow me to remove a shim, and get it closer to centre

• and the oil hole is not lining up. My plan is to grind open the rocker oil hole to meet the adjuster's. I can see these are not the favoured type, but I like the idea of getting oil all the way up to the valve, and I suspect others who have had issues probably didn't correct this problem.

One shim


Two shims


If I am not mistaken, getting the adjuster to thread in more will move the foot closer to center and allow for fewer shims.


An indication of the oil hole misalignment


But I can get a small diamond grinder to open it up.


So... Off to start grinding the rockers.

[fourier]

Take a look at the three diagrams in this post on the Samba... It saves a 1000 words

https://www.thesamba.com/vw/forum/viewt ... c&start=20

You want to aim for a set-up that resembles the last of the three patterns. The smallest amount of movement across the valve tip, either side of centre. This also results in highest valve lift.

[fourier]

Measure valve lift every time you make a change. That will show you if you are going in the right direction. You should end with maximum lift, closest to advertised cam lift.

[Tony Z]

Ah, the pain of setting valve geometry... another time consuming topic that most dont understand and can never fathom why a performance engine (done right) needs to cost so much.

Fourier and I have discussed this topic together many times when we have our "performance vw days" at my workshop. No-one knows this topic better than he does.
But, I do have 1 thing that look at from a slightly different angle, so I'm putting it in here just to confuse you further

The reason you aim for the rocker to be perpendicular at half lift is to reduce valve bending and guide wear.
The way I see it is that the guide wear will be less on the first half of the movement due to the spring pressures being low and then past half way, you initiate higher spring pressures making the valve bend more and wear the guide more than in the first half of lift.
So, I aim for the rocker to be perpendicular to the valve a little past half lift, usually towards 66% of lift. Now bear in mind that I've never physically sat down and done the maths on this to find the perfect point, but my head just screams 60% to 66% as the ideal spot.
I might lose a little lift but I feel I add longevity to the setup with my method.
This is probably one of the very few topics that Fourier and I havent agreed on yet. Oneday he'll see my logic

If the above confuses you and throws you into a ballpark where you dont know wtf to do, think of it another way.
Most people aim for perpendicular at mid lift. From a wear and tear point of view and methematically, things tend to point towards 60%to66%, so, you could use both and say that anywhere past half lift but before 66% is a good spot to end up - this might even make life a little easier.

Good luck and keep us posted.
Its frustrating now, but it'll pay off. You are busy with that 20% that others dont worry about. It'll be the difference between a nice engine and a great engine.

[Simmy]

Well, I definitely underestimated this part of the build... That's for sure. But I have the time, and I am more than willing to try and also learn.

I fully understand the logic. The spring load is proportional to displacement, so it makes sense to bias it toward full lift.

So one side done...

[fourier]

Good going with modifying the rockers! Just be very sure that you still have an oil supply to the adjuster.

I definitely don't disagree with Tony's logic I will go the same way, when using killer springs, but as stated before, I opt to run light springs and a light valve train. I therefore stick to stock geometry. It worked for VW, and will work for MY set-up. If you run a lot more spring, bias makes sense...

[Tony Z]

I will go the same way, when using killer springs, but as stated before, I opt to run light springs and a light valve train.

Wasnt it me that was telling you to ditch the heavy CB650 springs in favour for their cheaper duals 'cause they were overkill? I seem to remember you fighting this idea until I found the beehives.
???



Good going Nic

[fourier]

I will go the same way, when using killer springs, but as stated before, I opt to run light springs and a light valve train.

Wasnt it me that was telling you to ditch the heavy CB650 springs in favour for their cheaper duals 'cause they were overkill? I seem to remember you fighting this idea until I found the beehives.
???



Good going Nic

Mr Z, I remember it slightly differently... I never actually used the VW650's springs in a build. They were not bought because I needed the spring pressure, but only because they allow more lift than the CB duals. I needed to remove the CB dual springs as they ended up with coil bind, and wiped out a set of Scat ratio rockers with the cam I was running at the time (as mentioned earlier in this thread).

I've wanted to convert to light beehives years ago, but did not like the way guys were adapting them to our cars. They had to install longer valves to make them work, which hurt geometry in a big way.

Credit: When you discovered a beehive kit done correctly (thanks for that), I needed no convincing. Lighter springs able to do the job, will win every argument! I'm all for making things last.

Back to your thread Simmy... Apologies!

[Simmy]

No need to apologise! I didnt even know that beehive springs existed before your little debate here... so something gained!

Anywho, I got around to Volkspares on Friday. They were thoroughly useless - which is why I hate going there, but they do have the basics when you need them at short notice. They werent able to help with 0.1mm shims - but I found a place that can lasercut them from stainless shim stock. Has anyone come across a DXF/CAD file that I can reuse? Else I will draw up one myself later.

But I did get a rocker shim pack... so I managed to play around to see if I could maximise the lift and keep the adjuster parallel at half lift. For anyone else trying to do this... just a few pointers:

• When adjusting the pushrod, extend it to the point where the dial indicator just starts to lift off zero, and back it up just a bit. I found doing this by feel alone lead to variability in measurements
• Dont roll the crank over too quick, in case your indicator jumps and bumps the spring retainer. The light springs allow for a fair bit of play.
• Swap out and repeat with the same shim config several times... and check it makes sense.

And this is what came out of it:


Basically, very little shimming is necessary, which makes sense since the stroke remains stock, as are the rockers. I think I am going to settle on 0.017 since it has the same lift, the rocker axis is as close as I can get it to perpendicular with the valve stem, and even if the adjuster is no square as well, its only by 1.5deg, which means it will hit perpendicular in the 50-60% lift range, which is fine by me.




This is what the wipe patter looks like with the 0.03 shim - its about as small as I have seen it.

And here is a video of the same:
https://youtu.be/iYXpiI5m_Gw

[fourier]

Hats off to the effort you're putting in to get this right. Good going!

You have certainly gone through enough iterations, to arrive at these results. It is difficult to see from the pics, but if what you have, brings you closest to maximum lift, with the adjuster in line with the valve centre at half lift, and the smallest wipe pattern either side of valve centre... I'd say you're done!

After looking at the geometry diagrams from my earlier post, I realised why I would probably stick to 50% valve lift, to align the adjuster to the valve stem...

Looking at the first diagram below, which describes stock (50% alignment) geometry, you will notice that the wipe pattern either side of valve centre, is minimal. In the two diagrams at the bottom, the geometry is biased either side of 50%, resulting in a wider wipe pattern. The width of this wipe pattern, translates into radial force on the valve guide, as the rocker attempts to move the valve by the width of the wipe pattern, and the guide has to resist the movement. In short...a wider wipe pattern will probably cause more guide wear. (IMO)







On a different note. Just be sure that you still have enough clearance between the rocker covers and the adjuster tips, now that you have cut the rockers down. They tend to get close with this type of rocker modification.

[Simmy]

Thanks

There is still clearance at both 0 and max lift. I haven't measured it, but its probably 0.25-0.5mm.
Also checked the modified oil hole isn't blocked with compressed air aimed at the feed hole... You can feel the ball move when you do, so I am happy that the oil will make it to the same point, especially with all of the other mods.

Hopefully the other side will be a little quicker and easier, but I still need to get those barrel shims beforehand.