Does quench affect power?

[LotusElise]

By looking how Renault did things from the factory (which must be taken into account due to the rules) aand what the F3 head may look like, i think there are different ways to skin a cat...

They fill it with the upper intake access port or with both at the same time ?

[ptuomov]

...BTW, I validated my model of reduction of squish height by engine speed and heating at 3 examples, which fits quite well for alu and iron block systems. All 3 had evidences of sligthly light piston-head-kissing, which is the best validation case. From there I can definitely say, elongation of block and crank assembly is not compensated to zero, there is a leading difference, coming from temperature and inertia force driven portion.

If you have a elongation and thermal expansion model that correctly predicts the piston-to-head clearance at overlap TDC and then you add the gas pressure to that model at combustion TDC, my guess is that you'll have a pretty good model of how high the squish clearance actually is at compression TDC. Nitro2's point earlier that piston to head clearance is smaller at overlap TDC than at combustion TDC because of gas pressures should then be taken into account when computing things like squish velocity.

[ptuomov]

I concentrate only on the maximum squish velocity to see how big the impulse can be. My 0D-model uses the approach of G.P. Blair, which gives a rough idea what is happening in the chamber and when it happens. Of course a 3D-approach would clarify all the interaction, but this is out of my budget and time.

After spending some time staring at motorcycle pistons and large eddy simulations, here's how I'd design a "good enough for government work" high-compression normally-aspirated piston for an existing 4-valve head:

First, I would take a mold of the head combustion chamber with the spark plug in and digitize it or digitize the combustion chamber directly. This would be the starting point for the piston crown.

Second, I would compress the y-axis scale of the shape, anchored at the squish pad level, a small amount such all the piston crown surfaces above the squish pad y axis level would be angled slightly relative to the head. This would make sure there are no obviously trapped spaces etc. problem areas when the piston rocks. The piston crown top would now have somewhat more clearance at the spark plug level than at the squish pad level (where the clearance would be zero in the model at this point).

Third, I would create a cavity with a minimum clearance from the spark plug (say 5mm) for the case that the tip projects significantly and the required compression is high. If required compression isn't super high, this step will be made redundant by the sixth step.

Fourth, I would lower the piston in the model away from the squish pads in the head by the minimum required (cold) piston to head clearance.

Fifth, I'd take the camshaft profile, valve sizes, and locations, and the assumed margins of safety and would cut the valve reliefs to the model moving the pistons and valves in the model per the camshaft profile. The margins of safety require some rules of thumb and judgment calls.

Sixth, I would take a large-radius sphere with the x and z coordinates of that sphere's origin aligned with the spark plug center (but placing the y coordinate of the sphere origin obviously much higher than the spark plug) and remove all material from the piston crown that makes it inside the sphere. The sphere radius that I'd pick would be something like 10x the bore. I would lower the sphere origin y coordinate (and thus remove more material) until the (cold) geometric compression ratio requirement is met.

Seventh, I'd put a bit of a radius on all the sharp edges created by this geometric modeling strategy. This won't change the compression ratio enough to require repeating any steps.

I don't know if the above attempt to explain the design strategy makes any sense to the readers, but it basically amounts to a starting with a dome that exactly matches the combustion chamber and then cutting a spherical dish to it until the compression ratio requirement is met. If the required compression ratio is stupid high, you only cut a little bit of the crown off around the spark plug. If the required compression ratio is relatively low, you'll cut a lot of the crown out.

The company making the piston can do all of the above relatively easily in a computer model, at least the piston company I use can.

In my opinion, this will get you as close to chicken salad as the you can get with a particularly chicken $hit head you're starting with, or at least reasonably close to an acceptable compromise. The spherical dish on the crown will preserve the tumble as much as it can be preserved, the combustion chamber will be reasonably compact, and you'll get some squish motion near combustion TDC isn't going to hurt with combustion. Is it optimal? Of course not. Is it kind of there, close enough? I think yes.

Thumbs up or thumbs down?

[juuhanaa]

I concentrate only on the maximum squish velocity to see how big the impulse can be. My 0D-model uses the approach of G.P. Blair, which gives a rough idea what is happening in the chamber and when it happens. Of course a 3D-approach would clarify all the interaction, but this is out of my budget and time.

After spending some time staring at motorcycle pistons and large eddy simulations, here's how I'd design a "good enough for government work" high-compression normally-aspirated piston for an existing 4-valve head:

First, I would take a mold of the head combustion chamber with the spark plug in and digitize it or digitize the combustion chamber directly. This would be the starting point for the piston crown.

Second, I would compress the y-axis scale of the shape, anchored at the squish pad level, a small amount such all the piston crown surfaces above the squish pad y axis level would be angled slightly relative to the head. This would make sure there are no obviously trapped spaces etc. problem areas when the piston rocks. The piston crown top would now have somewhat more clearance at the spark plug level than at the squish pad level (where the clearance would be zero in the model at this point).

Third, I would create a cavity with a minimum clearance from the spark plug (say 5mm) for the case that the tip projects significantly and the required compression is high. If required compression isn't super high, this step will be made redundant by the sixth step.

Fourth, I would lower the piston in the model away from the squish pads in the head by the minimum required (cold) piston to head clearance.

Fifth, I'd take the camshaft profile, valve sizes, and locations, and the assumed margins of safety and would cut the valve reliefs to the model moving the pistons and valves in the model per the camshaft profile. The margins of safety require some rules of thumb and judgment calls.

Sixth, I would take a large-radius sphere with the x and z coordinates of that sphere's origin aligned with the spark plug center (but placing the y coordinate of the sphere origin obviously much higher than the spark plug) and remove all material from the piston crown that makes it inside the sphere. The sphere radius that I'd pick would be something like 10x the bore. I would lower the sphere origin y coordinate (and thus remove more material) until the (cold) geometric compression ratio requirement is met.

Seventh, I'd put a bit of a radius on all the sharp edges created by this geometric modeling strategy. This won't change the compression ratio enough to require repeating any steps.

I don't know if the above attempt to explain the design strategy makes any sense to the readers, but it basically amounts to a starting with a dome that exactly matches the combustion chamber and then cutting a spherical dish to it until the compression ratio requirement is met. If the required compression ratio is stupid high, you only cut a little bit of the crown off around the spark plug. If the required compression ratio is relatively low, you'll cut a lot of the crown out.

The company making the piston can do all of the above relatively easily in a computer model, at least the piston company I use can.

In my opinion, this will get you as close to chicken salad as the you can get with a particularly chicken $hit head you're starting with, or at least reasonably close to an acceptable compromise. The spherical dish on the crown will preserve the tumble as much as it can be preserved, the combustion chamber will be reasonably compact, and you'll get some squish motion near combustion TDC isn't going to hurt with combustion. Is it optimal? Of course not. Is it kind of there, close enough? I think yes.

Thumbs up or thumbs down?

What CID, VE and RPM?

[ptuomov]

What CID, VE and RPM?

I think it's a reasonable solution for any normally-aspirated four-valve engine that has a large bore relative to stroke and needs a high compression ratio. Those tend to be relatively high rpm engines. Whether it's a 150cc Suzuki cylinder, 500cc BMW cylinder, or 667cc Porsche cylinder, the same recipe should work ok.

[juuhanaa]

That sounds like a really nice playground for different approaches. Do you have a cylinder pressure indication system to find out more about the heat release over crank angle curve?

_MG_9390.jpg

Yep, like Fiat Tipo has close the same base plate than Alfa 156 No, i only have heard about TFX and i hate to say ive been saving for the competition, and that includes buying tires and things like that + i think thats just waay too big step for me right now.

Thanks for your response on the clearance. How much was fireland height? Do you have an idea what the BMW pistons does have on squish area ratio?

Youre welcome and thank you for interest. I have pics regarding chamber and piston dome volume. I dont know about squish area ratio, but i could take measurements about piston without valve pockets and post a picture, if that adds value for this topic.

My actual 180 hp/liter-NA approach on the Honda K-series need about a 12.5+:1 CR because of the cam duration needed at 10,000 rpm. The dome volume on that low stroke get really high and really tumble killer like. Therefore the squish design is very important as well as the chamber design to get the burn duration as far as possible down in duration numbers. I still need 130+ Nm/liter@10,000 rpm, means friction and combustion are my focuses.

Congratulations, you are a crazy man..

They fill it with the upper intake access port or with both at the same time ?

Renault F3 engine.jpg

Something like that. I dont know about that engine, but i could take extra port volume, better wet flow and bigger dish, because i have experienced how little flow numbers matters, and how two different velocities can achieve similar butt feeling for approx same rpm. I have junk pile of cylinder heads, so i just play with flow bench and proceed from there i think. I need to study how to port pistons too

[juuhanaa]

What CID, VE and RPM?

I think it's a reasonable solution for any normally-aspirated four-valve engine that has a large bore relative to stroke and needs a high compression ratio. Those tend to be relatively high rpm engines. Whether it's a 150cc Suzuki cylinder, 500cc BMW cylinder, or 667cc Porsche cylinder, the same recipe should work ok.

Large bore can be poison to combustion. What is high rpm?

[LotusElise]

If you have a elongation and thermal expansion model that correctly predicts the piston-to-head clearance at overlap TDC and then you add the gas pressure to that model at combustion TDC, my guess is that you'll have a pretty good model of how high the squish clearance actually is at compression TDC. Nitro2's point earlier that piston to head clearance is smaller at overlap TDC than at combustion TDC because of gas pressures should then be taken into account when computing things like squish velocity.

Validation of the model happens on the situation the piston kisses indeed the head, as this is the only time without insitu or very focused body noise sensor measurements get a clear measurement. The system of forces, gas mixture sided vs piston sided, speak out the likelyhood of an validation point at alternation of load TDC and less of ignition TDC. Despite the fact that the force of the cylinder pressure is low compared to the mechanical force at TDC on my example of an NA engine at 10,000 rpm. The effect on compensating a portion of the elongation at ignition TDC is maybe lower then some assume. The force ratio is less then 1/2, thence less then half could be compensated.

[ptuomov]

What CID, VE and RPM?

I think it's a reasonable solution for any normally-aspirated four-valve engine that has a large bore relative to stroke and needs a high compression ratio. Those tend to be relatively high rpm engines. Whether it's a 150cc Suzuki cylinder, 500cc BMW cylinder, or 667cc Porsche cylinder, the same recipe should work ok.

Large bore can be poison to combustion. What is high rpm?

High rpm in this context is whatever gets the intake port velocity to high enough level for tumble to work correctly. So larger the port cross section, higher the rpm at which that piston (and engine) would work well.

[LotusElise]

I don't know if the above attempt to explain the design strategy makes any sense to the readers, but it basically amounts to a starting with a dome that exactly matches the combustion chamber and then cutting a spherical dish to it until the compression ratio requirement is met. If the required compression ratio is stupid high, you only cut a little bit of the crown off around the spark plug. If the required compression ratio is relatively low, you'll cut a lot of the crown out.

Yeap, that approach is what the logic is telling one, once one want to reduce the 0-10 % mass fraction burned time and to stabilize the 50 % mass fraction burned timing at a certain small crank angle bandwidth, like 8°-12°, which would be really great.

The issue are the geometrical constraints as long as amount of cylinder is no field to play (= fixed). For a 10,000 rpm plus engine you need mechanical rigidity, which is achieved by material and lowering the stroke. That comes with compression volume constraints. Here we talk about a residual volume of 0.9 ccm which is free to play if all head volume would be filled up to squish height. From a piston fabrication standpoint and lightweight perspective the piston company would give one a no-go here once you claim < 350 g per piston assembly. The only geometrical way to keep the piston in a rational weight and combustion design is to weld the cylinder head volume, which opens a door of freedom, but also development effort.

Some data:

• Valve volume, cut by TDC clearance demand out of the piston, in total 33.6 ccm

• Compression volume to get CR on spot is 41.6 ccm

• Cylinder head volume of that 51° inclined valve angle head is 50.5 ccm

• aimed BMEP = 16 bar@10,000 rpm, thence VE, MFB duration, FMEP and so on have to be accordingly

That project would be much easier on a wedge head design, but the rigidity, stiffness, precision and capability of the Honda K20 basis is phenomenal. Achieving 370 hp at flywheel below 10,000 rpm on a 2-Liter-I4 engine is something pretty challenging on a NA induction concept, I know no single example. Maybe you guys know one? That not enough, I want to have 1.54 ftlb/cui from 5500-9500 rpm. I've already reached similar with an engine but from 4500-8200 rpm. That increase of 1000 rpm to the right is like a complete new development as nothing can stay beside the basic kinematic.

Here quench does indeed affect power as a proper tumble concept has bad constraints because of the piston dome volume.

[LotusElise]

Yep, like Fiat Tipo has close the same base plate than Alfa 156 No, i only have heard about TFX and i hate to say ive been saving for the competition, and that includes buying tires and things like that + i think thats just waay too big step for me right now.

Ah, thanks.

Youre welcome and thank you for interest. I have pics regarding chamber and piston dome volume. I dont know about squish area ratio, but i could take measurements about piston without valve pockets and post a picture, if that adds value for this topic.

That would definitely be of help.

Something like that. I dont know about that engine, but i could take extra port volume, better wet flow and bigger dish, because i have experienced how little flow numbers matters, and how two different velocities can achieve similar butt feeling for approx same rpm. I have junk pile of cylinder heads, so i just play with flow bench and proceed from there i think. I need to study how to port pistons too

These are steep looking ports. As I reported in the Larry Soft head, VOLVO went from 265 to 325 hp when they lifted the port angle to feed the I5 (2-Liter-NA-engine for BTCC) right. They bypassed the port welding rule by their attempt, a genius and brave strike. They surely added other points, but the major door opener was the inclined valve angle

[ptuomov]

51 degree included valve angle is a bit of a handicap from combustion point of view in a normally aspirated, high rpm engine. It’s not easy to weld the head to fully overcome that, in my opinion.

One question to answer is the true trade off between geometric compression ratio and combustion efficiency. Coming down from 12.5 to say 12 may (or may not) add power in net.

Another question is how much smaller can the piston crown valve reliefs be made if they are optimally cut to a specific cam/valve lift profile.

In terms of piston weight, interesting questions are how high can the piston wrist pin be brought and how narrow the top land can be made (ie, how high up the compression ring can be brought). One good thing about a box or box-in-box piston with a dome is that there will be a lot of room for the conrod small end even if the piston wrist pin is brought high up.

[juuhanaa]

I have lift to valve diameter ratio of .309 and 1.3mm cold squish clearance. Max lift and and piston TDC no contact.

Interesting to see how this response on a dyno to "slight" intake advance. I mean the rally car i have mentioned in other topic.. Its a coctail engine, but round and round we go

Screenshot_20220927-172405.jpg

[LotusElise]

51 degree included valve angle is a bit of a handicap from combustion point of view in a normally aspirated, high rpm engine. It’s not easy to weld the head to fully overcome that, in my opinion.

Yeap, that makes the chamber quite slow, therefore I will measure the cylinder pressures to see how combustion changes. By welding you can regain about 5 ccm, with some improvements on the intake, because inflow can be improved.

One question to answer is the true trade off between geometric compression ratio and combustion efficiency. Coming down from 12.5 to say 12 may (or may not) add power in net.

If the combustion can improved by added 2 ccm design freedom, then yes. I've tried to get the cam duration as short a possible, but the acceleration of the valves at 10,500 rpm need to match the 113 g mass which has to be moved with an spring which doesn't increase FMEP that much.

Another question is how much smaller can the piston crown valve reliefs be made if they are optimally cut to a specific cam/valve lift profile.

See above mentioned approach to short duration. The issue is the engine still likes 6 mm@TDC in the midrange (5500 rpm and up), which is necessary for the 130 Nm/liter from 5500-9500 rpm attempt.

In terms of piston weight, interesting questions are how high can the piston wrist pin be brought and how narrow the top land can be made (ie, how high up the compression ring can be brought). One good thing about a box or box-in-box piston with a dome is that there will be a lot of room for the conrod small end even if the piston wrist pin is brought high up.

A very interesting and complex point as valve relief depth dictates fireland height beside the temperature houshold of the 1st ring while the pin is limited by the oil ring. On a 6 mm@TDC valve lift a 22 mm would be a lower end even when using a 2-ring concept.

Actual my conclusion is still this engine is much depended on squish. Maybe I should open a thread to that as this may extend the quench to power relation but is an awesome discussion which I much appreciate.

[ptuomov]

In terms of piston weight, interesting questions are how high can the piston wrist pin be brought and how narrow the top land can be made (ie, how high up the compression ring can be brought). One good thing about a box or box-in-box piston with a dome is that there will be a lot of room for the conrod small end even if the piston wrist pin is brought high up.

A very interesting and complex point as valve relief depth dictates fireland height beside the temperature houshold of the 1st ring while the pin is limited by the oil ring. On a 6 mm@TDC valve lift a 22 mm would be a lower end even when using a 2-ring concept.

I think you can run the oil ring on top of the piston wrist pin bore. The intake valve relief does pin down the maximum heigh of the top ring, but I think only the top ring and second ring have to be clearly above the piston wrist pin bore. In a box or box-in-box piston, I think you can now machine fairly complex patterns under the piston crown to reduce weight which would not be possible with just a standard sort of unmachined forging. The piston oil squirters can be aimed to cool the specific points of the piston crown underside that are most at risk in the design. My instinct says that you should be able to make your weight with the piston assembly, as long as you can get the flow and combustion geometry to work adequately.