Dynamic Compression Ratio: Solving the Mystery with Excel
 

The link to the Tool

I have been working for the last few weeks to solve the mystery of why some can run huge static CR numbers on their motors, yet get away with it without burning up their pistons. It always seemed to me that the answer wasn't in what static CR these motors were producing, but in what the actual compression of the motor is, given the fact that cam overlap can have a major effect on the net realized compression the motor sees.

Of course engine management is important in mitigating detonation at the bleeding edge, but my goal was to create a tool so that I could baseline a specific motor configuration against various cam choices and later, turbo boost levels. As long as you are comparing apples to apples with the only changing variable being cam choice and boost, you’ll get a nice picture. Editing the spreadsheet will allow the Excel experts among you do some other really interesting things, too.

This work has resulted in my being able to calculate the dynamic compression of any 911 motor, including motors with forced induction, to get the real net compression realized under operating conditions, by factoring in cam timing and boost. This allows known configurations can be compared against theoretical configurations for validation.

This project started as an effort for me to validate cam and piston choices for my motor build. It has resulted in a tool that I hope will significantly benefit the air-cooled Porsche community at large.

You’ll need Microsoft Excel to open and use it. Enjoy!

http://i25.photobucket.com/albums/c7...Cam_Tables.jpg

I'd like to thank John Dougherty and Aaron Burnham for their insights and being sounding boards for my ideas. I'd also like to give credit to the various gearheads on non-Porsche forums that posted spreadsheets for disparate pieces I ended up assembling into this final spreadsheet.

OK, where do I find the sheet?

OOH pretty. Nice job Kenik. Now you need to graph the valve closing and then resulting cylinder pressure versus cam/crank angle to see what it is really doing.

Then expand it to VE and diff. EQ for changing VE with RPM and cam related peak torque and finally boost related peak torque. I have some of these numbers for static rpms but not yet fit for dynamic prediction. Numbers are fun.


What?! only 1bar boost? Where is your sense of adventure?

btw, That is a nice 3.3 spec motor.

I found a dynamic compression ratio applet here, in Visual Basic:

http://www.empirenet.com/pkelley2/DynamicCR.html

Good discussion of what DCR is and why you should know about it as well.

Tom

Pelican dropped my embed code; added a link to the top. Thanks for the heads up.

I've been workin on these, too, as well as creating a graph to overlay compressor maps. Help is always appreciated! :D

The VB based tools are opaque - Excel allows everyone to see and manipulate the math and add to the discussion. I did use the VB tool to check my math.

OK, I get 7.67:1 with John's Supercup 102s. What do I win? What do I do next?

What I really want to model is Bruce Anderson's scenario of the Short Stroke 2,5 with 911T cams- he says in the book that T cams work the best due to the higher dynamic compression. Should be easy to get the specs from the little white book.

http://forums.pelicanparts.com/uploa...1245173799.jpg

PS isn't this an argument for throttle-by wire? I remember John's point about the old NASCAR motors that had 17:1 static compression- you could only open the throttle once the motor was above 5000 RPM where the volumetric efficiency had fallen off to the point where dynamic compression was below the detonation limits. So you had to only give it part throttle on the pit road lest the engine blow.

With throttle by wire (or whatever marketing name Bosch uses, -- blank-Tronic--) you could limit the VE of the motor down low automatically. Maybe this is how the modern engines have 13:1?

All cam specs are on the 4th tab in the spreadsheet - of course there are no 'T' specs :rolleyes:. These came from John's website. One of these days I'll add a lookup so you can auto populate the cams, rod length and head volumes from a dropdown in the spreadsheet. :eek:

Our old friend VLOOKUP. The trouble with the factory specs is the lift is unspecified.

T cam specs from Elgin's site:

Carb'ed 911T

Seat Duration (I/E) 260 254
Lift (I/E) .263 .235
Lash: STD

This whole tool is very cool, and very professionally finished.

http://forums.pelicanparts.com/support/smileys/pray.gifhttp://forums.pelicanparts.com/support/smileys/pray.gifhttp://forums.pelicanparts.com/support/smileys/pray.gif kenikh

One question -- where are you getting the "Centerlines Intake/Exhaust" numbers from?

John's site lists the centerlines. I was calculating them off of the setting at one time, but since some of Johns cams list separate centerlines for the I/E cams, I made the guess that these are accurate.

Anyone know the head volume for a 69 engine? Piston dome volume for 2.0S pistons?

Would that be in Wayne's Engine rebuilding book? Its sitting on my coffee table at home, but I know the back has lots of tables - I've only really looked at the line for 930/04 (my engine) and never that closely at internal specs.

If so, I'd be happy to post later.

2 liter head is 75cc, IIRC. Just back out the dome volume until you get the right CR.

John 2,0 heads are 72cc. Dome volume for S pistons is a calculated 39.5cc.

That's it! :D

I probably just don't understand, so please be patient with me. Doesn't detonation occur because the cylinder pressure gets too high? Isn't dynamic compression ratio just a correction for for the valves being open for part of the compression stroke?

If the above is true then I don't thnk you can determine how high the peak pressure is with the "dynamic" compression ratio. Race cams for use at high RPM would appear to have a low dynamic compression ratio due to the valves being open for more time during the compression stroke. These cams however can allow high volumetric efficiency at high RPM due to all the things we haven't considered here. These would include intake and exhaust tuning primarily. If the engine is capable of a VE of 1.0 at 6000 RPM, then it is getting the effect of full compression ratio at that RPM, not the "dynamic" compression ratio calculated here.

What am I missing?

-Andy

Andy's right, without a full dynamic pressure model of the engine system you're still going to be missing a level of information.

http://images.absoluteastronomy.com/...sure_graph.gif

That being said, this model definitely is a step-up from where most of us were before. So is it perfect? No. Is it better then most of us were? Absolutely!

Absolutely right. This is what Cambgrinder and I were chatting about and a prolem to which I don't yet have an answer. My best guess is that it is a level of detail far beyond th eneeds of most, necessary only if you need to get numbers dead nuts for engineering purposes. Who else needs to know every pulse in every cycle along the entire rev range?

But you can use this tool to generate approximations, as the pressure and compression averages that result across discrete engine configurations can be assumed to similar.

That's my take, at least :p

Andy,

Of course you are right-- and cylinder pressure is determined with instrumentation. The value of this exercise is that it sets a benchmark for comparison among various compression ratios and camshafts, which helps when you're trying to figure out what to build.

For example, consider the 2,5 short stroke example I posted above. The low-compression 2,7 pistons result in a fairly low compression motor which is complemented by the early intake valve closing point and low overlap of the 911T cams. With the calculator, now we have a way to mix and match camshaft profile with static compression.

How about an engine that used 911T pistons, an offset bored small end bushing and a set of low-lift, low-overlap cams for a high dynamic compression ratio? The possibilities are interesting and this is another tool to compare them. Which is why Kenik set it up to compare multiple configurations-- and it's easy enough for the seasoned excel hacker to replicate the bore, stroke and rod information for each assumed case below.

Of course you have to know the VE to go all the way, but that requires another layer of analysis.

Detonation occurs when the intake charge becomes unstable due to excess energy and turbulence pre-igniting the mixture. Compression only has an indirect relationship to detonation since it can be mitigated using methods like water/methanol injection or intercooling and nitrous oxide injection which cool the intake charge.

You need to cool the charge more to manage the heat higher compression imparts, thus the relationship between compression and detonation. Make sense?

You are correct on your assessments of VE, though. VEs do play a factor, but again this spreadsheet is to get close, not perfect. We are still working on “perfect”. :D

In case folks are wondering, this where the spreadsheet gets fun.

The Supercup 102s on 9.5:1 pistons result in a pretty low DCR...or do they? Let's compare to a common motor we all have confidence in, say a 9.5:1 static CR, DC40 cammed engine, which, BTW, has a DCR of 7.23:1. Whoa!

You immediately see that the SuperCup 102s make a half point more dynamic CR than the DC40s, so you should actually drop the Supercup 102 motor to a static CR of 8.8:1 to get the same dynamic CR as the motor w/ DC40s.

You soon find that by matching the dynamic CR of an unknown config to a known config, you can draw a compression correlation between the two in terms of what the ignition requirements will be.

Are we having fun yet?

And we will always be far from perfect. ;)

I see this as an extremely useful tool. One thing to note is related to a fixed engine displacement/static compression ratio while changing the cam configuration. I found it useful to compare the offset of increasing duration and lift while offsetting the compression loss by widening the lobe centers. This would allow a relatively low compression motor to breath better and have high rpm torque.

I am sure that air density gradients, mass flow and ultimately air column momentum can be figured as part of the exercise. There just might be a few pieces of the puzzle missing before going that far.

This is good stuff!

To look more closely at John's question, here is a comparison of all OEM Porsche cams that were used for 2 liter applications, stock or race, on a 911S spec motor:

http://i25.photobucket.com/albums/c7...EM_2L_cams.jpg

Interesting, no? What would be more interesting would be to alter the engine config to the cam and see if static CR normalizes at a constant, by configuration.

Here is the data for dynamic CR for multiple engine configs, by cam:

http://i25.photobucket.com/albums/c7...f_positive.jpg

Most interesting is the fact that the 10.3:1 static CR motors needed twin plugs, the rest, did not. This is an interesting kink in the theory that you can draw these comparisons.

My only guess is that the 100% VE of the race motors at high RPM results in extreme cylinder filling, leading to much higher pressures at high RPM. This only reinforces my desire to get VE data integrated into this model to map the entire rev range.

Not so simple. Ugh.

I'm not sure I agree with this conclusion even though I agree with the math (BTW I used an 81 bore for my own calculation).

Across the board the engines with high-overlap racing cams have lower static compression ratios. So are you proposing that one could increase the static CR of a 906-cammed motor to something like 14:1 (I'm just making that number up, I already closed the sheet) to increase the dynamic CR to match that of the lowly 911T? It seems that that comparison doesn't take into consideration the volumetric efficiency of the 906 cam (totally disregarding ports and intake and exhaust diameters for the moment)

Also, the cool thing about the Supercup is the derivative of lift-- the modern lobe shape gets the valve to full lift earlier than a Solex and holds it there for more area under the curve, which isn't captured in the first-order duration spec.

None of which is intended to diminish the tool whatsoever, thanks for putting in the time and effort. I'll add what data I can from the little white books.

Agreed 100%, John. See above. :)

BTW, my calcualted cam timings don't match the numbers in Bruce Anderson's book, at all. When I enter his number manually, they don't make sense either.

The work continues...

Kenikh - great looking spreadsheet - well done

Some light reading for you from Bosch on engine simulation. Might help you further your spreadsheet development.

Regards
John

You're calculating dynamic compression correctly, but unfortunately dynamic compression is of limited use for calculating detonation. And calculating VE data requires making an engine sim, which is a huge undertaking. In the absence of real VE data, however, I think you might be able to make this model more useful for calculating detonation without making it impossibly complex.

One half step that might be to replace the advertised Intake Close Angle (a static value) with an "effective" intake close angle, which might be something like (ICA- RPM/100). This would attempt to approximate the fact that at high engine speed the intake closes so quickly that the air in the cylinder doesn't have time to escape ABDC before the valve closes, leading to a higher VE.

Applying this to formula to your RSR motor above, for example, would give you an "effective" compression ratio of 10.19:1 at 6000 rpm (sounds like it needs twin plugs). Compare this with the T motor at 8.57:1 at the same engine speed, which we know doesn't need twin plugs. Thus while this might not be the right formula an is imperfect, something along these lines might give a better indication of cylinder pressures without the complexity of a full engine sim. Run a calculation like this every 1000 rpm and generate a table and you'd get some idea of both what the peak effective compression is and where each cam makes power.

Just a thought. Good luck...

This is good logic, but Jamie Novak and I are working on actually creating VE calculations using some generalizations aout Porsche 911 intake systems that can be applied toward port and intake dimensions. Nothing too crazy, but should get us closer.

The more I look into it, the more it seems that we are going to end up with a poor man's engine sim package, albeit built in Excel. :)

That would certainly be better. You'll drive it from an estimated cam profile? I've used a number of engine sims with varying success, but one hopes making something specific to the 911 motor should be both easier and more accurate.

As a side note I'm also wondering how significant a variable the intercooler is on a turbo motor- in theory it allows you to get to a given compression with a lower air temperature than normally aspirated because you can remove the heat generated during some of the compression. This may well be more than canceled out by other factors, however- perhaps your model will suggest if that's the case.

Looking forward to your next revision. Good stuff.

Intake & exhaust design and tunning can add a supper charger effect at some points and the effectively increase compression.

Also, restrictions in the intake or exaust might effectivly limmit CR. To small of ports, restrictive air filter, low lift cam...

...And in some cases can produce a pressure less then ambient, or at least less then the pressure at the open exhaust valve which can reduce torque (and by definition HP), or even cause reversion at certain RPMs.

Philip Smiths "The Scientific Design of Exhausts and Intake Systems" is the bible on that subject as far as I know. It's also quite readable. It would be interesting to see someone hook up a 911 engine to the Morrison Multi-Point Indicator that he describes in his book. Think of it as an oscilloscope for pressure at multiple points in the intake and exhaust systems.

Huge! I have the calcs figured out on this one, too. Haven't added it yet.

At what RPM is the DCR being calculated in this?

I suspect DCR is a calculated value based on instant 100% valve opening at at given time and instant 100% closing at a given time with air fill responding instantly. If the intake valve is closing at 15 deg BTDC it reduces the DCR geometrically for doing so even though at a given rpm the phasing will treat it as if it was closed past TDC.

However, most of this is way over my head.

Then we have the supper charging effect that is a result of the mass and movement of air flow that effects the true CR also.

Thus, I am guessing it would take some very hight level math to calc this. The testing the actually combustion chamber pressures or backing into it from the TQ curve might be the only way to get there.

Thus, we might add a cam that we think lowers the DCR by one point and think we can get away with bumping the CR by one point to find we get ourselves in trouble.

http://forums.pelicanparts.com/support/smileys/wat6.gif

My understanding is the DCR is the maximum compression ratio obtained by the 'ram-air' effect of the intake (or lack thereof for an N/A car) minus losses from time spent with valves open and the piston compressing the chamber volume. Is this right??

I *think* the valve closing instantaneously does not matter--I believe this "adjustment" is captured in Kenik's model for whatever cam you use, because it is taking into account the degrees of the opening. Instanteously is also not really useful because even if the cam "dropped off" there would still be some time for the valve to spring back to closed.

I think what you are trying to do is think about the cylinder pressure as a function of RPM...kind of like what I was asking about if the DCR changes as a function of the RPM or not. Either way I think you could do that with just calculus, provided you have data.

Interesting stuff to play around with but I seem to be confused with your statement. When I widen the lobe centers on any particular cam with the spreadsheet, dynamic compression goes down . My mind tells me that as overlap is reduced by widening the lobes, the affect should show dynamic compression going up, not down. What am I missing?