SwedeSpeed - Volvo Performance Forum banner

Status
Not open for further replies.
1 - 20 of 125 Posts

·
Registered
Joined
·
157 Posts
Discussion Starter · #1 ·
Block Crack… A technical overview…

Well, I was going to add this to the painfully long block guard vs darton sleeve thread, but that turned into mindless willy-swinging match, so instead… a fresh start.

Hi. I’m new here. I am a mechanical engineer, currently working in the Aerospace sector. Previous to that, I worked at an Automotive OEM in the Midwest in their Engine development program. Originally heavily involved in intake/Port design and CFD work, and later designing in reciprocating parts for DI engines. (Piston, rods, and rings). I will admit I never was directly involved with the design of cylinder blocks, but obviously still involved with that aspect, as they heavily influenced my pistons. (mostly focusing on bore distortion as related to piston ring sealing and piston skirt scuffing, It is surprisingly complex. And guess what? Cylinders are never round! Never ever…They are oval, sometimes kinda triangle-ly, more often they are almost clover leaf shaped, (4th order distortion, on the order of a handful of microns, but it is enough the rings which are much easier to keep round, don’t seal properly.. Months, nay years, of my life was spent fighting the comformability of rings against funky shaped bores…I still have nightmares..…) Anyways, here are some things I learned along the way, some applicable to production engines, some to race engines….

First, the basics. Open deck-vs closed deck. Pictures speak a thousand words…



This is a 4G63 from a turbo mitsu (dsm/GVR4/evo). Note there is a continuous amount of material from the cylinder bore, all the way across the top of the block to the walls. The “deck” is closed. Also, the block is iron, not aluminum. This is very robust and stiff. It’s great for sealing head gaskets, and keeping the cylinders from shifting around under high loads. Much like Chevy’s LS engines, doubling, tripling or quadrupling the power coming out of a mitsu 4G63 is only slightly more difficult that falling off a log. Stock 4G63s can easily handle 400-500hp all day long. Toss some decent forged rods and pistons in one, and 700-800hp is almost reliable enough to daily drive… However that closed deck block, well, It sucks for heat transfer out of the combustion chamber, as it leaves a large part of the hottest portion of the combustion chamber (the very top) not near any coolant.



This is an open deck block. It’s a Honda K24. The tops of the bore are not connected to the walls of the block. This lets coolant surround each bore all the way up to the cylinder head. This is great for cooling and knock toughness, and primarily, these two considerations are why 80% of the engines made today are open decks. OEMs focus on any tiny improvement to BSFC (brake specific fuel consumption) to improve fuel economy, and open decks blocks are a great way to achieve more combustion stability through leaner mixtures and more timing.

Taking that a step further, a lot of open deck blocks will also have a little saw-cut between each bore in the Siamese area. This cut allows coolant between each bore and helps maintain a more uniform temperature and distortion in the top regions of each cylinder. Some newer blocks, like the newest AMG 4 cylinder take this a step further and include a matching little saw cut / slit into the cylinder head. Anything in the name of the almighty god of Fuel Economy, I guess….. Our 2.5L Volvo blocks, only have the saw cut in the Siamese area between each bore. And this little feature is the root cause of a bunch of people acting like dicks on the internet.
Never-the-less… Moving on to more important things....



More Cowbell! When people starting exceeding the original power levels of an open deck block, lots of things can go wrong. I think, most common, is that people start to experience head gasket problems, probably. Open deck blocks are a bit tougher to seal properly than their closed deck counterparts. With higher BMEP from more boost, or higher rpms, the very top of a cylinder bore can start to shift around a little bit. Higher temps can also cause thermal distortion that ruins a head gasket . Lots of detonation or pre-ignition can make those cylinder liners shimmy around like an upside down cow-bell too. That movement ruins a head gasket quicker than Will Ferrell can get naked in the movie Old School…

(Interestingly enough, it appears that before our Volvo 2.5L blocks experiences this phenomenon, they tend to crack at the saw cut area… I’ll get to that in a bit..)


Since the beginning of time, people have tried a lot of modifications to fix moving cylinder liners in an open deck. The first ideas had people “pinning” or “bolting” the liners in place. (I have actually seen this, not making it up!) What they do to limit the movement of the bores is, from the outside of the block, drill and tap a holt about an inch down from the top of the block and thread in a bolt that just barely touches the cylinder sleeve. This gives each cylinder a solid touch point to restrict movement in that direction. A variation of this is to press fit solid dowels (about 20mm long between the cylinder and outside wall of the block (this got popular again with a lot or open deck Subaru blocks somewhat recently..) . If I recall correctly, primarily people do this against the thrust side of the block, and leave the anti-thrust side untouched. It kinda works..but there are better solutions... There is a great risk of inducing unknown stress and strain if this is done poorly. There is a great risk of the dowels loosening and shifting around as well. You could end up distorting the sleeve and creating a worse situation than you started. I do not recommend this method.


Taking things a step further, lots of people have developed “block guards”. As we have already seen, they are a big piece of machined aluminum (I hope Alum and not Steel or something…) that surrounds each bore and fills in the gap going out to the rest of the block. These have been very popular with the Honda crowd, and also some Nissan V6s and I think the Porsche V8 (in the 928) crowd. If well designed, these are great. They solidly anchor the top of each bore against the rest of the block. It doesn’t take much at all to hold those cylinders steady, so if you go this route, just be sure the guard you choose has a lot of open passages to allow coolant all the way up and around the cylinder. If you are repeatedly blowing head gaskets, and not due to the head lifting (use ARP studs for that…) These are a good solution. I would personally avoid a guard that had the entire top surface flush with the block. These designs are why block guards have a reputation of causing “hot spots”. However, with our Volvo blocks, unless that saw cut in the Siamese area between bores is filled in, these may not really prevent a block from cracking in the manner we see on the forums..
This one below seems to have some small coolant passages and a channel to allow coolant to circulate around the combustion chamber.... I would maybe even want some more passages.....


As mentioned, the “nuclear solution” is a fully sleeved solution like the Darton MID sleeves. These reinforce both the entire length of the cylinder and brace the top edge against the block. Theoretically, this should be indestructible and perfect. But they are not. The success in implementing one of these is highly dependent on proper installation. Tight machining tolerances and assembly conditions must be followed to ensure these work. If your shop isn’t heating the block in an oven and freezing the sleeves in dry ice prior to assembly, I’d be worried. And just because your machining shop has a shiny pretty surface finish on their parts, doesn’t mean they know a damn thing about building an engine. Quite frequently, the super awesome indestructible sleeves end up failing after some time. (Go ahead and google search “ engine dropped sleeve”. ) Poorly installed sleeves won’t be held correctly in the block and can actually fall into the block deeper by a few thou. Properly installed sleeves, over a few hundred heat cycles can fall into the block. It doesn’t make sense to me to machine out a factory sleeve only to replace it with one the same sized one but pressed in. I think you end up weakening the block too much to hold the sleeve properly. To be frank, sleeves are best left to dedicated racecars that receive routine tear-downs. I would never daily a car with sleeves. It is just asking for trouble.

To be frank, it is extraordinarily rare that any aftermarket company can achieve the same sleeve adhesion that an OEM can do with a cast in liner. It is extraordinarily rare that an oem block needs the sleeves reinforced to prevent them from rupturing half way down the bore. This solution is overkill and simultaneously potentially can cause a whole new world of issues..…

So… if it was up to me.. What would I do to the Volvo block? Nothing. My S60R is a daily driver with 2 car seats in the back that will probably never get off the stock turbo..

But if I did…….. well, first off, I would attempt to polish down the sharp edge at the bottom of the saw-cut to remove that little stress concentration. (Assuming I could get some sort of grinding wheel down that far or a polishing cord.. Second I would not shim the gap, especially with a steel shim. ( if you are going to shim, at least try to fit a aluminum shim in the gap to maintain some sort of uniformity in thermal expansion between the 2 walls. I would guess the steel shim would move around a bit once everything heat cycled a few times. )

I would bet the safest strategy would be to weld the gap between each bore. Aluminum melts much lower than steel, so I bet you could weld the gap without affecting/distorting the steel liners. The deeper you could fill that gap the better… If you are feeling ambitious, I would even go so far as try to weld the gap in a few areas around the water jacket area. Afterwards, re-hone (with a torque plate) and deck the block, cross your finger and turn up the boost.

[Side note: this exact process, polishing away the stress concentration, and welding the top of the gap is used on a certain manufacturer’s Indy Blocks, with about a 95% success rate in keeping the sleeves from cracking in a similar manner]

Or Instead of that, if it was too difficult to weld, I would combine the block guard and a shim. That ought to be fairly robust too, and maintain the sleeve adhesion to the block without the risk of pressed in sleeves moving about… Even better, would be if the shim area was integrated into the guard.

And now a fair warning. I am terrible sleep deprived and have ingested enough cold and flu meds over the last few days to tranquilize a wildebeest. I could be very wrong in a lot of my assumptions. Take everything I wrote with a grain of salt. I could be completely wrong. I could have made all this up as I wrote it… Also, I really want to see a picture of the crack close up, of the actual cracked surface… that could tell us a lot more about what is happening in our blocks..

So that is my 2 cents on the subject, feel free to call me an idiot in the comments below!

TLDR:

Shims prevent blocks from cracking, but do nothing to prevent the cylinders shifting and blowing head gaskets. (if that is a problem)

Block guards are great to prevent cylinder sleeves from moving around and blowing head gaskets, but only help brace the weak point between bores a small amount.. It isn't the main benefit of installing a guard.

Darton sleeves can fix both, but also host a whole new world of potential problems. They are not as perfect as everyone thinks..
 

·
Registered
Joined
·
157 Posts
Discussion Starter · #2 · (Edited)
But until then, am I full of ****?
I modeled up a quick dummy aluminum block with steel liners, I put in an arbitrary saw cut to compare against a cylinder without one. Then I added a shim (I choose aluminum for materials), and after that, I CAD’d up a quick block guard… and then made an approximation of what polishing the groove and welding it could achieve… and Then I tossed them in some FEA software I should only be using for real work purposes,,,,

Please note: spent 2 minutes refining the mesh. It is sloppy and coarse… I took an arbitrary value for cylinder pressure. It is probably on the high side of what a turbo engine would ever see… I applied the pressure evenly through out the entire length of the bore… (in retrospect, I should have added a piston in the block, and limited the depth the pressure is applied, but what ever… No thermal expansion or material property changes due to temp was included… there are a million other things that could be refined…oh well..)

Use this as an A to B comparison, not as exact numbers. It ought to illuminate the relative effectiveness of each countermeasure, only on the stress seen at the sawcut area. The block guards ability to keep the cylinders from ringing like a cowbell, (which they will do) was not evaluated..

My conclusions:
Countermeasure Sleeve Stress Reduction Block Stress Reduction
Saw cut Base line Base line
Shim -17% -30%
Block guard -11% -4%
Weld /Polished grove -17% -30%
BLockGuard + shim -21% -34%
Control (no cut) -30% -45%

So..

A dummy block with steel liners.


Stress concentrations from the saw cut area
Cut-section through a bore.. stress in the cut is higher on the outsides, stress in the steel liner is higher right in the middle where we see cracks. I'd guess the steel liner cracks before the aluminum block does...



Liners don't like to bend...



block guard with shim... happy


Block guard with open cut... umm not so happy:



For reference, my dirty modeling of a block guard..
 
  • Like
Reactions: Jesse.Skulls

·
Registered
Joined
·
911 Posts
Speechless.

P.S. I want some of that cold medicine.
 

·
Registered
Joined
·
869 Posts
 

·
Registered
Joined
·
157 Posts
Discussion Starter · #5 ·
 

·
Registered
Joined
·
1,605 Posts
Oh man, I actually learned something today! Good post!
 

·
Registered
Joined
·
170 Posts
block guard with shim... happy


Is the shim integral to the block guard?

As part of the thread discussing this there was a block guard designed in this manner.

Block guard with open cut... umm not so happy:


That is showing the stress on cylinder 1, the block guard will be weaker as you go in to cylinder 2,3 and 4. I believe those are the ones that crack.
 

·
Registered
Joined
·
2,367 Posts
Nice write-up with some solid logic. Your FEA confirms my suspicions. I suspect an integrated shim/block guard is no better than having them separate as long as they properly fill all voids. But I'm another automotive powertrain systems engineer so may be talking out my ass as well... and I'm also on cold medicine. ;)
 

·
Registered
Joined
·
8,055 Posts
Nice modelling, but to better represent the max cylinder loading, I would prefer to see you modelling about 1500 psi in only the top half inch of the cylinder along with a side load from the piston about 1.5 inches down in the bore of at least 250 lbs. Once the crank is well past TDC all the good uppper cylinder stresses are way below what they are right after TDC.

May also want to add some downward compression forces on the cylinders too to represent the effect from the head being clamped on as well. Till then we really don't have an accurate representation of what the upper cylinder stresses are, especially due to the stress risers involved.

Besides that I love what you are doing here, if I had the time and a full legit copy of solidworks with the analytical libraries I would love to try this myself.
 

·
Registered
Joined
·
157 Posts
Discussion Starter · #14 ·
Nice modelling, but to better represent the max cylinder loading, I would prefer to see you modelling about 1500 psi in only the top half inch of the cylinder along with a side load from the piston about 1.5 inches down in the bore of at least 250 lbs. Once the crank is well past TDC all the good uppper cylinder stresses are way below what they are right after TDC.

May also want to add some downward compression forces on the cylinders too to represent the effect from the head being clamped on as well. Till then we really don't have an accurate representation of what the upper cylinder stresses are, especially due to the stress risers involved.

Besides that I love what you are doing here, if I had the time and a full legit copy of solidworks with the analytical libraries I would love to try this myself.
I have to go back and check, but I think I was just a tad bit higher than 1500 psi for Pmax. but yeah, I could have spent a lot more time in the FEA set-up. It is far far from what actually happens, but it should at least show the trends people are guessing occurs inside. No thrust from piston skirts included, no thermal effects. I purposely cut out the specific numbers of input loads and total stress numbers to try and cut down on people nitpicking afterwards.. It shows shows some trends...
 

·
Registered
Joined
·
157 Posts
Discussion Starter · #15 · (Edited)
Nice write-up with some solid logic. Your FEA confirms my suspicions. I suspect an integrated shim/block guard is no better than having them separate as long as they properly fill all voids. But I'm another automotive powertrain systems engineer so may be talking out my ass as well... and I'm also on cold medicine. ;)
I agree, the difference between a guard with integrated shims and leaving them separate is probably insignificant....
 

·
Registered
Joined
·
157 Posts
Discussion Starter · #16 ·
/\ Hell yes! Btw is that solidworks? I take it you work at the Everett B*eing facility?
I work very close to there! but I'm not the big B. And not SW, we use a different CAD package. The analysis portion is a stand alone package, Ansys,
 

·
Registered
Joined
·
8,055 Posts
I can live without the thermal stresses, but at the least I would want the max cylinder pressure on only the upper portion above the lower ring at TDC, piston side loading and head clamping forces, since those are the top 3 acting on the area in question. Everything else is just noise in the cyclic stresses that drive the crack.
 

·
Registered
Joined
·
9,226 Posts
Nice read...shim flush with deck or pressed down leaving gap at the top??? :)



also i like the way this pic depicts blockguard fully fills the area to the shimmed portion...that is the way it should be if you deck it...

instead of some just making up theoretical nonsense to justify the flaws some sellers of blockguards made designing it . ;)



 

·
Registered
Joined
·
170 Posts
I agree, the difference between a guard with integrated shims and leaving them separate is probably insignificant....
If you have a block guard with no integrated shims would there not be deflection/deformation in the middle portion of the guard allowing the cylinders to spread and cause the crack.

I'm thinking of concrete forms as an example. Try and brace concrete forms from the outside with no ties between the forms and there will always be some spreading of the forms. Tie them together even with very small pieces of metal and the forms don't spread. Same kind of idea.

Just to round out the discussion, could we get an analysis with just shims?
 

·
Registered
Joined
·
8,055 Posts
If you have a block guard with no integrated shims would there not be deflection/deformation in the middle portion of the guard allowing the cylinders to spread and cause the crack.

I'm thinking of concrete forms as an example. Try and brace concrete forms from the outside with no ties between the forms and there will always be some spreading of the forms. Tie them together even with very small pieces of metal and the forms don't spread. Same kind of idea.

Just to round out the discussion, could we get an analysis with just shims?
Good analogy based on some common sense.

This is the reason I think the blockguard itself is useless since because it doesn't fully loop around the cylinder it doesn't provide any real support to the cylinder. When pressures are over stressing the loop of steel the cylinder represents, some easily distorted aluminum on the backside of that loop can't do anything to stop it.
 
1 - 20 of 125 Posts
Status
Not open for further replies.
Top