Topic: Getting more out of your 4 squirrels

[Wrench]

First of all, I am here to help those running a stock or mostly stock 2.5.  I have no intent for this thread to turn into an ego/pissing match about how to build an engine.  I have read nearly every page on this forum and found nearly everyone asking the same question (how to get more from your squirrels), and the only "real answers" seem to be to mod the crap out of it $$$$$$$$$$  or go V8.  I am here to give witness to an alternative.

I started my 2.5 build several years ago and found a few things along the way that I thought I would share.  I have stuck with OEM internal components in my build to see if more power can be made if modded in the right area. 

My current setup is: 1984 2.5, stock/OEM crank/rods/pistons/cam/valves. I am running a late model long runner MPI intake with an adapter to run a Weber 32\36.  I am running a home-made tri-y header that feeds a 2.25" Flowmaster, though I suspect a 2" would net better low-end power on a mostly-stock engine.  For the ignition system, I am running a Summit Racing digital ignition box with an adjustable rev limiter.  Rev limiter is currently set at 6k rpm.  I decked the block for a sub-0.030" squish clearance, and have to run premium fuel.  This engine flat-out pulls harder than any other 2.5 I have run with, and have even beaten several 4.0's in the mud boggs.

I believe the Jeep 2.5 and 4.0 heads both suffer from ports that are too large (for stock displacement that stays under 6k rpm) and poorly shaped.  I have been successfully running epoxy-filled intake ports for several years now, and the effects have been VERY good.  In my search for more power (on many different engines), I have found that port velocity has a very profound effect on overall power of an engine.   I would NOT advise that you remove much material at all if porting this head.  What keyed me into this was a question from my engine builder when I handed him the head early in my build (before I tried epoxy), "Those ports are pretty big!  You spinning this thing over 7k rpm?"  For reference, this guy has been building SBC race engines for over 30 years now.  I replied, "I didnt change the ports, only sanded the rough castings smooth".  So, here are a few things I have found to help this:
1) Advance cam timing.  I used a Mr. Gasket#720-988G 4 degree key on my crank, available at many parts stores dirt cheap.
2) Good multi-angle valve job.  The stock valve lift is a major restriction.  A good valve job will increase flow at lower lifts.
3) Better cam with more lift:  Clay Smith still lists cams for this engine http://www.claysmithcams.com/hydraulic-flat-tappet/
4) Not so inexpensive: Roller rockers.  Others have found that the stock stamped ones flex, at the cost of lift.  I would stick with 1.6 ratio as I suspect the 1.7's put too much strain on the 5\16 rocker bolts.  Best price I have found is on the Harland Sharp S40196.  I hear you can buy them in pairs from HS.
5) Last, but not least: epoxy.  You would be absolutely amazed at how much better this engine runs with less than $30 in epoxy put in just the right places.

If sticking with the stock cam, I would recommend the 4-degree offset key and roller rockers.  If you go with a good aftermarket cam, no need for the offset key as the advance is most likely built into the cam (check specs to be sure.  If the LSA and centerline are not identical, the difference is the advance).

I have also found that the stock valve springs are inadequate, for an engine that runs past 5k rpm, even with the stock cam.  I experienced valve float, which nearly dropped a valve in my engine.  Luckily for me, the keeper ring land shifted just enough to cause a ticking sound that sent me looking for trouble.  I upgraded to hardened keepers (Howards Cams locks pn HRS-93020 ) and two sets of springs (actually listed for a V8) from Comp Cams pn CCA-901-16.  These springs are 353 lbs/in spring rate with a seat pressure of 101 lbs at stock 1.65" installed height.   

I have also experimented with bumping the compression ratio, more in an effort to get a tight squish clearance than anything else.  I have found:
1) The factory squish clearance (piston to head) appears to be in the 0.100" range (depending on head gasket).  Completely and totally ineffective.  Even if you were to get a hold of a 0.020" metal shim gasket and use the 4.0 pistons with taller compression height (Silvo-light 3242HC), you wont get it under 0.040", where squish actually becomes effective.  You have to deck the block.
2) FWIW, the guy at Clifford stated they take 0.060" off the head to make a smaller combustion chamber, and don't even worry about squish clearance.
3) The thickest head gasket is Victor Reinz, followed by Felpro.  I cant quite remember the specs, but I believe the Victor gasket was around 0.070" and Felpro around 0.063" compressed.  I also found a 0.020" metal shim head gasket on my spare 1986 engine when I took it apart, but cant seem to find a listing for it.  I believe this is the stock gasket that came on it.
4) Decking the block to achieve a 0.040" squish clearance seems to be the sweet spot that nets a good squish clearance benefit, but is right on the edge of pinging on 87 octane/stock cam/stock pistons.  If you use the Silvo-light piston 3242HC, it has a deeper dish which should allow for a slightly tighter squish/higher compression ratio.  Different cam timing will also effect octane tolerance.
5) On a tight engine, you cannot go below 0.026" squish clearance without risking contact with the head.

My quest for a tight squish clearance also led me down the path of finding more conservative ignition timing.  Not much to note here except that the stock Jeep I6 distributor advance weights and springs are considerably more conservative in timing than the stock 2.5l ones.

Last, but not least: for anyone concerned, the piston to valve clearance on these engines is HUGE with the stock dished piston.  I dont see a piston ever hitting the valve unless the valve drops from retainer/keeper failure.

Hope this helps some of you.  Sorry, I don't have any real-world experience with larger throttle bodies on this engine, mine is carbureted.

[VA_YJ]

Paul,

Interesting tech.  Do you have any pictures of the epoxy work in the intake ports?  Thanks.

[Wrench]

Paul,

Interesting tech.  Do you have any pictures of the epoxy work in the intake ports?  Thanks.

Sadly, I dont have any pics.  I did all the work in a rush the week before a local mud-bog race and didnt stop to take pics.

On the long radius side, I filled in behind the ridge below the valve seat so the roof of the port seamlessly flows to the seat with no ridge.  I also filled the squared-off corners of the port roof and made the roof round.

On the port floor, I actually began the fill in the intake.  The stock MPI intake angles up for a few inches, then levels off for about a quarter inch before contacting the head.  I continued the upward slant that was started in the intake all the way into the head, and raised the entire port floor around 1/4".  This allowed inside radius to be larger and blend much more progressively to the valve seat area.  The port floor was left flat, so when you look at the entry on my head, it has a very distinct "D" shape.

I also necked down the valve guide to lessen the protrusion into the port. 

I used construction paper templates to "blueprint" the profiles with all the ports.

I have used Propoxy, JB Weld, and JB Quik.  All appear to be bonding just fine to the rough-sanded, clean cast iron.  I hope to have it apart again soon to do more adjustment and recheck the epoxy.  Since I did this work (about 4 years ago), I have learned a whole lot more about epoxies and how to work with them easier.  I plan to post my findings and pictures then.

This engine is in a 1984 Cherokee.  In the mud-bog, I took first place in the 4-6 cylnder class, beating 3 other Cherokees with 4.0's, not to mention 8 other rigs.

Funny thing is, the entire week before the race, I was out in the garage covered in dust grinding on the ports.  My wife comes out two days before the race and asks, "Is it worth it?"

I brought home the trophy and told her it was MORE than worth it...  

[VA_YJ]

Thanks for taking time to provide details.  Just a couple more questions: Did you do anything to advance the ignition timing (e.g., relocate the crank position sensor) or is this accomplished with your Summit box?  My '95 4 banger's MPI intake is aluminum - sounds like the late model 4 cylinder intakes are cast iron?

Maybe part of your sucess is that the lighter weight of the 4 banger helps keep your front end up (better weight distribution) and you get more floatation in the goo.  Of course, high RPM HP helps a lot with muddin'.  Have you noticed any loss of low end torque (the rock crawler's friend) with your mods?

[Wrench]

Ignition timing:  My (1984) engine runs a standard distributor that originally ran a Ford DUI box, the curve is built into the weights and springs on the distributor.  I removed the Ford box and replaced it with the Summit box.  The only thing gained was a rev limiter and msd below 3k rpm.  With my high compression ratio I had to retard the ignition timing to prevent the pinging (initial timing and total timing), but I cant remember how much.  I am not familiar with the later model ignition systems or how you can adjust them, and my engine does not have a crank position sensor.

The intake I am using is the latest MPI setup (~98?), the horseshoe shaped one, aluminum.  The cast iron reference was about the ports in the head.

The #1 improvement with the smaller ports is that the engine runs dramatically cleaner off-idle (remember, I'm carbureted) with excellent throttle response.  It has healthier midrange, with no loss on top end.

I'm running 33's with 4.88 gears, also.  I dont really like rockcrawling with this setup, as I am running a 5-speed.  It would be perfect with an auto, but really needs deeper gears to crawl well.  4.88's are as deep as I can go with my axles, though.

[chardrc]

interesting ideas with the port velocity. I would venture to guess that results will very from your carbed injection to mpfi since the quality and speed of atomization is significantly different between a carb and modern fuel injection. I'm going to have to dig out the internal combustion engines textbook and brush up a few things...

[Wrench]

interesting ideas with the port velocity. I would venture to guess that results will very from your carbed injection to mpfi since the quality and speed of atomization is significantly different between a carb and modern fuel injection. I'm going to have to dig out the internal combustion engines textbook and brush up a few things...

I wonder that myself.  Since most the critical changes happen downstream from where the MPI injects the fuel, it should flow the same as the carb setup?   I have found that the airflow characteristics seem to get more critical the closer they are to the valve. 

[sharpxmen]

you lose VE at high RPM with smaller ports and/or valves, same idea with 2 vs 4 valves and such, dunno about the rest (carb,mid-range, etc) but the high revs VE is pretty much known to be restricted by the size of the intake valves (and cam lobe profile as well) unless you have restrictions in other places (as in exhaust, carb or tb, air-filter & ducts) in which case that part goes out the door. The better the head flows the better VE at high RPM, a flow bench would be the place to get definite answers in this case.

Now, adding material does not necessarily mean there's a restriction in the "run" compared to before, the cross section could still be at least the same as the smallest or most restrictive x-section of the run which is usually at the valve seat so by adding material and "smoothing" the airflow it might result in less restrictive flow.

also, the 91-95 and probably later cams as well the timing of the centerline is altered to favour higher rpm (i think it's retarder something like 6 deg but can't recall exactly).

[Wrench]

you lose VE at high RPM with smaller ports and/or valves, same idea with 2 vs 4 valves and such, dunno about the rest (carb,mid-range, etc) but the high revs VE is pretty much known to be restricted by the size of the intake valves (and cam lobe profile as well) unless you have restrictions in other places (as in exhaust, carb or tb, air-filter & ducts) in which case that part goes out the door. The better the head flows the better VE at high RPM, a flow bench would be the place to get definite answers in this case.

Now, adding material does not necessarily mean there's a restriction in the "run" compared to before, the cross section could still be at least the same as the smallest or most restrictive x-section of the run which is usually at the valve seat so by adding material and "smoothing" the airflow it might result in less restrictive flow.

also, the 91-95 and probably later cams as well the timing of the centerline is altered to favour higher rpm (i think it's retarder something like 6 deg but can't recall exactly).

Saying that you loose VE with smaller ports and valves as a stand-alone fact is a loaded statement and not entirely true.  The standard law of fluid flow properties states that a well shaped venturi can dramatically reduce the cross section while not impeding volume of flow, only speeds it up.  A port designed for the best overall power has way more to do with it's shape and contour than just cross section and valve size.  As a test-bed, I have been using a small, aircooled, horizontal 177cc Honda/Kawasaki clone engine and found that even reducing the port cross section by up to 30% with a well-tapered run has shown absolutely NO reduction in peak power (actually a small increase), but a dramatic increase in low-midrange.

It has also been shown over the last few years that flowbench data is very one-dimensional and not the end-all be-all measurement for how much power an engine can make, only theoretical peak numbers.  VE/power output at all RPMs has far more to do with port velocity, which is altered by cam timing, ramp rate, lift, the taper of the port, length of port, plenum volume,  etc, etc. 

What I am saying is that with this engine, running stock displacement, and stock cam, I have found the major restrictions to be: valve lift (too small/restrictive) and port shape/size (too large/slow/turbulant).  The valves themselves are plenty large enough for this engine displacement and rpm.  I believe with my setup that I have been able to not only enhance peak flow by streamlining the port, but also increased port velocity, which has a very noticeable effect on overall power.  This info is not meant for building a "race engine" tuned for a peak hp output in an RPM range that is useless in a Jeep.  Nearly everyone seeking more power from this engine references not having to shift down on hills, which indicates they want better throttle response and low-end torque.

If you could get this engine to have a better valve opening rate (ramp rate) or larger displacement, then these ports would become more effective.  Sadly, the ramp rate is highly inhibited by the poor/heavy/slow valve train design, and there is no reasonable stroker kit for this engine to increase displacement.

One thing I have found that I can say for absolute certain: Changing the port cross section back and forth several times is detrimental to overall power.  One of the ways I see people do this is when they "gasket match" their intake to their head by opening both up.  There needs to be a linear progression from the large inlet (at plenum), to the tightest choke point (at the down-turn), and linear back open with no wavy variations in between.  The 2.5 intake itself has virtually no taper through the runners, but they are smaller than the port on the head itself.  This sudden opening-up, followed by a squared-off inside-radius downturn, followed by a restrictive valve lift basically gives any port velocity a triple-whammy of suppression.

I have also found that opening up the port behind the valve seat will hurt your power (found this on several different engines).  Not sure why, my guess is that it squares off the turn from the port to the combustion chamber at the valve seat and creates turbulence.  If you try to blend this ridge with the existing port on our heads, you will end up with a very poor transition from port wall to combustion chamber.  Filling the low area of the port to blend this ridge with the port instead does not create this problem.

As for cam centerlines, I can only tell you that the Melling OEM cam (that I run) from Napa is ground factory straight-up with no retard.  It is the 0.424" lift/270 grind that came on the earlier engines.  I have never seen a positive effect from retarding cam timing.  I believe there reasoning for it has to do with lowering emissions.

[sharpxmen]

if you retard the timing for the intake valve you give more time for the airflow to fill the cylinder (and will work at high rpm, not low rpm where it would actually hurt performance).
I was not disagreeing with you (if you read my second paragraph), but it is a fact that you have less VE with less intake x-section or better when you have 2 intake valves compared to 1 as a result of a larger x-section (high rpm/higher piston velocity), i didn't come up with it but just the way it is, now - if you have something that is a poor design compared with a good design (from a fluid flow perspective) then you can't really compare the two - i'm talking apples to apples comparison.

[Wrench]

I wasnt trying to start any arguments, but my findings seem to contradict a lot of "conventional information" out there.

My original build included a custom-ground cam (from some place in the Sea-Tac area) that had more duration and 0.010"  less lift.  If you compared actual intake valve events between this cam and the stock cam, the closing of the intake valve was about 8 degrees retarded from where my current stock cam is with the advance key.  The stock cam has more power at every RPM than the custom ground cam did. Conventional thinking says that the custom cam with more duration and later intake valve events should net me a noticeably better top end due to the more volume it would theoretically allow into the cylinder.  It didnt make sense to me at the time, and was the second large clue that led me in the direction I went.   It seems that retarding the intake valve events will increase top end only if the previous timing was a form of restriction in the first place.

This page has a ton of good info in there, and it has helped me progress a lot: http://www.strokerengine.com/ports.html

I have also been reading as much info from the late Professor Gordon Blair that I can possibly wrap my feeble brain around.

I will disagree with your statement that less x-section=lower VE.  If that was true, the FZR1000 from the 1990's would have more power than the R1 of today, along with dozens of other engine designs that have reduced their ports in the later years (Jeep 4.0 among them).  That just isn't the truth.  There is far more to VE than port cross section, and peak flow of a port is only one small factor.

I am also a motorcycle guy. I used to own a street-legal two stroke bike that had about a 70hp peak (highly modified NSR250).  I have also owned an Aprilia SXV550 that had the same peak power.  Both bikes weigh the same, but that Aprilia would absolutely decimate the NSR any day of the week.  Why?  Better throttle response, and more power at every rpm except the peak. Torque GALORE!  This is the direction I would like to go with my tuning on this engine.

Aw, screw it.  I'll just wait for Koenigsegg to come out with a solenoid operated valve retrofit kit for our heads.  That'll solve a whole lot of issues with this engine...

http://www.youtube.com/watch?v=Bch5B23_pu0&list=PLHa6PXrV-yIgnXSYFT07BouKhEhyFuWnf

[sharpxmen]

i don't know if i didn't explain it well maybe, I'm referring to the same cam, same intake manifold, etc. once you change the cam you can't compare 2 heads by the end result with different cams. Anyway, my point is (to make it simple) if you compare 2 pipes with same bends and lengths but different diameters (say 1'' vs 2'') and do a flow test the larger diameter flows better for the same pressure differential. Like i said, i wasn't disputing your results, just put out a comment in the discussion regarding flow with smaller (diameter) runs. If you have a pipe with a nice radius bend and one with a 90 deg angle in it with no radius the latter would def flow worse, so i agree with you on making the runs flow smooth and you might find better performance there with less diameter.

[Wrench]

My changes were incremental, leaving all other components on the engine the same.  Initial setup ran with custom cam and "ported and polished" head, swapped to stock cam (advanced as listed) and run for a while, pulled head off and added epoxy to port floor and run for a while, later pulled head, filled and reprofiled the entire ports as they are now.  Each increment was a noticeable increase in power, even at peak.  Most the power was found in the midrange, 3500 rpm.

No, I never ran the custom ground (higher duration) cam with the epoxied head.

Have you ever tested, say, a 2" pipe that has a well profiled venturi in the middle that necks down to 1"?   Was there a pressure drop after the venturi?  The laws of fluid dynamics says no, unless that flow at the venturi reaches the speed of sound.  http://en.wikipedia.org/wiki/Venturi_effect

I dont know exactly why it works but it seems that if you put the venturi at the bend of the port and force the flow around the outside of the curve by filling the floor more and making it flat, it maintains more flow through the curve than if you left it the same cross section as the rest of the port.  

[motorhedfred]

As a possible alternative to decking the block, I'd like to suggest a Sealed Power W403P piston. They're the same bore with a 1.640" compression hieght, which would put them .048" higher in the bore. The D-shaped dish is perfect for establishing an effective squish zone as long as you put it below the plug side of the chamber. You can often have more compression without increasing octane requirement with effective squish because you need less advance when it's set up right.

They do have a slightly smaller wrist pin bore at .928" compared to the stock Jeep 2.5L at .931", but that can be cured with a brake cylinder hone, a couple of block of wood and a bench vise. You'd also have to check to make sure the small end of the connecting rods don't rub the insides of the wrist pin bosses. Being as the rods are press-fit to the wrist pins, the pin bore sizing needs to be accurate as that's what controls piston rock in the cylinder bores. I've done this operation on other engines and it works fine.

MHF

[sharpxmen]

..., a couple of block of wood and a bench vise.

are you making a fire?

j/k, I think you can run the .928 pin, can't remember the exact press fit on the stock pins but i'm almost sure it's >.006 (so would still leave you with about 3 thou press fit on that), I recall something about .008 but like I said can't remember exactly.

what's the gasket thickness though (compressed)? I honestly think you'll need a + thickness one if those pistons are indeed 50 thousands over the deck.