Hi,

The changes were +- ~0.4 AFR at different RPMs. Datalogged with an LM-1 on the same stretch of road on the same day. Made 2 runs before to see run-to-run variations (< 0.2 AFR) and after.
The old boot was pretty hard.

Regards,
Klaus

LM-1's are nice, glad to see you were using some good equip!

Maintaining the CIS is very important and your post shows exactly how important! Sounds like an inexpensive way to keep a CIS running correctly.

Thanks!

If you race the engine from idle, you should be able to see the boot flex...top of it will move down. But as the RPM increases at a steady rate (under WOT), I doubt the boot will do much to dampen any pulsation beyond what the mild cams can create.

The flow path downstream of the meter is also terribly tortuous....I really don't think the meter itself is a problem. If you really want to take the meter resistance out of the equation...re-engineer the fuel distributor so that the plunger does not have to act against the system pressure...use a booster between the fuel dist. and the injectors to get the fuel pass the injectors. The WUR would also have to be modified so in the end, you control pressure and your system pressure are very very low...but the injectors are still seeing the same DP and flow...

or put a TPS sensor on the lever..and a solenoid on the plunger...(shaking head)...

Better yet...redesign a plenum/manifold, move the throttle body to the center of a larger plenum... (shaking head again) ..wait..that's kinda like a 3.2 right :)

there are lots of things you can do, but again, all of it seems hardly worth the effort when you can just build a hotrodded engine with EFI or simple carbs. Sure I like my CIS, but somethings are just not worth the effort....glory..sure...there's plenty of glory in solving a problem that now one else has ventured to do...but by the time you are done dicking around with re-engineering a CIS engine, one could have built plenty of monsters.

Go back to the cams....that's where your going to get the most return on your effort.

Thanx, I designed them. :D
I posted this before, so those guys who read it, forgive me. But it works really well.
On the CIS Lambda you can use an LM-1 or LC-1 to make it programmable. Disconnect the WOT switch wire from the throttle body switch. This makes the Lambda computer stay in closed loop all the time. Then program the second analog out of the LM-1/LC-1 to simulate a NBO2 sensor with a switchpoint at 12.8 AFR or whatever. Use the WOT switch to drive a relay that switches the O2 sensor input of the Lambda computer between analog out1 (regular O2 sensor simulation, 14.7 AFR) and the second output. The lambda computer then regulates at WOT to the programmed AFR.

I did NOT use that when checking the boot effects :)

Regards,
Klaus

Souk,

I agree with you. Cam design is going to be the biggest power maker, but I do plan on making small changes to the plenum size (maybe a simple spacer) and improving the exhaust flow at the same time.

I honestly would have thought that there would be more interest/information on this subject!

"even managed to piss off Randy..."

- huh - well, I don't remember it... (just don't call me no engineer...)

I agree to focus on the issues and not personalities -- I'll often say something like "the poster above" rather than use a name for just that reason.

- I am curious if 'Souk' is in any way related to the famous Souk carpets.

Why are SC piston domes shaped the way they are?

So what you are saying is; CIS is slow responding . . . but dont use hotter cams because CIS is too sensitive ...too quick to respond and will pick-up the cam pusling.

. . is that right? :cool:

That does sound like limited responsiveness.:rolleyes:

I think I remember that from a Porsche tech document a long time ago (the mind is the second thing to go, and I can't remember the first):

CIS injects continuously. During ~3/4 of an engine cycle the intake valve is closed and the fuel cloud just sits there. When it gets sucked in, a lot of turbulence is needed to mix properly during the intake and compression stroke. The dome-shaped pistons increase turbulence. Specifically a CIS problem during low load conditions.

Regards,
Klaus

Shape; it's a balance between mixing(air/fuel - toroidal swirl) and burning (flame propogation)

Isl - I'm saying dont use hotter cams because the pistons used in CIS engines will likely hit if you do.

Also someone asked for a source for that -- IIRC it's in BA's book.

PS - I think it's great if somebody wants to put in the work to see how much lift/ramp angle you can do on a cam w/o having to change the pistons. It might be a poor use of time, but that has never stopped anybody else with a 911 - including me...

:) Ah. sure.

fwiw, the conventional wisdom is that CIS is to reactionary to the pulses of hotter cam (over-lap, not lift)

Klaus,
I agree that the mind is the second thing to go. I previously answered a question with a question on the domes; the mix needs the turbulence to atomize in order to burn efficiently. Everything through to the early SC cars is pointed towards the reduction of emissions...the pistons, the K-Jetronic (which has proved it's durability and simplicity in years of VW use).

The camshaft profiles and timing are made to work with the "low response" CIS and provide enough power to get by, but not much more. The big picture becomes "how to increase the volumetric efficiency and raise the compression ratio", which have the biggest impact on power production. The proposition eloquently put forth by 911sTarga seems to be "what is the limitation of the K-Jetronic?". Difficult to quantify, especially when the cam options are either A, B or C....exactly why there is a PRACTICAL limitation of the CIS system. I have pondered this myself, and said "this looks easy to screw around with", but as Souk and others with a heck of a lot more experience than me have replied, it's really not. So, it ends up a matter of atomization and cam timing.

It's why I have a set of (plan B) Webers in the garage.
Pat

For us shade-tree mechanics following along, when y'all talk about "reversion" what does that mean exactly? Does it refer to the tendency of the air sensor to oscillate due to the suction variations as the valves open and close? I'm more of an electronics guy than an automotive guy, so it makes me want to get out a o'scope and watch the air flow sensor waveform. Anybody got one they can post?

Souk, 20/21's will pass emissions here in smoggy Southern California, so I'm sure you would be fine in Chi-town.;) These are the "wildest" CIS/DME cams that I know of that still will pass emissions, but there may be other grinds out there that do as well.

My 3.5L twin-plug motor recently passed with Steve's chip, 20/21 cams AND headers (but just barely). Throwing the heat exchangers and cat back on it would pass with flying colors.

Somebody asked for 964 cam specs, this is what I have in my notes from Andial's cam doctor:

964's are timed at 1.25mm. Intake is .470" lift and 240° duration and exhaust is .430" lift and 230° duration. I don't have the lobe centers (so shoot me!) but it will be on the wider side obviously.

As a comparison, these are the 20/21 specs according to Web-Cam:

Timed between 1.9mm-2.2mm with .10mm lash

Valve Lash intake: .004
Valve Lash exhaust: .004

Valve lift intake: .485"
Valve lift exhaust: .452"

Duration: 258° intake and 246° exhaust
Duration at .050": 238° intake and 226° exhaust
Lobe Center: 113° intake and 112° exhaust

Intake opens 6° BTDC and closes 52° ABDC
Exhaust opens 45° BBDC and closes 1° ATDC

Valve timing is checked with zero valve lash at .050" of valve lift

Minimum piston-to-valve clearance is .050" intake and .080" exhaust. I used the RSR style piston for my twin-plug application rather than the CIS/DME piston so I had PLENTY of room. I think I measured something like 2.5mm for the intake and exhaust with these pistons!

Y'all go back to analyzing CIS, I'm enjoying a beer and sitting back in my chair while perusing this enjoyable thread.SmileWavy

Ralph

reversion:

Both the intake and the exhaust resemble 6 oboes (or something) next to each other so close that they interact. If one was isolated off by itself you could plot a sine wave for its pressure vs. time. Since they are near each other, one low pressure region will interact with the pressure from the tube next in the firing order (right before it, and right after it). For an exhaust system, this is usually called scavenging and for race engines, the post-WWII engine builder will try to enhance it at a favored rpm. The low pressure from one exhaust pipe helps pull the exhaust out of the one next to it. Reversion usually is used w.r.t. the intakes - on some engines a cloud of vapor particles will actually hang in the air above the intakes.

- There is also the tuning of each pipe for a given frequency...

- not exactly (for which you'd write the equations) but should be close enough - I hope it makes sense - not sure i did such a good job... wave physics sucks.

Yes. My understanding is this:

To enhance high RPM performance, it helps to have the intake valve start to open before the piston reaches TDC on the exhaust stroke to begin filling the cylinder with fuel air mixture sooner. If the intake and exhaust valves are open at the same time, this is called valve overlap. You can see that if the intake valve is open while the piston is still travelling out, this can cause a pressure reversal, or reversion, in the intake. When this momentary pressure pulse travels back up the intake, it can cause the sensor plate to 'flutter' up and down, which plays hell with the mixture and causes power loss due to the momentary lean condition induced.

Clear as mud?

ianc

The biggest limitation of CIS is the single plenum.

There is a discussion of cam selection with single plenum intake here, (on page 2): http://forums.pelicanparts.com/showthread.php?threadid=161071

Another discussion of CIS cam selections, with a chart here: http://forums.pelicanparts.com/911-engine-rebuilding-forum/96467-3-0-cis-cam-grinds-964-vs-web-20-21-a.html

Another problem is fuel delivery, if you modify too far beyond the design parameters of the engine. Past results have shown CIS to tolerate power increases fairly well, with Dane's 3.4, My and many other short stroke 3.2s, Noah's C3 making 200+ HP with a 2.7 fuel distributor, a few aftermarket turbo kits, among others.

The cams chosen for the CIS 911s were designed to make the engine produce decent power, nice smooth idle, acceptable fuel economy, and low emissions. It met these goals quite well, especially if you consider the competition when they were built. Porsche got 180 HP out of a 3 litre engine, the early '80s Corvette only had 190, and that was from 5.7 litres.

With these design parameters in mind, the guys at Porsche sat down with the guys from Bosch, and put together a fuel injection system to meet these needs. Porsche, like every German car builder at the time used whatever the current Bosch fuel injection system was. CIS can provide as much or as little fuel the engine designers required, to meet the expected needs of the engine.

If you want to stick with CIS, make more power, and drive the car on the street, you are really limited on what you can do.

Displacement: for a 3.0, 98mm pistons/cylinders = 3.2L, add a 84-89 crank = 3.4L
Slightly hotter cam: conventional wisdom is Web 20/21 is as good as it gets. If you go a lot hotter, low speed drivablilty will suffer (see the discussion in the first link above).
Higher compression: depending on fuel availability, this may require twin plugs.
Better exhaust: If you want heat, SSIs are pretty much it.


rdanes's 3.4 CIS is a great example of how far you can go with CIS if you are willing to spend the cash. He got ~220 at the rear wheels. His engine is also for sale: http://forums.pelicanparts.com/porsche-911-used-parts-sale-wanted/226319-fs-3-4-twin-plug.html

My SC makes ~175 HP at the wheels, with 964 cams, 9.8:1 98mm pistons and SSIs, with CIS. My engine has the smaller late intake runners, which hurt top end, but boosted my torque to 186ft lbs @ 4200. My car has as much torque at the wheels as a Motronic 3.2 has at the crank. My dyno results and discussion are here:

http://forums.pelicanparts.com/porsche-911-technical-forum/158331-dyno-results.html

Another discussions of hot-rodding CIS engines:

http://forums.pelicanparts.com/porsche-911-technical-forum/113153-cis-gurus-help-me-safely-inject-my-hot-rod-3-0-a.html

Don't neglect your ports.

Tom

Jay, there is a lot of interest on the subject, but I think prudent action is to live with the system (those that understand and accept it) or go to carbs or EFI. And if one is going to go to the large effort (time and $$) one might as well build a better engine from all angles. Cost benefit is a popular factor on this board, IMO. If you can make a cam with more duration without stretching the overlap too far, you'll have the next best thing for a CIS/DME engine! Like I said...concentrate on that.

I would graph the 964 (or 20/21) cam profile on a chart with the piston to cylinder clearing on that same chart, that'll give you a good visual and a nice graphical tool to plan your work. I'm not a cam guy and I haven't the time to play with it, but I would really like to see a graph like I described for the various cam profiles.

Randy...Souk is all things and everywhere :D ;)

Ralph...thanks for the post on the 20/21. I think I'm going to have a 964 cam set for sale very soon :) I'm actually working a deal to build a better engine than the one I set out to do..one with no EPA limits..just a cash limit, but I promise, it'll be a hell of an engine in NA form... this is the stuff I LOVE!

Interesting thread.
The 2.7 cis pistons are close with SC cams. I believe the 964 cams are get too close on the exhaust side. I usually tell customers they will require piston valve pocket machining.
911 Rocker arm ratio modifications are very very expensive. I have seen a few people try it, but the results have been minimal.
Adding duration to the exhaust side of the camshaft will decrease exhaust port velocity, and add to the overlap problem. Even though the exhaust port only flows 75-80% of the intake port, the port size is big. I have only seen 2 "single pattern" cams work, one is for a turbo the other a factory racing cam.
I think by increasing the compression ratio, moving the intake closing and exhaust opening points, and keeping the overlap to a minimum a good HP increase can be had. I made a post a couple of months back with a few cam timing charts attached, for some reason I couldnt find it this morning.
Valve lift... Most 911 heads flow past .400" to .450" valve lift. After .450" the valve is no longer the restriction in the port. Some prople will say its a waste of time lifting the valve further than .450". I believe the opposite. During the intake cycle, the point where the connecting rod and crank are at a 90 degree angle, is where maximum piston velocity is seen. It is usually 75 to 80 degrees atdc (after top dead center )(depending on crank stroke and rod lengths). The pressure differential between the cylinder and atmosphere is highest at this point during the intake stroke. But the valve has to wait until 112 degrees ATDC to achieve full valve lift. By opening the valve to .490" at 112, the valve will be open further at 75 to 80 degree ATDC.
Of course if you find severe turbulence during your flow test and airflow decreases after a certain point as you open the valve, you will have to restrict the total lift or fix the port.

Excellent! Now we have John in the discussion!

Here is a bit of information that I've dug up so far. A big thanks goes out to John and his website for most of this!

These numbers clearly show that a cam with a wider center line is the "key" to CIS camshaft design. What I really need know are the opening and closing events. That way we can tell overlap and get a good idea on when these valves are starting to move.

Anyway, I think we're starting to get somewhere!

C/L –
CIS = 110
S = 98
SC = 113
964 = 113
Web 20/21 = 113

Duration (@ .050”) –
CIS = 220°(I) / 200°(E)
S = 263°(I) / 235°(E)
SC = 228°(I) / 218°(E)
964 = 238°(I) / 226°(E)
Web 20/21 = 238°(I) / 226°(E)

Lift
CIS = .405"(I) / .350"(E)
S = .455" (I) / .395"(E)
SC = .450"(I) / .395"(E)
964 = .470"(I) / .430"(E)
Web 20/21 = .485"(I) / .452"(E)

Information like this should be stored in a database on this site somewhere. I honestly think this thread will help people understand cam choice with the CIS by the time we're done. I also hope it'll show what other modifications can/can't be made to make a CIS equipped engine really run (without hearing "CIS sucks, just trash it and buy carbs")


Also, here is the lonk to that thread John.
Carrera 3.0L Race Engine Questions

By the way John, I like the cam you named "Custom #2" in that thread. #1 might be a bit easier on the CIS though. I wouldn't mind trying both designs out, but man those opening and closing points sure have shifted around a bit. :D

Thanks!

Jay

Here is a simple timing card I made in Excel.
Timing Card
The yellow fields can be changed and the file does the math.

That is a GREAT file! Thanks!

Would level of compression would you suggest with the custom cams you described in the other noted thread? (2.7L narrow body) I'm thinking somewhere around 9:1 to 9.5:1?

Much of it is, as compiled by John Luetjen, in one of the links I posted above.

Here it is again:

http://forums.pelicanparts.com/showthread.php?s=&threadid=96467

Tom

Compression ratio should be the last thing chosen. It is based on fuel octane, volumetric effciency, engine loading etc.
Remember, nothing in the cylinder is compressed until the intake valve is closed. As you can see by the timing card, the valve isnt closing until 50 plus degrees after bottom dead center. You lose 30% of your swept volume. A typical 10-1 static compression engine will have 7-1 dynamic.

uhmmmm, last I remember, air is a mass, and subject to inertia. (read: lots of matter in motion on the way in . .. reversing that direction takes time ... why we (advance) time these things)

I understand clyinder bleed and the effects of overlap. That is why I asked about a recommended compression ratio with regards to those cams you engineered. Your statement actually backs up my question about a recommended compression ratio too.

The most important valve event and the most sensitive to the induction system used on ANY engine is the intake valve closing. The more fuel/air mixture that can be forced into the cylinder, the higher the performance will be. So IVC is normally delayed until well into the COMPRESSION stroke. But if IVC is delayed too far, the building pressure in the cylinder due to the piston upswing will exceed the induction systems ability (through pressure waves and gas molecule momentum) to hold back the pressure and fuel/air will flow back out of the cylinder (reversion).

As with the exhaust, a pressure wave will be generated in the intake as well. In this case, an expansion wave is generated, although will less amplitude than the exhaust pressure wave. The strength of this wave will be determined by the amount of suction that can be created in the cylinder resulting from the piston downswing and the exhaust scavenging wave.

When the expansion wave reaches the end of the intake runners (the plenum in the CIS), it is reflected back as a compression wave. By the time this wave reaches the cylinder, the intake valve is closed and the wave bounces back out. This wave continues to oscillate in the intake system until the next time the intake valve opens (this is what plays havic on the sensor plate).

Since the length of the intake runners are typically significantly shorter than the exhaust headers, the frequency of the pressure wave is considerably higher – usually two to three times higher – so by the time IVO occurs, the wave has bounced back and forth several times.

As with headers, the intake system must be tuned for a particular RPM to deliver the most benefit from this pressure wave oscillation. The air horns on some induction systems (Webers, TWM, Kinsler) are designed to spread the reflection wave so that it will provide benefit over a broader RPM range.

The area most cam companies error on is the exhaust side. Simply put, on an N/A motor the intake aircharge is not assisted. (leaving wave dynamics of the aircharge out for a moment).
After the combustion stroke there is tremendous pressure in the cylinder. As soon as the exhaust valve cracks open it flows a LOT of air. It's basically boosted out of the cylinder if you want to look at it like that.

Having the exhaust valve open too early not only costs heat (power) and velocity through the exhaust runners, it also empties the cylinder before the intake valve is open enough to take advantage of the pressure differential. (in a limited overlap camshaft this is especially true). This causes exhaust reversion and is one of the key factors in surging problems. When the airflow reverses course, it is loosing a lot of it's inertia. Typically this is overcome before peak torque however. So only low-speed issues are present.

This is why I wanted to discuss this "CIS vs camshaft design" in detail. It's nice to see people who have the knowledge/hands-on experience jumping in and giving their suggestions/knowledge. I'm hoping we can design a few cams/CIS mods to allow us to finally make some power.

I'll bet you guys can improve on the ramp profiles that P AG generated in the mid-70s for these cars too. This would be true even if there was no cheap way to improve the lift... again, whether it is worth it or not....

It should have read...
Remember, nothing in the cylinder is compressed by the piston until the intake valve is closed.

It would be interesting to see how much pressure is in a cylinder right before the valve closes. I would think at peak torque on an engine with 100 % VE you might see some pressure.
Not too many N/A engines are seeing 100 % VE.

I agree with all of 911sTarga's last post. Lots of good info there.


I try not to recommend compression ratios for camshafts as an absolute until I know more about the engine combination. Someone can read it on the board, order pistons and then find out the previously quoted ratios were with twin plugs and race fuel etc.


Most 911 cams are usually 10 to 20 degrees shorter on the exhaust side. I think you will find carb engines like a late exhaust closing and cis and motronic engines prefer an earlier closing. Actually I should say Weber carbs like the late closing. PMO Carbs are somewhere in between. You should do some airflow and velocity testing at low lifts. I think piston shapes and the airflow at low lifts can be manipulated in the right direction to help increase the scavenging effect without increasing the overlap.

FWIW... VE = volumetric efficiency, right?

Correct, Randy. It can be 100% or even greater, but very difficult on an N/A car, as John D says. I found a short practical discussion of VE in this link

http://www.installuniversity.com/install_university/installu_articles/volumetric_efficiency/ve_computation_9.012000.htm

Pat

Thank you very much for the kind words. It means a ton, especially coming from someone like yourself!


I think I know where you're coming from with the CIS camshaft theory. Have you tested any different profiles on the CIS yet? We might have to talk about doing a CIS cam project together. Since Babe (Erson) passed away a fews years back, I don't have a specific shop that I use anymore. I know his old asst. Steve (Long) started his own deal afyer Babe passed, but I haven't tried his personal work yet.

Patkeefe,

That is an excellent link for VE info! I was just about write "way too much info" until I clicked the link. Saved me a ton of writing!

Halfway down the page here:
Combustion pressure graph
there is a graph of combustion pressure. I like this graph because it is taken from pressure sensors, not computer modeling software.

I have done a few "one off" cams for cis engines. I made one cam with the Super cup intake lobe and 964 exhaust lobe on 115 centers. The customer is still in the building process, I think he has been down a year or more now. IF I remember right he was builing a cis 3.0 into a 3.2 with 10.3-1 compression. My idea was to bleed off a little cylinder pressure with the Cup intake lobe and widen the lobe centers to 115 to help reduce the overlap.
I think most companies shy away from changing lobe centers and mixing cam lobes etc. But I have no problem doing it.

The most aggressive I use for the 2.7 and 3.0 (8.5-1) litre engines is the 964 profile and the 3.0 (9.3-1) and 3.2 litre the DC20 profile.

Hi,

Just an add-on to that VE link (sorry, may be a little OT).
If you know the fuel mass flow rate (inj. duty cycle and rated inj. flow at 100% at specified or measured fuel pressure for EFI) and you know AFR, you can also calculate mass air flow and consequently VE. (air_mass_flow = fuel_mass_flow * AFR)

On a CIS it would be very interresting to characterize injector flow rate vs. injector pressure. From logging that pressure, AFR and IAT you could then also calculate VE without a characterized MAF sensor and it's associated laminar flow requirements.

Regards,
Klaus

-- I like this graph because it is taken from pressure sensors, not computer modeling software.

As Billy once said:

"When you measure, you know ... something."
- William Thompson (Lord Kelvin)

for whom the absolute temperature scale is named.

John,

I would be very interested in any updates you hear on that engine. I'd love to know how the CIS reacts to the cam/compression.

I am going to stay away from a larger engine for now and practice these ideas on the smaller 2.7L. These are some of my thoughts so far. All input or ideas are appreciated.

*Custom ground cam that is close to the "964" or "20/21" profile, but with 111° centers.

*Custom intake plenum or build a spacer for the air door. My main goal is to increase the plenum size to help avoid reversion, but also keep it small enough to save low speed operation. I can easily make one from Carbon or 6061. 6061 might be better, as CF might create added harmonics.

*Have a custom set of pistons made by some friends at Arias. I really need to do some flow testing on the head and the CIS as a unit before we get this far though. Compression ratio will be close to 9.3:1 (static).

*Custom designed header system using a different collector system I designed back in 1999. Worked awesome on the inlines, but I'll have to rethink it with the firing order of the Porsche. (162435 vs 153624)

*Plus the engine will have the obvious upgrades to the heads and such, but mods will be based off of flow data and not past Porsche thinking.

My main goal is to concentrate on combustion efficiency, I/E port velocity, and I/E port scavenging. I know that this idea and testing seems dumb to most, but I'm luckly enough to have friends that'll help me out and/or the ability to make the prototype stuff myself.

If I am wrong, then I am wrong. If I am correct.....oh boy! Just think, we might be able to use a hotter cam on a stock CIS by just adding a little volume to the plenum. Bump compression and/or change the header design and now we can do even more. The limiting factor (besides absolute airflow) for a performance based CIS engine is reversion. If we fight reversion using other methods then just cam overlap.....anyway, we'll see.

I'm wide open for suggestions on this guys and very serious about this testing. My first step might just be a new plenum design. Flow test it against stock and then dyno test it without any other changes. Then change to a 964 cam with a little tighter lobes and see what happens.

Who wants to help attempt to figure this out? Right or wrong, just toss out some idea or theory you have and we'll discuss it here.

Thanks!

Jay

I think your first move is to get out the STP can and use some of the "Racer's Edge" (modelling clay) to see how much lift can be tolerated at what degrees...

Everything else will be subject to that I think...

Im a little late to the game, but I have the Bosch Kjet manual if anyone is interested. PM me;)

http://forums.pelicanparts.com/uploa...1121369695.jpg