That Dyno chart shows gobs more area under the curve for the 911s cam. I know this only shows full throttle performance but I'd have a hard time giving up all that wot grunt! Are you claiming that the k45 cams make more torque at part throttle and that is what we can't see based on the Dyno chart?
I wonder if Mr. knight will weigh in.

Dan,
Are you also running a lightened flywheel?

Do you have full cam specs for both cams? The most single important cam event is 'Intake Close Event' (IC) you need to fully understand the importance of this event.

If a cam closes the intake valve later then it means that cam gives up a portion of the compression stroke and dynamic compression is reduced! You must understand this concept and it's consequences. This is why you can NOT decide your static compression ratio until a cam is chosen, pro builders build an engine around the specs of the cam.

So, I'm very interested to know when the intake valve closes for both these cams?

My notes have the 911s cam as these specs
Porsche 911S
Duration at 0.050" 267/235
Lift E/I .459"/.396"
LCA 97

This cam hits it's IC close point at about 50deg ABDC of compression stroke. It gives up 50deg of the compression stroke before compression can start.

This 50deg IC is very similar to that of a 964 cam and we know that using a 964 cam in a single plug engine needs about 10.5:1 compression.

To help visualize the cam events this tool is super helpful! It really helps you understand the relationship of all 4 valve events. Intake open and close, exhaust open and close. The intake valve is not close at the start of compression stroke, it's still open for 50 deg of the compression stroke and that stroke only has 180 deg available.

http://forums.pelicanparts.com/uploa...1726674069.jpg

Enjoy!

I just read this that you wrote:

The lobe separation of the 911S cam is 98 degrees. The lobe separation of the K45 cam is 105 degrees. The valve duration and lift are very similar between the two cams.

So if the K45 LCA is 105 and the duration is about the same as the 911S it means you further delayed the Intake Close event by 12 more degrees! You now are giving up 62deg of compression stroke vs the 50deg with the 911s cam. The K45 cam needs more static compression to achieve the same dynamic compression with the 911s cam. This is likely why you are down on power with the K45 cam compared to the 911s cam.

Further complicating this discussion is that the 105LCA vs the 98LCA causes valve overlap to be reduced at TDC, if the LCA goes up and duration stays the same then overlap goes down. Reducing overlap helps idle quality and low RPM drivability. So I'm no surprised it idles better and takes off better at low RPMs with the K45 cam since it has more LCA.

The missing piece here is what the static compression ratio of this motor? That K45 cam likely needs 10.5:1 to 10.8:1 compression for very best results :)

Best book ever written on cams, octane and compression ratio is this book!
https://www.amazon.com/Tuning-Engine-Definitive-Performance-Economy/dp/1859606202

Enjoy.

You can get this nifty free excel sheet here
https://www.tomei-p.co.jp/tool/valve_timing_graph_e.xls
Just enter the cam specs of your choice.
I create multiple tabs from the default tab, that way I can see each cam in it's own tab. Just play with LCA number and you'll see what that does to cam events.

Photos of the Race Factory's fabrication work connecting the Knight headers to my existing Bursch sport muffler:

http://forums.pelicanparts.com/uploa...1726705217.jpg

http://forums.pelicanparts.com/uploa...1726705217.jpg

Not to be a Debbie Downer on this great thread, but that header contacting the oil cooler is a problem waiting to happen. You have options, but I would never run that setup as is.

I have not mentioned this, but pushing 230 Wheel HP with a 2.8L engine is really amazing. Nice work so far.

The ITBs you are using, can they be made longer in length? Can you get a longer horn to extend the length of the ITB? I ask because the overall length of the ITB helps dictate the peak torque sweet spot. Longer ITB length lowers the peak torque sweet spot.

No, I mentioned 911S cam has combined LCA = 98 deg (96 & 100 - combined 98).

CR = 9.8 for this motor. I patterned after the 67-73 911S specs (Excellence Magazine) of 9.8-9.9.

So I talked with William Knight and thought hard of all of the various options for modifying the right side header to allow for 1/4 inch clearance under oil return line on the oil cooler.

Given that I had just gotten custom fab work completed to connect the Knight headers to my muffler, I did not want to make changes that would move the collector from its current location. So I started thinking about heating up the #6 primary tube and deforming it from a circle to an ellipse to get the clearance I was looking for.

I did a little math to look at the cross sectional area of the current tube compared to the cross sectional area of the flattened tube (approximated by an ellipse so I could easily do the math).

The OD of my header primary tubes are 1.625 inch with 1/16 inch (.0625 in) side walls, so the ID of the primary tubes is 1.50 inch. The radius is then 0.75 inch, and the cross sectional area is pi x r^2 = 1.767 inch^2.

If I deform the circle into an ellipse with a minor axis (a) of 0.625 inch and a major axis (b) of 0.875 inch (I am squishing the diameter from 1.50 in to 1.25 inch to get 0.25 inch clearance, so the radius goes from 0.75 inch to 0.625 inch in the vertical axis (minor) and 0.875 inch in the horizontal axis (major)), the cross sectional area is pi x a x b which is 1.717 inch^2.

So heating up and pounding down the #6 primary tube to get 1/4 inch clearance under the oil cooler will reduce to the cross sectional area of the deformed regions by 2.77% and dyno data from the Engine Masters show on TV has shown that there are negligible effects to torque and HP, so I decided that was the corrective action I was going to take.

I bought a 3 ft section of 1/8 inch thick, 3 inch aluminum angle stock and fabricated a fixture to hold the primary tube head flanges to prevent them from moving out of alignment with each other.

http://forums.pelicanparts.com/uploa...1726887068.jpg

I had clearly marked on the #6 primary tube the two areas that needed to be worked on.

http://forums.pelicanparts.com/uploa...1726887068.jpg

I worked on one section at a time. Using a MAP gas torch I heated up the area to be deformed, then laid a piece of thick wall gas pipe on the heated area and struck the gas pipe with a single jack sledge hammer. I repeated the heating and pounding process until I had achieved the desired amount of deformation. Then I repeated the process on the second section.

http://forums.pelicanparts.com/uploa...1726887658.jpg

I let the header and fixture cool back down to room temperature, then removed the fixture and reinstalled the right side header and reconnected the headers to the muffler.

I left the insulating wrap on the oil return line and verified that I now had 1/4 inch clearance between the #6 primary tube and the oil return line.

http://forums.pelicanparts.com/uploa...1726887794.jpg

http://forums.pelicanparts.com/uploa...1726887794.jpg

I'm sure you will be ok with this fix. Thanks again for such a detailed thread.

The headers just needed to drop down another 3/8", so the down tubes should have been slightly longer :( but they do look well made.

Man I hope William made it right with you on that. I know those headers were not cheap and should fit better. I am glad you got it fixed to your satisfaction and you should be up and running

Final Dyno Results
 

I just got back from the dyno shop. Here are the test results of my rebuilt 2.8L engine with K45 cams and 1.625" Knight headers versus 1.625" Bursch headers.

Gary was able to tune out the 3500 rpm to 4000 rpm torque dip we had with the Bursch headers by adding fuel to the VE table and 1-1.5 degrees of advance to the ignition timing table with the Knight headers.

With the Knight headers, we were also able to recover half of the max wheel torque we lost going from the 911S cams to the K45 cams.

All that and I ended up with a car that is MUCH MORE enjoyable to drive than when I was running with the 911S cams.

http://forums.pelicanparts.com/uploa...1727233293.jpg

Now, if you are like me and having a hard time keeping all these dyno results straight in your head, see the following summary table:

http://forums.pelicanparts.com/uploa...1727233406.jpg

Enjoy.

Previous Dyno Results
 

For a recap, here are the previous dyno results:

Stock 2.7 engine without smog, ITB EFI next to CIS

http://forums.pelicanparts.com/uploa...1727241236.jpg

2.8 engine with 911S cam / 1.750" Bursch headers vs stock 2.7 w/o smog and ITB EFI

http://forums.pelicanparts.com/uploa...1727241236.jpg

2.8 engine w/ 911S cam / 1.750" Bursch headers vs 2.8 engine w/ 911S cam / 1.625" Bursch headers vs Stock 2.7 engine w/ ITB EFI

http://forums.pelicanparts.com/uploa...1727241236.jpg

2.8 engine w/ 911S cam / 1.625" Bursch headers vs 2.8 engine w/ K45 cam / 1.625" Bursch headers

http://forums.pelicanparts.com/uploa...1727241236.jpg

2.8 engine w/ K45 cam / 1.625" Bursch headers vs 2.8 engine w/ K45 cam / 1.625" Knight headers

http://forums.pelicanparts.com/uploa...1727241236.jpg

Dyno results summary table

http://forums.pelicanparts.com/uploa...1727241236.jpg

Enjoy

Excellent Work great data for the community. William

The area under the curve (average power) looks much better with Williams headers/cams combo!

Very grateful that you have taken the time to carefully share all your testing with us!

excellent
 

Dan,
Nice summary of your results. As i followed your posts, I appreciated the blend of data and personal driving results/comments, illustrating that the highest dyno numbers do not always translate into the best driving experience......congrats to for going down this particular slippery slope.

regards,
al

Nice work getting rid of that 3700rpm dip! A lot of us with ITBs have experienced this dip. In my case, i was able to see the root cause was the timing dropping below my target. Played with some noise cancelling settings for the signal from the cam/crank pickup and then added a few more degrees timing and got it to flatten out.
I think there is some inherent resonance with the crankshaft or something at that RPM.

Table of Contents
 

This forum thread is long. Below is a table of contents to help you navigate thru it.



CIS to ITB EFI conversion: pg1 post 1-20, pg2 post 21-27

Brake conversion / weight reduction: pg2 post 28-30

Wiring improvements: pg3 post 41-42

Crankcase tear down: pg4 post 62-66

Heads tear down: pg4 post 68

Head stud removal: pg4 post 69

Engine wiring harness: pg4 post 72, 76

Parts cleaning / refinishing: pg5 post 81, 90, 96-99, pg6 post 100-105

Distributor modifications: pg6 post 106-107

Engine compartment seals: pg6 post 109-110

Engine lid seal: pg6 post 115

Rod and chain holder tools: pg7 post 126

Convert to moisture resistant connectors: pg7 post 129-132

Ground connections: pg7 post 136

Updated wiring diagrams: pg7 post 139-140

Piston weight: pg8 post 141

Engine air baffle modification: pg8 post 142

Updated sprocket supports: pg8 post 146-147

Piston / cylinder assembly: pg8 post 148-150

Fan, fan housing, alternator assembly: pg8 post 151-152

Deck height and compression ratio: pg8 post 153

Ignition timing set up - more info: pg8 post 154-155

Rod weight and piston/cylinder assignment: pg8 post 160

Crankcase clean and assembly: pg9 post 161-178

Valve train mock up: pg9 post 179

Piston to valve clearance: pg10 post 181

Engine assembly: pg10 post 182-194

Engine / transaxle installation: pg11 post 202-207

Camshaft break-in: pg11 post 209

O2 sensor controller calibration: pg14 post 263-266

Camshaft replacement: pg14 post 267-275

Final headers / exhaust fabrication: pg14 post 276, pg15 post 287, 292

Dyno tune / test results: pg15 post 295-296



Hope you find this thread to be helpful.

Good luck with your DIY project!

Hi Dan,

I have found your thread very useful and interesting.

Could you by chance post a picture of your acceleration enrichment tables?

Thanks,
Rutager

http://forums.pelicanparts.com/uploa...1727709552.jpg

Awesome stuff, Dan!
Ive also been watching with great interest ( I'm in the researching phase for my own EFI).
Question - If you are using Alpha N ie throttle position as your main load axis, and could you also post your latetst VE table Please? Also, what fuel injectors did you end up with?
Thanks so much!
Doug

Hi Doug,

With the stock 2.7 CIS motor on the original CIS to ITB EFI conversion, the vacuum signal was good (55 kPa at idle) and I used speed density (MAP sensor) for the load monitoring signal (see pg1 & pg2 of this thread for the details including the tune maps).

The rebuilt 2.8 motor with high lift / high duration cams had poor vacuum signal (72 kPa at idle) so I had to convert to alpha-n (TPS) for the load monitoring signal.

As you can see on the tune tables below the vertical (Y) axis is labeled "xxx load % TPS".

I am using Bosch 62672 (0280155831) fuel injectors (31.4 lbs/hr @ 3 bar, 329.8 cc/min, 12 ohm). The injectors are originally for Volvo 4/5/6 cyl turbo applications.

http://forums.pelicanparts.com/uploa...1727726042.jpg

http://forums.pelicanparts.com/uploa...1727726042.jpg

Thank you so much Dan! it’s a great reference. Its interesting that your max VE is 200. Is that a typical AlphaN / ITB thing to use more of the table resolution? Still learning :)
Doug

Hi Doug,

I know the VE in VE Table stands for volumetric efficiency and a naturally aspirated engine should operate up to 100% volumetric efficiency, but in TunerStudio, the numbers in the VE table cells are really a percentage of the "Required Fuel" value you set in the "Engine and Sequential Settings" page.

http://forums.pelicanparts.com/uploa...1727733994.jpg

For my set up, I have the "Required Fuel" value set to 6.00 ms. If I use TunerStudio's Required Fuel calculator that takes into account the engine displacement and the fuel injector flow rate, the Required Fuel calculator will set the "Required Fuel" value to 9.60 ms. Which is a problem with my engine because at idle with % TPS at or near 0.0 the values in the VE table cells along the bottom two rows would be in the range of 10-20 for the AFR target (from the AFR Table) of 14.0 to 14.2. VE table cell values less than about 30 make too big a shift in the injector pulse width going from say 12 to 13 for example to hit an AFR target of 14.0. A change in the VE table cell value from 12 to 13 will shift the AFR by 0.5 or more (and you can only make whole value changes, 12 to 13, not 12 to 12.5 for example).

So, I lower the "Required Fuel" value from the calculated value of 9.6 ms to 6.0 ms which means in ALL of the cell values in the VE table need to be increased by 70% to hit the AFR target values in the AFR table. My idle / no load values are 36 so my max rpm / full load values go to 202.

Long story short: in TunerStudio, the VE table cell values are very dependent upon the "Required Fuel" setting. The cell values are only indicative of the volumetric efficiency, not a measure of the engine's actual volumetric efficiency for a given engine speed and load condition.

Thanks for the explanation Dan. I knew the VE numbers are relative rather than absolute for your Volumetric efficency, but now i see why you did what you did..I’m leaning towards using a Haltech 750 Elite and they do all that differently from what i can see in the software... Been playing with the software on every system i can to make my final decision. I’m using PMO TB’s btw. Thanks :)
Doug

Interesting information Dan,

I did notice that your fueling tables were much higher than mine and I’m less than 20 in my bottom two rows of my tables. I am using MAP as my cams are mild and I get a good signal.

So if I set my required fuel 70% less and then do a whole VE table increase of 70%, that would be basically what you did to get better resolution?

Thanks so much for your sharing of info, you clearly have a very good grasp of the software.

Rutager

Hi Rutager,

I was just trying to explain how the VE table works in concert with the Required Fuel setting. If your tune is working well for you, LEAVE IT BE.

If you change the Required Fuel setting, you basically have to re-tune the VE table as a simple % change applied to the entire VE table won’t be correct. You’ll end up either making repeated tuning runs on the street while running auto tune and recording datalogs on your laptop to ensure you're running at your target AFR's or spending $$ at the dyno shop to do the same on the dyno.

CIS to ITB EFI conversion
 

Just to chime in and add more signal… or noise…

The required fuel setting is also relative to the injectors you have and how they perform/work…

FWIW, when looking at someone’s VE table in TunerStudio, basically just ignore the cell values and look at the 3D map to see how and where they’re adding or removing fuel.

Rutager- if you’re looking for more resolution there are a few TunerStudio hacks that I was playing with to effectively turn a 16x16 table into a 16x32 table with either the x axis (rpm) or y axis (load) doubling. This is very helpful if you are running speed density or %baro and spend most of your driving time above 80kpa or 80% baro which is pretty much a given with ITBs and want more resolution to really refine your tune.

If I can get to it this week I’ll post a thread about it. I think I posted in another thread somewhere a few months ago (July?) how to do it.

Rutager,

You are running speed density (MAP sensor) so your VE table is going to be different from my current VE table as your idle / no load conditions are probably not in the bottom two rows of the VE table.

Shown above is the VE table from my car before I rebuilt the engine. I was running speed density (MAP sensor) and my idle / no load condition was 50 - 55 kPa which is in the middle of the VE table.

And as Showdown points out the values in the VE table depend on not only the Required Fuel setting, but also the injector size and performance. And when looking at someone else's VE table look for trends like fueling increases with rpm and load, etc.

The VE table above was used with the MAP sensor for load with a Required Fuel value of 8.0 and 25 lb/hr injectors.

My current VE table in post 306 is using %TPS for load with a Required Fuel value of 6.0 and 31 lb/hr injectors on an engine that is making 20% more power.

Julian,

Yes, you did post how to enlarge the table on another thread- very clever!

Dan,

Thanks for the great explanations, as always, that makes sense, I only dip down to the lowest few cells on deceleration. My car is running pretty good right now, just a few minor drivability issues when doing slight throttle corrections like in traffic.

Best,
Rutager

Thank you Dan. So much great information that is sure to help other navigate these upgrades.

The car is running good. I have driven about 2000 miles on the rebuilt engine and after retightening valve cover nuts, oil feed fittings, etc the engine is oil leak free.

I drove the car over to Don's Autowerks in Campbell, CA last week to show my work to Don Wise. Don is my Porsche go-to guy, I have known him for 20 years now and he is a friend and mentor that has always encouraged me to get in and work on my old Porsche.

Anyway, I asked Don to take the car for a test drive - I have been the only person to drive the car since the engine rebuild. After he got back, he said the engine ran smoothly, sounded good and had plenty of power which was nice to hear from a Porsche person I trust.

After all the engine work, the rear decklid gas-filled strut was getting weak - it would still hold the lid open, but when I pulled the release handle the decklid would not longer open slightly so I fashioned a tool to keep the release handle pulled while I walked around back and opened the decklid.

http://forums.pelicanparts.com/uploa...1730309067.jpg

http://forums.pelicanparts.com/uploa...1730309067.jpg

I also bought a new strut, but instead of buying the gas-filled strut, I bought the Rennline mechanical strut which will lock the decklid open but offers no spring assist to help lift the decklid. It works as advertised just fine, but again as with the weak gas-filled strut, I have to use my release handle tool to hold the release open while I walk around to the back of the car and open the decklid.

http://forums.pelicanparts.com/uploa...1730309473.jpg

After more thought, I decided to just go back to the gas-filled strut. In fact, I ordered new struts for the front trunk as well.

Copying the Table of Contacts to end of the thread.

It's not the strut that opens the lid. It's the rubber bumpers.

I talk about it in my thread: https://forums.pelicanparts.com/porsche-911-technical-forum/1126971-zoidberg-my-1978-911sc-10.html#post12270585

Thank you sharing the link to your thread, I did not know about the rubber bumper adjustments. On my car the right side bumper is stuck, but I was able to unwind the left side bumper and decklid latch released when I pulled the release handle.

I am going to replace the decklid rubber bumpers with new ones. Thank you, jrwilson!

On Pelican Parts, the rubber bumpers are called "Hood Stop Buffer"s, part number 9P1-823-499-OEM (I ordered the Porsche part, there are other OEM brands available).