combustion analysis
 

So has anybody here used this technique combustion analysis? I know its been around for a long time in the automotive research literature, and I would have guess deep inside GMs research program :eek: or the like, but not out in the 'wild' per se.

Mapping the pressure developed during the combustion process in each cylinder seems like a really brilliant way to actually get to the heart of getting the most from a engine. As the article says, BMEP is the root of all and cylinder pressure is a direct relation to that metric.

Im really curious as to 'real world' experience. Just thinking about this technique it would be the only way to actually use and set individual ignition and fuel trims that come on most high end boxes... let alone helping speed the overall tuning process.

It makes me really curious as to how much money one could save having 'non-intergrated' information data set for a given motor tune. In other words, a standard dyno run 'sees' every change averaged over the whole motor. This seems to give you a that same data, but distributed up into the pieces parts of the combustion process. Thus you can distiguish between an intake problem vs a spark problem vs an exhaust problem...

Thoughts greatly apprecieated.

t

http://forums.pelicanparts.com/uploa...1330023550.jpg
http://forums.pelicanparts.com/uploa...1330023567.jpg
http://forums.pelicanparts.com/uploa...1330023585.jpg

Would love to have a set of the "special sparkplugs" along with a datalogger to keep track of cylinder pressure vs crank angle in real time for all cylinders. That would be the ultimate in EFI fuel and spark tuning.

I have read about such systems. many times these are used to help design intake and exhaust manifolds to prevent (more like reduce when you consider the usual LOG intake) uneven air mass distribution.

Might as well load up your spark plugs with ion-sensing, high-speed thermocouples and a few other metrics while you are collecting data.

Love the test engine on the dyno.

Here is the link to the company that makes the parts. Listed in the Ed. Note.

http://isystemsperformance.com/

Very cool!

Never seen it on a gas engine but used all the time on big to very big diesel engines. We used to have a chart recorder that attached to a valve on each cylinder to make a recording of the cylinder pressures on ships I sailed on.

W

http://www.gouptoday.info/avatar1.jpgLove the test engine on the dyno.

Yeah, nice looking rig....looks like they light it up with MSD units.
I swear though, when I see a paint job like that I think MotorShysters

I worked with a similar cylinder pressure measurement and data aquisition system at my last job with Waukesha Engine.

This measurement technique is not intended to replace the dynamometer. A dyno is still required to apply a load to the running engine at a given rpm. The cylinder pressure data just compliments what is ultimately absorbed by the dyno, and reveals some other interesting and previously, difficult-to-obtain, data.

In addition to individual cylinder pressure-volume data, indicated mean effective pressure (IMEP) is readily available. As brake mean effective pressure (BMEP) is calculated from the power absorbed by the dyno, we can now calculate the frictional mean effective pressure (FMEP) lumped with the pumping mean effective pressure (PMEP) from the relation;

(IMEP) - (FMEP) - (PMEP) = (BMEP)

Now with the cylinder pressure history being known for the exhaust-intake-compression strokes, the data system can easily calculate the PMEP. With that being known, the most elusive parameter, the mechanical frictional losses (FMEP), is calculated. Thus, the combustion data logger in concert with the dynamometer, reveals accurately some parameters that were previously just best guesses.

As the article states, there are several key variables that effect individual cylinder combustion rates and peak pressures attained. And what is seen at one given load and engine speed, can be entirely different at a different speed/load condition. Some wimpy cylinders suddenly become the stronger of the lot, and the strong become weak. This is usually due to the air handling ability of each individual cylinder due to the influence of tuning effects in the inlet and exhaust manifolding. The insight into the actual cylinder pressure history thus opens up a more microscopic view of the engine process that can mean more work for the engine developer but in the end, will result in a superior-performing product.

Dave:
Nice synopsis.

I've gone through both volumes of Taylors "The Internal Combustion Engine in Theory and Practice". Both are a good read, but I enjoyed the second better just because as a chemist the fuels/combustion is a more directly comfortable zone :D. Anyone suggest some other books if similar vein?

Funny part is that I do explosives research and we do similar closed volume pressure measurements for energetics performance testing. However our stuff only goes boom once! But I can really see how this could be really useful.

What intrigues me is along JPs thought line. Could you data log while driving the car and back calculate to hp and torque numbers real time while driving?

Some of iSystems previous handy work here on pelican... Even with great colors as well :eek:. 202 HP at 7500 with muffler aint bad for a 2 liter with stock small valve heads.

http://forums.pelicanparts.com/porsche-911-used-parts-sale-wanted/566662-2-0-engine-excellent-performance.html

t

Funny. I never connected the screen name until now.

btw, did the new owner of that engine ever get it in a car in the PNW?

btw, I too have read Taylor's volumes and not found anything else that compares. Would love more reading.

FWIW,

The comment on development instrumentation was spot on. By example a system does not help with timing unless you have the means to adjust spark cylinder by cylinder. The same is true for fuel mix, we can monitor the AF ratio for each cylinder, some systems you can tune per cylinder (carbs) some EFI yes others not.

With all that said if the engine is well developed (Porsche) then changing the timing on two cylinders by 1 degree is barely going to be read on the dyno. If the AF ratio between cylinders is really off more than likely some thing else is wrong (sick injector plugged carb passage, flat cam etc).

I guess what I am saying is on a well developed engine that does not have a tuning "problem" do not expect to find BIG power by using such tools even IF you can make corrections.

PFM,

To further your comment, the same holds true with valve timing. Being able to adjust cam timing to each cylinder for optimum efficiency would be another step.

They use a spark plug like that in Aviation all the time to check cylinder pressures.
The holder of an STC has to prove to the FAA through a DER that an intercooler package is performing to what the STC states it will do to cool the engine heads, and lower peak combustion temperature and pressures.
Nice article.

I guess the question im working over is "will it work for me". I really see few levels of engine building on this board.

1. Home builds. Assemble and run.
2. Home builds/professional builds that actually go to the dyno for a 'one time' tune.
3. Club racers that do a seasonal build/tune cycle.
4. Those spending real money to go fast... high level amature/professional.

Four is obvious. They are spending what it takes. Three seems like they could advantage since they are going to the dyno anyway. Why not get the 'extra' while your there. Might even save a few pulls given the additional data. One obviously is not interested.

Two is where I'm at and is a personal question driving the post. I've been collecting parts for my 2.8 for about 5 years now (thanks to work, moving twice, chemo, ect). I'd guess I'n north of $12k but south of 20 so far for my goodies. Some dyno time is a no brainer once the thing is assembled. Question is, is using this as a 'one time' set up going to be of benfit. So far, I think the answer is a solid yes to get the most out of my collection of non-stock parts. I don't think this will personally lead me to modify the HB intake runners or tweak Herr Weiners head work, but heck, even micro-squirt has individual trims. It just seems like given all the cash spent on a hot motor, even for the street, leaving hp on the table, given that your on the dyno anyways, seems weird. Like everything else the last few are more expensive :D.

To PFMs post, although I'm a chemist, I find it hard to belive that most factory motors are build to 1% spec. I mean thats like less than three hp spread amongst cayannes on the showroom floor (275 hp IIRC). Is this really true?

Given the CNC ground cam from the 2.0L engine that sold, I would say lindys comment is spot on. As far as I have heard its got its MFI but is still sitting on the ground. I think its headed for the 69S. It was a screaming deal for a turnkey though.

t

Tadd, on your read of Taylor, that two-volume set is one of the finest works ever devoted to the understanding of the IC engine. Truly a classic, but now a bit dated. For a more current text along the same vein, I recommend the following:

INTERNAL COMBUSTION ENGINE FUNDAMENTALS by John B. Heywood, McGraw-Hill

This is a well-written book that's now used by most colleges that have an IC engine course.

Tadd,

Maybe my point was missed, on a single 4 brl carbureted V8 the air flow to cylinders is greatly affected by the intake manifold. It is known the middle 4 cylinders are well fed and the outboard 4 cylinders not so well. On the Porsche for example a stock off the shelf 3.6 manifold we tested was within 6 CFM from best to worst on well over 300 CFM, so as I said well developed. As for AF ratio if the airflow is relatively well balanced the AF should be as well (each cylinder getting equal air and fuel if everything is working correctly). So during a dyno tune we set the AF ratio that makes best power at each set point, not the same AF ratio the one that makes best power. Normally this is done with an O2 sensor at the collector(s) and moving all injectors in an adjustment. If you tune to this level then stop put an O2 sensor or load sensor as you wish in each cylinder and go to individual cylinder tuning you will not see much gain (2 to 3 HP), on the less well developed induction system tuning the mid 4 cylinders and if possible the out bound 4 cylinders (yes with cam timing if possible too) you will find more gains.

If Porsche tuned every production motor this way your comment about 1% spec would be more valid, Porsche does not tune each engine on the dyno, rather they pass smog by de-tuning the engines to meet those guidelines.

In a more direct answer to you if the opportunity to better tune your personal project is available at the dyno shop, in cylinder pressure, individual cylinder ignition and AF control and you have the budget for the dyno tune to the best you can but be prepared to de-tune that ultimate setup unless you will always drive on the dyno fuel used as that last 1% tune will be that critical.

BTW When I get to the dyno with my project you know I will use every tool at my disposal to tune all I can get.

Combustion measurements
 

Hi all,
By day, I am the Technical support engineer for AVL. We make these type sensors and more. This is indeed the only way to go if you really want to understand what is going on inside your engine. By night I build 911 race cars and used to be an Indy car mechanic in the 80's with Truesports and Bobby Rahal. I welcome any questions about these products and usage. My personal e-mail is pmaehling@provide.net or post here and I'll try to answer in a timely manner. Tommorow I can post some photos of the sensors.
All the best,
Peter

If you google Toad Hall racing, you will see that levon pentecost (bmep), one of the authors, has his current activities/knowledge based on "hands on" experience............

http://forums.pelicanparts.com/uploa...1330150149.jpg


regards,
al

PS: Jamie....I think that Kenik ended up with that fine motor featured in your link....hmmmm..........whatever happened to it?...........Kenik????

Combustion analysis
 

Hi all,
As promised I am uploading a good photo of our AVL 19 mm flat gasket spark plug sensor. These come in different thread reaches, flat and taper seat, and numerous heat range/ tip extensions as they are application specific. You can see the actual probe that goes inside the body of the spark plug next to the plug adapter.

The problem here is cost. To perform combustion analysis of this nature costs big money in anyones scale. I service the big three, race teams, universities, and everything in between. These devices are Piezoelectric which means they output a charge in picoColoumbs per force unit (Bar, PSI, etc.) This signal must be entered into a charge amplifier which converts it to a useful scaled DC signal that can then be read by your data acquisition. You also need an engine position sensor in the form of an encoder or else rather sophisticated software to interpret your crank wheel such a 58x (60-2) if you want to do any meaningful work.

So where are we?
6 sensors * $3000 each ~= $18,000
6 channels of charge amp ~= $15,000
Angle encoder ~= $1,000 to 5,000
Data Acquisition ~= $500 to 15,000

So you can see it is beyond most enthusiasts at this point. Not to discourage anyone, as I am working on assembling some of this equipment for myself from secondhand and trade in sources. Even E Bay has stuff. BUT in order to fully understand what is going on inside the engine this is the right way to go...oh, by the way as pointed out in a previous post, this is all done ON the dyno, so you need access to that as well.

If that isn't enough, I also work with an even higher end system where we perform combustion research by optical analysis of the flame intensity, but that is another story.

All the best,
Peter
http://forums.pelicanparts.com/uploa...1330179037.jpg

sub

Tadd,

So if the dyno you use has such tools to "tune" your engine be happy, if not for about $50K you can get the dyno guy these tools. Now if you are not going to move or replace your cam on the dyno or change the entire induction system then you are likely going to "tune" the AF ratio and the ignition timing. When / if you go to cylinder by cylinder tuning on a well tuned well sorted engine combination you my find at best 10 HP for a $50K investment.

Peter,

You are working with these nice tools all the time, what have you seen in a tuning only environment as gains using these in cylinder sensors? Again starting with a well tuned and sorted combination how much HP / TQ gain do you see in tuning AF and ignition timing on a cylinder by cylinder level?

If as one of the posts above mentioned you are doing development work where everything is open to change, combustion chamber shape, piston dome shape, valve size and placement, cam, induction system then these tools are of much more value.

72RSR:
That's a really awesome looking piece!

I do agree start up costs are not to be trivial, however I would guess you could sneak it in such that the data aq equipment would only double the cost of a dyno install? I had a buddy buy a new dynojet in the 2002ish time frame and that was 30-40k so another 30 to add CAS? Regardless, that's the service providers problem :D:eek:.

For us customers it's what is the 'rent' cost vs what we learn. Those pesky utils.

So on the DIY front... When I go blasting I got a pelican box with a NI Vixy box, a laptop, a signal conditioner, the transducer, and coax to string it all together. This sounds ALOT like what you describe. A recient project utilized this transducer from PCB:

Model 105C12 Spec Sheet

We get these in the $350-450 range. It's a 10-32 mounting thread, so pretty small. I haven't drawn up a part, but given our airheads use a M14-1.25 plug (if I'm remembering this right), it seems like one could fit a 10 mm plug along with the 10-32 transducer in a M14 'bung'. I bet it will be really tight, and things would need some serious placement love. One saving grace would be the fire port which would push the transducer further into the wrench section which is wider. An M10 plug is the smallest I see on the Bosch plug translation chart.

That particular transducer typically has a 1.7-1.9 micro second rise time. Plenty fast i should think given that an 8k motor is around 20 micro seconds per crank degree of rotation if I did my dimensional analysis right. I go bang only once :D, so I don't know if the decay time is fast enough, but typical rule of thumb seems to be 2-3x rise time. So that leaves us still less than 10 micro sec... Starting to get too close, but not undoable I would think. The fire port will add a functional delay depending on th length, but that's just a label correction in the data set since the pulse getting to the transducer is the previous or 2nd previous. Heck, that's recording every degree. Move to data every 2 or three degrees and there is a really comfortable margin.

I'm not sure what's in the signal conditioner box but for the 9V batteries they eat. Can't be much, but they aren't that expensive to buy outright. I'm not an EE jock. I leave that to others more twisted than I :D. Especially the good RF guys :D:D.

That leaves data aq.... Which makes me wonder about using n of these. Amazon sells them. Open source. $99. I haven't dug into it, so I don't know if it will output to a computer realtime.

Amazon.com: DSO Nano V2 - Pocket-Sized Digital Oscilloscope: Electronics

Some software via lab view and instant Combustion analysis :D. Easy peasy :D:D.

I'll think I'll rent time for now, I'd rather be driving :rolleyes:.... Or wim the lotto so I can play for a living.

PFM: Well, iSystems has this stuff on their dyno. Hence I'm trying to get at if this would be worth it to me personally. May be, may not. That's why I'm asking. It might also be useful to someone else so I posted to inform myself as well as others.


t

% gains
 

PFM,
You are on the right track with your questions. Starting with a well tuned engine, you are looking at very small gains Maybe on the order of 1 to 3%. You can optimize individual cylinders via fuel and spark with these tools, but it may be easier (read cheaper) to use such things as EGTs for fuel. I am working with old school stuff (carbs and distributor) so it is a bit tougher. I am looking to try to optimize timing by evaluating 50% mass fraction burns on individual cylinders, but I will only have one timing adjustment. High dollar NASCAR teams have been known to spend the time and money to have individual cylinders offset timing with the distributor not being equidistant between cylinders. High tech development on low tech devices.

The idea of using this as a tuning device is certainly possible, but it is more intended for combustion chamber development, alternative fuels, and major component development such as port designs affecting flow tumble and swirl.

With current digital EMS systems, we have more control of the individual cylinder events, controls and timing. So here you have more flexibility, but starting with a well tuned engine, I wouldn't look for major gains. This is like the kids with tuner cars and their laptops like you see on TV. Usually they are not starting with a well sorted system, so they can make larger gains to a point. Then they need this type of more sophisticated data and feedback else they are spinning their wheels.

I'd also like to compliment Levon, as he is one of the few to attempt this with our Porsche air cooled stuff, and it isn't easy! I encourage all of you to keep studying and moving toward this level of work. It can be rewarding.

Peter
72 RSR clone vintage racing in the Midwest and Northeast

Easy peasy ;>)
 

Tadd,
You certainly have the right ideas. I am a sensor snob as that is my specialty, so I only regard two companies as genuine players...AVL and Kistler. That said, others do make sensors. You can make adapters and such, we do it all the time. You must be careful of two factors: the indicating passages must be designed so as to not introduce Helmholtz resonances (think organ-pipe) into the data. This is often an issue and particularly with the older designs of equipment. The other is thermo shock the diaphragm of the sensor causing cyclic drift, which may or may not recover by the next firing event. This is why we use flame arrestors on the tips of the sensors to protect the diaphragms.

The lab view solutions are very possible. Many universities use it because it is cheap and they have slave labor (students) to spend months writing the code. I'd rather be playing with my car!

I have uploaded two more photos for you. One is our 10mm spark plug solution with integrated sensor as the body is too small to offset and install a separate probe. Note the diaphragm on the side of the body. The other is the finest sensor in the world bar none...and note it is only a 5mm sensor (photos are way different scales). So maybe size doesn't matter?? ;>)

Great discussions and I'm headed to the garage for the afternoon. I'll check in again tonight.

Peter
http://forums.pelicanparts.com/uploa...1330196316.jpg
http://forums.pelicanparts.com/uploa...1330196333.jpg

Combustion Analysis
 

After much coercion, threats and pleading I respond. First one comment; "a well tuned and sorted out" engine would have to be one open to much interpretation. Remember, one mans junk is another mans treasure; nothing implied. How much can be gained with combustion analysis? A very good question; if I could answer that in one sentence I could fix our economy and I know no one else that could answer it either. But, give the individual peak cylinder psi at crank angle, individual cylinder cyclical pressure variation, individual cylinder psi rate of rise, mass burn per cylinder crank angle and the p/v loop per cylinder and some of us may take a stab at answering. The thermodynamics that generate cylinder pressure is not a given and all engines do not read the books we do so making a blanket statement is most difficult. I would like to thank everyone for their interest and comments.:rolleyes:

Levon (I assume it's you)

Thanks for joining us here and for elevating the level of discussion. Just a couple questions if you don't mind.

I see you used a crank trigger on the 2,0 liter engine with the green fan. What did you discover about spark timing or spark scatter? Did your ignition control somehow compensate for timing differences among the different cylinders?

I see you used PMO tall manifolds and what appear to be 40 webers. What pressure differences did you observe in Cylinders 1, 3, 4 and 6, vs. 2 and 5, the cylinders affected by the curved intake runners? The original Solex carbs, and certain ITB setups, have a straight shot into the intake port.

Were there any other glaring issues with the baseline configuration of the 2,0 that would be a good starting point to look for improvement?

Thanks for joining us and for whatever knowledge you'd be willing to share with us.