How to modify the 3vz-fe cylinder head (and others).

[4v6]

How to modify the 3vzfe Cylinder head, a guide by 4v6.


The disclaimer.
You, yes you, are entirely responsible for any work you decide to perform on your engine or vehicle.
The procedures described in this article can be accomplished by most anyone with common sense and a fair degree of patience, skill and time.
It cannot be stressed highly enough that the three components above are ESSENTIAL to producing a quality job that performs as expected.
If you do not posses the necessary qualities, or you doubt your ability to perform the following actions, then do not proceed to attempt them.
In short, if you destroy your engine or hack your own head off with a grinder in the process of making modifications based on procedures described in this article, then its not my responsibility, no one made you do it!
Use common sense and employ safe working methods when making any modifications.

I have tried to ensure as much relevant detail as possible is included but this is still a rather condensed account of how to proceed because the flowtesting part isnt included and errors and omissions can unfortunately creep in, so if in doubt at any point, re-read and if still unsure, ask!
The information contained herein is based on extensive testing I have carried out myself therefore any quoted figures may or may not compare directly with others experiences due to differences in either testing regime or flowbench differences.

Now thats out of the way, lets see what tools are needed.

Youll need either a high speed air or electric die grinder (variable electric is best) although an electric drill can be used and at a pinch a dremel type tool- however theyre generally unsuitable because the tools are smaller and remove less material.
Better to use that for detailing/cleaning up jobs later on.
A supply of 80 and 120 grit abrasive cloth on a roll.
A carbide "Oval" Burr.
Sanding Rolls 80 to 120 grit.
Some wd40.
Engine degreaser.
Paper towels ( lots of them)
A vacuum cleaner would be useful also.
A plastic storage box for washing purposes.
A valve spring compressor to remove the valves.
A piece of ¼ inch rod around 6 inches long with a slot hacksawed into one end.
A dust mask and some eye protection.

At this point ill assume you have enough knowledge to have removed the heads off the engine,
removed and ordered the valves correctly.
Read the engine manual if youre unsure how to do that and get familiar first!
Clean the head in the plastic box with Jizer or some other engine cleaner that has a corrosion inhibitor in it otherwise the guides and seats will rust.
Once done with the majority of cleaning, dry the head and blow it off with compressed air if available and use your can of WD40 to coat the guides both inside and out and the seats to further protect them along with the cam cap retaining studs.

Unless you want your hands torn to shreds, its a good idea to deburr all sharp external edges with a carbide cutter or sanding roll, so do that next and save your fingers.

Ok, if you look down the intake ports youll see all manner of interesting stuff.
Casting marks, valve guides, the bosses and ridges left by machining ops at the factory etc.
The first thing to note is this: Be guided by the fact that youre not going to be making these ports monstrously bigger so resist vigorously the urge to go bananas with your carbide cutter!
If youve got a pile of chippings as deep as your armpits when youre finished, then you may have gone a little too far.....

This first photo shows a view down the port as itd normally appear minus the valves of course.



Things to note.
Rough surface texture, guide bosses and guides, fairly wide splitter, machining ridges left by the factory.
Incidentally, the reason theyre not removed ( on newer engines they are very much better finished) is because the engines built to meet a specific aim in terms of power and cost which, even in this state it does so, plus its a design thats over 15 years old and was never really intended to do what we're asking of it.


The next shot shows the chamber, the seat and bowl areas.
Ive sprayed these with a welding product called flaw finder/developer, it highlights contours quite well.
You can clearly see the ridges and changes in port cross section thats been left by the factory machining ops.
At the other side of the port theres a sharp edge.



Ive already taken port moulds of these heads on numerous occasions and you can clearly see how uneven and poorly finished the surface comes as stock.





So where to start?
First attach a vacuum cleaner to the port and use it to keep the dust and chippings down.
If you cant use a vacuum cleaner, then wear a dust mask, aluminium dusts not supposedly good for your health ( alzheimers) so protect yourself just in case.
I usually start at the port mouth and smooth the surface, just removing casting lines, flash etc.
Its important to run the tool fast enough to cut the port material but not press on it as itll clog up rapidly and end up slowing you right down.
Just let the tool cut steadily and keep it moving.
Youll develop a natural sweeping motion as you progress.
Once youve done all the port floors or roofs, turn over the head and do the opposite sides, just removing marks and ridges.
By now youll start wondering if youve done the right thing and worrying how long itll take.
Its not a fast job and thats where the patience part comes in.
This is how itll look as you proceed.







After a few hours youll have removed the casting marks and generally smoothed the surface.

The valve guide bosses on these can give you a couple of cfm for just a few minutes work, so its worth spending that time to reduce their size.
Just remove the material at either side and lower the height of the surface blending it in to meet the guide, making it more aerodynamic.
Try to get both sides of the boss symmetrical and mirroring its neighbours but dont worry too much about surface condition at this point, minor changes wont make any noticeable differences and in any case the surface will be addressed lastly after everything else is done.
Use the photos as a guide for what to aim for.

Have a look at the following photo of an unmodified intake port.
Its an experiment I did some time ago using smoke trails to get a visual clue where the air was going.
Look at the right hand trail. Its got a definite bump to it that proves the pressure is higher as it travels over the guide boss, its almost a dead copy of the profile in that area.
What happens is the air is deflected up by the high pressure (well higher than elsewhere) created as the air is flowing.
The pressure is raised in that place and diverts the air, stopping it from going where it would like- straight.
You obviously cant eliminate this effect entirely but by making the bosses less pronounced, that pressure rise isnt as great and Its very easily measured on the test bench.



Assuming youve got all the guide bosses reduced in height and nicely shaped and the ports lead in sections all roughed out, we can turn the head over and start from the other side.
Position the head so the manifold face of the intake port faces upwards.

At the bowl area use the oval burr to remove the edges of the alloy that overlap into the airstream, blending the material so it ends up level with the valve seat insert.
Youll notice at the sides of the port a deep ridge that alternates on each set of ports, caused by the factory machining.

[4v6]

Feather these into the seat insert and port as youve done with the rest of the bowl.

Now turn the head over ( prop it up on a length of timber if needs be) so the intake manifold face is pointing downwards.
The short side radius turn is next up.This is opposite the bowl that youve just blended in.
If you look at it, youll notice an apparent lip, almost an overhang where the factory tooling has cut into the port to produce the seating area for the insert, its the "arch" shape you can see on this port.



This sharp lip is detrimental ( very) to getting air into the cylinder because air dosent like to take sharp turns and when its moving at close to supersonic speeds as in a port it really dosent like to change direction because air has mass. Hang your hand out the car window at 80mph and feel the force.
The air coming off that lip creates vortexes and back eddys that make the stream turbulent which stops air further back up the intake from going where we want it.

When reshaping this turn we want to be making it aerodynamic, as in an aerofoil section of a wing, which is what the profile will resemble when we've finished on it.

Ive already done the hard work for you here and made a copy of the profile to aim for in 3D out of silicone rubber.
Compare it with the mould of the stock port on the right for the differences.
You need to try to copy the shape as closely as you can and use it to check the shape of the turn as you go.
This is one of the hardest parts of the job and itll take quite some time to get it just right but the payoff is way better flow with the valves in than youd get just by rounding that edge off.





Once youve carefully shaped the short side radius, you need to apply some efforts to the sides of the port walls where the SSR meets the vertical parts.
If you position the head on its end and look back up the intake from the valves, youll notice a "hump" type shape as the port wall makes the turn.
Its present on both outer and inner walls ( splitter side).
Youll need to be careful how you approach this bit of the job, what youre doing is making room for the air as it slows and goes around the SSR and blending the sides and SSR together, but not cutting into the already modified SSR.
In my case I transfer the head back and forth to the bench, test it, modify the area then retest to check the changes, takes ages.
Youre not going to be able to do that, so just aim to smooth and flatten that hump on both sides of the port, blending it in to the SSR at the apexes.
A gentle touch with a finger can usually detect what the eye dosent, so use that as a guide and stop when it feels smooth and flat.

Now youve done all these, turn your attention to the port divider/splitter.
The basic idea here is to slim the bluff nose of it down, reducing its width at the leading edge and smoothing the surfaces as it proceeds further into the port.
No need to knife edge it, just apply a small radius so its not dead sharp and dont work solely on one side, itll end up biasing the flow one way or the other which affects overall flow, again patience and care will see you right.
Use a pair of dividers to keep it central in the port by measuring.

You can see the basic shape to try and reproduce on the following photos of the port moulds.





Once youve completed all the major works in the ports you can start to apply a surface finish.
I use 80 grit sanding drums and 80 grit emery wrapped around a rod to further flatten and smooth the surfaces and impart a nice flow friendly surface.
Its best to run this at a moderate speed as too fast will just make it uncontrollable and the media wears faster.
You can use wd40 to help but I dont find it always necessary except in the chambers.
Start at the port mouth and work around the periphery, youll feel and see the effects of it almost immediately, continue that method all the way .
To get that cross hatch effect, stroke the tool back and forth covering the entire port, but be sure to not dwell in one area, keep it moving until youre satisfied with the results and you have a uniform finish.
This is what youll end up with.




In the chamber, use the sanding drum to carefully dress the surface and remove the myriad of pits and marks in the alloy.
Youre just aiming to remove sharp edges and blending surfaces.
Try to form smooth contours where the alloy runs down to the seats.
There are lots of edges in this part of the chamber and I use a homebrewed cutter to remove them evenly, but its possible to use old valves with the outer diameter reduced and the head ground with a bevel to protect the seats while youre working in the chamber with a sanding drum( its what I use).

Aim for a 120 grit finish in the chamber. You can go smoother but it takes a lot of effort and time and its likely not worth it.
Aim to get the ports and chambers as shown in the photos.







The exhaust ports are far simpler to work on believe it or not.
They too suffer from all the intakes casting flaws and can be improved quite markedly.
The 3vz fe has a compromised head design especially on the exhaust ports because theyre not identical across all three of them.
Two are of a "dog leg" design whereby one side of the port is a straight shot out with the other runner coming in at around 40 ish degrees and merging.
This creates turbulence and a loss.
All we can really do here is improve it so it flows better.
The "odd" port out on the end of the head is more of a siamesed design with almost equal length runners and this definitely flows better to the tune of around 10cfm over the other exhaust ports.
That kind of leads to a bit of a quandry.
Improve it or not?
Well yes, but bear in mind itll easily outflow the other exhausts if you apply the same measures to it as all the others.
I tend to develop the dog legged ports to flow equally and as highly as I can and then do that "straight shot" port last, balancing the airflow to match the others.
Takes time and effort.
All youll be able to do, is make them resemble each other as closely as possible and leave it at that.
I have it on great authority that balancing the exhaust flows isnt as critical so im happy to bow to much greater knowledge on that score.
Heres the 3vzfe's exhaust ports in silicone rubber, not a pretty sight.





Ok, proceed with the exhaust ports as per the intakes, removing casting lines and imperfections at the port exit (at the face and working inwards).
Theyre awkward to do due to their shape, but stick with it and youll get there.

The guide bosses are a fair size on these.
Reduce them as per the intakes, mostly on the widths and the "ramp" at the front of them leading to the guide, smoothing and blending them in.

Once youre happy with these, turn the head over and examine the bowl area ( under the valve head).

Youll notice some material just on and below the seat which needs removing and a rather large overhang.
I try to blend the overhang into the seat insert/ port but its pretty much impossible to remove it entirely as youd have to burrow into the material far deeper than really is good and its purely an aesthetic operation to remove it which you wont see when its all together anyway.
In this case, just smooth it as best as possible.
The sides of the ports at the divider have a similar "hump" to the intakes, so again its a case of basically flattening them and blending in to the rest of the port.

[4v6]

The exhaust also has an SSR which heavily influences how well the port can flow.
This mould shows the kind of condition the factory leaves the heads in.



A finger prodded into the port will expose a sharp edge which creates major restriction and turbulence just as on the intakes.
Once youve addressed that and modified it to more closely resemble the next photo, apply a similar finish to the exhaust port surfaces as the intakes using the methods youll now have developed.
You can finish these to 120 grit as itll help to reduce carbon buildup than a rougher surface which can alter the flow of the ports over time.

Modified exhaust port shape.










If youve done a good job, you should have ports that are closely matched for both flow and give a good improvement over stock.
Typical bare port flow for a stock 3vzfe is 204cfm@28", the result of your efforts if you work diligently should enable an improvement to approx 230cfm.
Typical exhaust flows are 154cfm@28" on exhausts, yours should rise to at least 165-170 after mods.

At this point in time, ive done limited tests regarding valves and seats, so its far from clear what the ultimate seat angle, width and number of cuts plus optimum valve shape is for these ports, that will come later when I have more time to spend on them.

I hope this article is useful.

Copyright 2009 T.Warren ( 4v6)

[U-235]

I am leading to believe that the "flaw finder/developer" is simply developer compound for penetrant inspections?

[4v6]

Originally Posted by U-235

I am leading to believe that the "flaw finder/developer" is simply developer compound for penetrant inspections?

You are 100% correct!

That photos around 2 years old, it was part of an experiment (along with the smoke tests) to see how the air was moving in these ports.
The whole port was lightly sprayed with the developer ( that white powder) then the port attached to a vacuum and the flaw finder/dye introduced via aerosol into the stream- but very lightly indeed and across the port mouth at around a foot away.
It shows up as red staining on the port walls where the dye drops out of suspension (a bit like fuel) and although its not 100% representative of whats going on it does at least give some visual assistance.
Incidentally, most all of it falls out of suspension at the splitter just at the turn and into the bowl.
Turns out the airs not actually doing much there in comparison to the rest of the port and it shows up as being so when you modify the splitter area- not a lot happens.
Most of its going over the SSR and round the outside of the guide and over it.

[328FTW]

Where was this thread when I did my 3sge head purely using logic and pointers from questionable honda threads? I like that I did almost exactly this just by thinking about what would improve it so that puts my mind at ease.

Good write up man its really complete for the length.

[4v6]

Originally Posted by 328FTW

Where was this thread when I did my 3sge head purely using logic and pointers from questionable honda threads? I like that I did almost exactly this just by thinking about what would improve it so that puts my mind at ease.

Good write up man its really complete for the length.

Thanks for the comments.

I did do some work on the 3sGE heads as well, i found that although its possible to gain quite a bit of flow on the bare intake port, once the valves go in you end up with about 5cfm of extra flow at the upper portion of lift but at least the same loss at the lower end of its curve- obviously not what i wanted!
Im surmising that valve shapes/seat angles may have a large influence on these GE ports as theyre pretty good as they are.
Again, more investigations required on these.

[mr2good]

Im just curious if you dynoed before the port job and after the port job? Cause to get the proper flow and to know if you actually improved it instead of hurting it, it should be wetflow bench tested right? Alot of peoploe port there stuff and actually hurt it instead of doing good. I think DART heads is the one who pioneered wetflow bench testing cause they found out just doing a flowbench test wasnt accurate enough. Anyone can hog out a head and increase volume, but if it doesnt flow right its not going to help at all Not saying youre doing it wrong or anything just wondering if you actually dynoed before an after to seee the real results?

[4v6]

Originally Posted by mr2good

Im just curious if you dynoed before the port job and after the port job? Cause to get the proper flow and to know if you actually improved it instead of hurting it, it should be wetflow bench tested right? Alot of peoploe port there stuff and actually hurt it instead of doing good. I think DART heads is the one who pioneered wetflow bench testing cause they found out just doing a flowbench test wasnt accurate enough. Anyone can hog out a head and increase volume, but if it doesnt flow right its not going to help at all Not saying youre doing it wrong or anything just wondering if you actually dynoed before an after to seee the real results?

Fair questions.

Best result to date is 215.8 bhp using precisely the methods employed here but i didnt build the engine , i just did the heads so ive no way to know whether it was set up correctly or not.
Im expecting to see a fair bit more than this shortly once ive tested a few other things that are currently in the pipeline.
Worst result was 207bhp again, i only did the heads (done to the same spec as the 215.8 heads) and i do know the exhaust system was well below par being a mismatch of pipes and bends etc which couldnt have helped.

Ive never been an advocate of the "hog out" method, indeed, i think i did say at the start that i wasnt going to make them bigger.
Id rather stay with the stock-ish sizes and alter the shape.

Ive actually destroyed two cylinder heads ( on purpose) to test out the results of making them bigger and i gained precisely squat in terms of flow by doing so.
I built my own flowtest bench so id have less guesswork and more hard data to rely on, otherwise i can tell you, id not have put this article up for scrutiny.

The wet flowbench- yeah thats a bit beyond me at the present and since the FE heads arent in any way a race quality item, i dont think going to that extreme is going to get us much on a street engine, but i may be wrong.
Race or drag heads sure, theyre after every last ounce of power so it makes it worthwhile for those guys.

[U-235]

Originally Posted by 4v6

You are 100% correct!

That photos around 2 years old, it was part of an experiment (along with the smoke tests) to see how the air was moving in these ports.
The whole port was lightly sprayed with the developer ( that white powder) then the port attached to a vacuum and the flaw finder/dye introduced via aerosol into the stream- but very lightly indeed and across the port mouth at around a foot away.
It shows up as red staining on the port walls where the dye drops out of suspension (a bit like fuel) and although its not 100% representative of whats going on it does at least give some visual assistance.
Incidentally, most all of it falls out of suspension at the splitter just at the turn and into the bowl.
Turns out the airs not actually doing much there in comparison to the rest of the port and it shows up as being so when you modify the splitter area- not a lot happens.
Most of its going over the SSR and round the outside of the guide and over it.

I honestly never thought about using Type II dye that way. I'm an NDT person by trade and it's always cool to see stuff like this.

[4v6]

The dye thing was a bit of a brain storm and wasnt planned at all...just had one of those sit up in bed at 3am moments with it going round in my head.
Mind you it does tend to pong a bit and is a bit messy.

[deelymc]

great write up. i will atempt this starting this weekend. i was reading the manual and it said to remove the head bolts carefully or you might damage the head. is there some particular order? also is there any easy mistake i might make grinding that will completly destroy the head?

thanks. also i received the ps mount about a week ago-it looks beautiful, great work.

[4v6]

Originally Posted by deelymc

great write up. i will atempt this starting this weekend. i was reading the manual and it said to remove the head bolts carefully or you might damage the head. is there some particular order? also is there any easy mistake i might make grinding that will completly destroy the head?

thanks. also i received the ps mount about a week ago-it looks beautiful, great work.

Hiya, glad you got the mount ok!

I have seriously bored out the ports on these heads to find out just how much material exists between the ports and water jackets and theres no way youll ever break through if you just follow the guide.
Ive also cut them in half to have a look.
I didnt manage to get a break through at any point so its very unlikely unless you get a casting thats got flaws in it but so far no, theres no way youll scrap the head just doing this.
Just be careful to avoid hitting the seats at all otherwise its off to the speedshop for a valve job.
If you get stuck at all, let me know.

[MR JAS]

my only question is how did you make those molds they dont seem like hey would want to come out in one piece after hardening

[4v6]

Originally Posted by MR JAS

my only question is how did you make those molds they dont seem like hey would want to come out in one piece after hardening

Two part castable silicone is your friend in this case.

You can use latex but its not as durable and can tear, the silicone can be stretched to something daft like 600% of its original length and still springs back.

I made sure the ports were clean and dry before using a little furniture wax sprayed down them, then seal the ports off with the valves and slowly pour the silicone in, let the air bubbles rise out, top it off and let it set.
It starts to go "green" ( just setting) after about an hour or so but its best to leave it 24 hours before removing it as itll have reached full set by then.
I just push it out steadily from the combustion face and et voila!

[brianbooth]

The moulds are great for seeing the actual shape of the air flow. Thanks a ton for this thread.

[MR JAS]

somebody sticky this

[hfmt93]

Very good writeup!

This is very good info, but it is easy to get carried away with all this. 4v6 is serious when he says it takes patience. It takes a lot more time than anticipated. I really do not even see how you got into the bowl so good. Can you show a pic of all the tools? Especially the one used for the bowl area.

I messed up by trying to contour the casting marks in the side of the tunnel. You get carried away and ready to be finished and stop thinking. I have now found a new set of heads and will probably be leaving them stock. For now. This pic shows the casting marks I messed up on. Use caution:

Originally Posted by 4v6

http://i82.photobucket.com/albums/j2...203816x612.jpg

Originally Posted by mr2good

Im just curious if you dynoed before the port job and after the port job? Cause to get the proper flow and to know if you actually improved it instead of hurting it, it should be wetflow bench tested right? Alot of peoploe port there stuff and actually hurt it instead of doing good. I think DART heads is the one who pioneered wetflow bench testing cause they found out just doing a flowbench test wasnt accurate enough. Anyone can hog out a head and increase volume, but if it doesnt flow right its not going to help at all Not saying youre doing it wrong or anything just wondering if you actually dynoed before an after to seee the real results?

Mr2good makes a good point. From looking at all the data 4v6 has come up with I believe these will make power, but the only real way to tell is drag times and dyno in my opinion.

One thing I noticed when messing up mine is the valve seat insert's machining made the insert a "step" into the airflow. Did you see this on any of you heads? Did you mess with it?

4v6 this is amazing work and data. Beautiful curves on the finished ports. I always enjoy reading your threads like this. Please, keep up the good work and thanks!

[4v6]

Originally Posted by hfmt93

Very good writeup!

This is very good info, but it is easy to get carried away with all this. 4v6 is serious when he says it takes patience. It takes a lot more time than anticipated. I really do not even see how you got into the bowl so good. Can you show a pic of all the tools? Especially the one used for the bowl area.

I messed up by trying to contour the casting marks in the side of the tunnel. You get carried away and ready to be finished and stop thinking. I have now found a new set of heads and will probably be leaving them stock. For now. This pic shows the casting marks I messed up on. Use caution:







Mr2good makes a good point. From looking at all the data 4v6 has come up with I believe these will make power, but the only real way to tell is drag times and dyno in my opinion.

One thing I noticed when messing up mine is the valve seat insert's machining made the insert a "step" into the airflow. Did you see this on any of you heads? Did you mess with it?

4v6 this is amazing work and data. Beautiful curves on the finished ports. I always enjoy reading your threads like this. Please, keep up the good work and thanks!

Hiya, thanks for your feedback.

Yeah the access issue can be a hard one for someone doing this at home, but most of the tools are fairly simple.
A quick description of what i use.
Oval carbide burr- thats the item shown removing material from the port.
Thats the most useful shape ive come across so far because it tends to blend surfaces quite easily.
Important not to press on it tho, just let it cut.
I run it at around 7krpm ( slowest speed on my grinder) because it becomes a problem with clogging at higher speeds and less controllable.

I also use sanding rolls from standard abrasives- they do a small kit which has lasted through at least 10 heads so far and still going.

One of the most useful items is a split fork deburrer.
Its just a length of 5 or 6mm thick rod, with a slot in the end- i have two, one has a flat end the other is shaped like an artillery shell.
You just tear off a strip of emery, fold it and slot it in equally then use it like a sanding roll.
Its a vital piece of tooling for finishing as its extremely gentle even in 80 grit emery yet it can still remove material and further blend hard to access areas of the port.

Ill take some photos later of what im using to better illustrate them.

These 3vzfe heads do have lots of steps and dimensional changes that have an effect on the airflows, but ill tell you this; theyre very easily made to flow.
Most people can do them, but where it gets tricky for the homeporter is knowing whether or not theyre balanced and all flowing the same.
Ive found its very easy to get large discrepancies in flow if youre not consistent in how you approach these, its very likely the same with others too.

I think ive probably modified most all aspects of these ports by now, from making them bigger, smaller, wider, higher, changing the angles the air flows in, waisting the valve stems, changing the chamber shapes and on and on....
Thats not to say i know every aspect of whats possible with these cos i dont think im even close yet but its nice to have actual data to refer back to.
One area i do want to focus on before too long is valve and valve seat geometrys.
Its far from clear how well these are doing as regards flow and i feel certain there are some nice gains to be had in that area alone.
One thing the 3vzfe intake valves do have that makes them flow quite well is a 30 degree back cut on the reverse side. so Mr T's engineers have obviously done some work here.

[4v6]

As promised a few photos of the tools i currently use.

Metabo high speed die grinder goes to about 27,000rpm uses the flexi shaft from the same maker and is rated to the speed of the grinder.





Has a compact head and is very user friendly gets in to the tightest of spaces but heck its expensive.
Cost almost as much as the grinder itself.
Quality item tho so no complaints.



Two split fork deburrers.
The one on the right is for final finishing.
The left one can be used to smooth the surface of carbide cutter marks.
You just have to be careful to start the grinder with it in the port tho or it whips around.





Oval carbide burr- the best shape for general work in the ports that ive found.