48 replies to this topic

[Pong]

Engine improvements:
- would it be possible to weld the cylinder head to the engine block? It can be done for high pressure equipment inorder to save weigth so why not for F1 engines? It would save both mass and space precisely where it matters most.

[david_martin]

A head is slightly more complex than the end of a pressure vessel. The problems involve in welding a head are enormous -- just think about tolerances, thermal expansion problem, oil and coolant galleries, liners, post weld heat treatment, post weld machining, NDT.... The list of issues is long. And in the end, how much does a set of head studs and a pair of gaskets weigh anyway?

[Greg Locock]

Big problem is the restriction on valve size and orientation that is imposed, if you join the head to the block before machining. I know that various people are looking into uniblock (? did I make that up?) construction again for production engines, for various reasons.

It would be very difficult to make the join via welding, since all the internal galleries would need to seal as well. The scrap rate would be horrific.

[Pong]

I was more thinking about machining the head and the block separately and then weld them together instead of bolting them together.
This is doubtless quite problematic - especially the thermal expansion caused by welding could be a real hassle. But with good temperature control, welding robots and possible electric or laser welding I do not really see why it would not work.

[Greg Locock]

Take the head of a watercooled engine. Look carefully at the top of the block. There are some big holes, with uppy downy things in them, and some smaller holes , with diferent sorts of gloopy stuff in them. Many people spend hours and years making sure that the things in one hole don't get into the other holes. HOW THE HELL ARE YOU GOING TO KEEP THEM ALL SEALED?

[david_martin]

Quote

Originally posted by Pong
I was more thinking about machining the head and the block separately and then weld them together instead of bolting them together.
This is doubtless quite problematic - especially the thermal expansion caused by welding could be a real hassle. But with good temperature control, welding robots and possible electric or laser welding I do not really see why it would not work.

You don't mean you intended to weld the fully assembled head onto the block, do you?
If so, were you planning on depositing weld metal on the inside surfaces of the bore and head? Do you think that might require a spot of machining in-situ? Or were you just planning on not having any sealing between the block and head around the bore in an engine with a compression ratio of about 12:1?

To amplify what has been already pointed out about oil and water galleries, in a conventional arrangement with head studs and a head gasket, sealing is acheived by putting the gasket material and the two mating faces of the coupling under not inconsiderable compressive stresses. How can you possibly acheive the same effect with a weld bead around the edges of the two halves of the coupling?

And as for "possible electric welding", what else did you have in mind? Controlled atmosphere brazing? Friction stir welding, perhaps?

[Pong]

How not to "get things from one hole into the other".....

At present this is accomplished with a seal. This seal is compressed by the bolting the head onto the block. If all goes well the precompression holds up and prevents medium from passing between sub-systems.
With a welded head you would need to apply a load to get the precompression of the seal - then weld from the outside without causing damage to sealing.
Hopefully the precompression would be retained in the weld.

[saudoso]

Quote

Originally posted by Pong
with a welded head you would need to apply a load to get the precompression of the seal - then weld from the outside without causing damage to sealing.
Hopefully the precompression would be retained in the weld.

And if you burn the seal you waste the engine? You'd better cast the head and the block in one piece.

[McGuire]

Quote

Originally posted by Pong
How not to "get things from one hole into the other".....

At present this is accomplished with a seal. This seal is compressed by the bolting the head onto the block. If all goes well the precompression holds up and prevents medium from passing between sub-systems.
With a welded head you would need to apply a load to get the precompression of the seal - then weld from the outside without causing damage to sealing.
Hopefully the precompression would be retained in the weld.

That depends on the rigidity of the castings being joined together. How rigid are they? For the purposes described above, not very. It's better to imagine a block or head casting as a cage or skeleton -- like a sheet metal box -- rather than a solid block of material. After all, its interior volume is mostly air, eh.

You will note there are more than two or four bolts on the corners retaining a cylinder head to a cylinder block. For example, the Chevrolet V8 uses 17 screws per bank -- which works out to five per cylinder. For a four-cylinder deck, it's hard to get by with less than ten (four per cylinder). If you draw yourself a cartoon of the deck you can see how it works out.

It would be easy to apply a compressive load on the assembly and then weld it in place. However, the compressive loads would only be maintained in the area of the welds. You would have to devise a means to weld the interior locations as well...which would easily take as much space as conventional screws.

[hydra]

Before you guys keep giving him any more flak, remember that the Porsche Indy engine of the 70s/80s had its head welded to the block... Granted it didn't set the world on fire, but still...

[McGuire]

It was tried again in the late nineties in CART and F1 engines, and once again found to be not worth the trip.

[david_martin]

Quote

Originally posted by hydra
Before you guys keep giving him any more flak, remember that the Porsche Indy engine of the 70s/80s had its head welded to the block... Granted it didn't set the world on fire, but still...

Its a bit easier when the block is air-cooled and only the head has any coolant (and each head had a close coupled water pump as I remember it). But did the '78 spec 936 cum Interscope Porsche engine really have a welded head?

[desmo]

Quote

Originally posted by McGuire
It was tried again in the late nineties in CART and F1 engines, and once again found to be not worth the trip.

Really? Any examples you can cite, particularly F1? This is news to me.

[GBarclay]

Just to stir the pot a little.

I don't think this would work in F1 or other top series.

However the top competitors in tractor pulling are know to weld the heads to the block. I am assuming (yeah I know) they are running compression ratios far in excess of F1 or CART motors given that most are using diesel motors.

i am fairly sure they are doing it as a means to keep the head and block together. weight savings does not seem to be a high priority in tractor pulling.

It's got to make R&R'ing a motor a lot more hassle.

Don't ask how they deal with the water and oil passages, I have no idea.

Grant

[Pong]

McGuire:

allright! 17 stud bolts per block in a V8. That translates to approx. 20 - 25 in a V10 F1 block? In other words there must be a way to weld the head to the block from the inside out. A serious technical challenge in other words..... Calls for a very high tech small size welding robot.

[McGuire]

Quote

Originally posted by Pong
McGuire:

allright! 17 stud bolts per block in a V8. That translates to approx. 20 - 25 in a V10 F1 block? In other words there must be a way to weld the head to the block from the inside out. A serious technical challenge in other words..... Calls for a very high tech small size welding robot.

It's been done...but roughly speaking, you need a weld station (and access to it) everywhere you would ordinarily place a head bolt. So what is the advantage, other than a slight weight savings from discarding the fasteners? Meanwhile an entire building full of mechanics and machinists will curse you for all the resultant headaches, and their children and grandchildren too probably.

As a point of interest, decades ago there were engines totally built up from cast sleeves, forged or fabricated head caps, and sheet metal or electro-deposited jacketing and enclosures, all welded together (usually one cylinder or pair at a time). Mercedes-Benz for one. They are beautiful to look at, and marvels of the metalworker's art. However, modern casting processes and metallurgy have pretty much erased any technical advantage. The better way to accomplish what you have in mind is probably unit-cast cylinder case and heads, rather than welding them together...but even then I believe the drawbacks would overcome the advantages considering the current state of the art.

...I would not dismiss ease of disassembly/assembly as an important factor in the success of a racing engine. They are creatures of development as much as design...probably moreso. And there is always the KISS principle to keep in mind, plus the various corrolaries of Murphy's Law to deal with. This idea has been tried before and didn't prove out. Are you sure you can do better and if so, why? Every engine program has finite resources. Is this the best place to spend them?

[Greg Locock]

The fundamental issue is that in reality you aren't joining two lumps of metal. You are sealing a myriad of joints in fluid transfer ports, and providing a load path for some big loads. The block itself is largely a sealing device, and a spacer/load transfer path between the head and the mains. If you redesign the engine taking that attitude then you don't end up with a conventional block, you end up with a sandwich, where the cylinder liners are the meat, and the mains and the head are the slices of bread.

[Pong]

it seems this idea goes into the "theoretically possible but in practice not so good idea"

[McGuire]

Quote

Originally posted by Pong
it seems this idea goes into the "theoretically possible but in practice not so good idea"

On the contrary: I would say it's a very good idea (nothing wrong with trying to save weight and space!) but in this case extremely difficult to execute.

[McGuire]

Quote

Originally posted by GBarclay
However the top competitors in tractor pulling are know to weld the heads to the block. I am assuming (yeah I know) they are running compression ratios far in excess of F1 or CART motors given that most are using diesel motors.

i am fairly sure they are doing it as a means to keep the head and block together. weight savings does not seem to be a high priority in tractor pulling.

It's got to make R&R'ing a motor a lot more hassle.

Don't ask how they deal with the water and oil passages, I have no idea.

Grant

These are big diesels running unholy boost levels (and sometimes propane injection). They will run the full complement of head bolts, AND weld the head to the block. The purpose is to keep the head from moving around on the deck surface of the block.

[ZoRG]

I've not heard of a F1 head gasket problem in a long time ... not saying it doesn't happen, however haven't heard a car go out with a head gasket problem ever.

Those Tractor pulling machines don' have water going through them at all... they run dry because they run so short... especially being diesel I don't think they get as hot as a top fuel dragster.

[desmo]

Do F1 engines use head gaskets? I seem to dimly remember being told they don't.

[ZoRG]

Desmo very interresting...

[McGuire]

Quote

Originally posted by desmo
Do F1 engines use head gaskets? I seem to dimly remember being told they don't.

Not necessarily a conventional one-piece head gasket, but they do run metallic fire rings around each cylinder and seals on the oil and coolant passages, which serve the identical purpose.

[rgagne]

Quote

Originally posted by McGuire


Not necessarily a conventional one-piece head gasket, but they do run metallic fire rings around each cylinder and seals on the oil and coolant passages, which serve the identical purpose.

As i have worked at a diesel engine rebuilding facility for years, i have seen this myself. Detroit diesel engines has been using that system for years and years. And we are speaking of compression ratio way above any gasoline engine.

RGagne

[rgagne]

Quote

Originally posted by GBarclay
Just to stir the pot a little.

I don't think this would work in F1 or other top series.

However the top competitors in tractor pulling are know to weld the heads to the block. I am assuming (yeah I know) they are running compression ratios far in excess of F1 or CART motors given that most are using diesel motors.

i am fairly sure they are doing it as a means to keep the head and block together. weight savings does not seem to be a high priority in tractor pulling.

It's got to make R&R'ing a motor a lot more hassle.

Don't ask how they deal with the water and oil passages, I have no idea.

Grant

I have been at a few tractor pulling contest and never seen that. But i assume they "spot weld" the head to the block after it has properly been bolted.

RGagne

[Greg Locock]

"Not necessarily a conventional one-piece head gasket, but they do run metallic fire rings around each cylinder "

Mills rings, or Dykes rings, from memory. Very handy on all aluminium engines in the past, which were too unstable thermally to use a conventional head gasket.

[mudpuppy]

I'm pretty sure that the Porsche 962 and the earlier 911/935 had electron-beam welded heads to the cylinders. This was done to cure head cylinder leaking and warping due to heat.

I need to check my copy of 911 Performance Handbook, which is at home.

Someone else please help with this?

[McGuire]

Quote

Originally posted by Greg Locock
The fundamental issue is that in reality you aren't joining two lumps of metal. You are sealing a myriad of joints in fluid transfer ports, and providing a load path for some big loads. The block itself is largely a sealing device, and a spacer/load transfer path between the head and the mains. If you redesign the engine taking that attitude then you don't end up with a conventional block, you end up with a sandwich, where the cylinder liners are the meat, and the mains and the head are the slices of bread.

That's a very useful way to look at it. Since much of my personal experience involved linerless cast iron cylinder cases, my special obsession was always maintaining the concentricity of the bores. With four screws per cylinder, the bore will tend to be pulled into a square or rectangular shape when the head is torqued down; the Chevy's five bolts per cylinder form a pentagon.

When the cylinder case is used as a stressed member of the chassis it gets more complicated yet. Then it is helpful to think of the block as a big cylinder in torsion. In this regard the Buick V6 was pretty good, because it was so short and was fully skirted below the centerline of the crank. Though it was so adaptable to so many other applications, the SB Chevrolet V8 was never really much good as a stressed member, either in iron or aluminum. (The first thing you need is a massive sump girdle.) On oval tracks especially, it is easy to underestimate the loadings -- both Honda and Toyota were reportedly caught out by it early in their CART programs. The problems can be rather easy to misdiagnose...what sorta looks like oil starvation under prolonged high g-loadings might actually be the main bearings being pulled slightly out of alignment due to the simultaneous cornering forces, eh.

[ZoRG]

On the honda motors, which is a all aluminium motor, they run multi plate steel gaskets, works really well.

[Pong]

Here is what might be a potential solution to the problem of welding the head to the block:
Both the head and the block must be machined with a inclined surface where they would normally fit nicely togeteher. This allows for the plasma weld to actually reach deep enough in order to weld around the cylinder liner. Now the remaining space must be filled with material strong enough to handle the stresses of racing as well as to provide cooling fluid to circulate in the galleries. Here it might be possible to use laser welded powder metal technology. "Rapid Prototyping Machines" already makes use of this technology. I am not claiming this could be done without some serious research & development - but just in theory......

[ZoRG]

I'm just wondering what the benifits are?

[david_martin]

Quote

Originally posted by Pong
Here it might be possible to use laser welded powder metal technology.

Would you care to elaborate on that?

[Pong]

laser welded powder metal technology:
on a surface an extremely thin layer of glue is sprayed in the desired pattern. Thin metal powder is sprinkled on top of this. Laser welds evaporates the glue and welds the metal powder to shape. The process is then repeated until the final form emerges. By using or not using glue on specific spots is is possible to produce things like interlocking chains.

The whole process can take days to reach completion.

[dosco]

Quote

Originally posted by Pong
laser welded powder metal technology:
on a surface an extremely thin layer of glue is sprayed in the desired pattern. Thin metal powder is sprinkled on top of this. Laser welds evaporates the glue and welds the metal powder to shape. The process is then repeated until the final form emerges. By using or not using glue on specific spots is is possible to produce things like interlocking chains.

The whole process can take days to reach completion.

Yes, the parts are "sintered" together.

The part you forgot is to HIP the thing after the laser process (Hot Isostatic Press). That packs the sintered powder together and forms more or less a homogeneous mass.

[dosco]

Quote

Originally posted by Pong
Here is what might be a potential solution to the problem of welding the head to the block:
Both the head and the block must be machined with a inclined surface where they would normally fit nicely togeteher. This allows for the plasma weld to actually reach deep enough in order to weld around the cylinder liner.

I think you'd want to use an e-beam weld in a hard vacuum. E-beam welds can acheive remarkable depth-of-penetration and do have very narrow heat affected zones.

Post-weld cracking would be a huge problems....you'd have to do a lot of research into pre-weld heating....

So what do you get for all this cost and development effort?

[Yelnats]

The solution is to make the head and block continuos and do away with the gasketed interface. It make no sense to construct two separate components if a permanant attachment (the weld) is to be used. The welding of block/head interfaces on modified engines is a stop-gap solution.

The fixed head design as a solution to the gasket problem has lost favour due to improvments in gasket design and the ease of construction and maintenance afforded by removable head designs.

[UStifosi]

I seem to remember that the 1984 Toleman had a Hart 1.5 liter turbo engine what had an integral cylinder head hence it was called a monobloc design.

[D-Type]

Wasn't the classic Offy also a monobloc?

[desmo]

It appears both the classic Offy-Drake and the Hart 415T were indeed monobloc designs.

Offy Drawing

[McGuire]

Quote

Originally posted by D-Type
Wasn't the classic Offy also a monobloc?

Sure was. So was the 1913 Peugeot, the Miller 8, the Duesenberg twincam, the Novi, and the six-cylinder "variants" of the Offy (the Sparks-Thorne and the Lencki). They all cast the head and cylinder block in unit -- it was the standard architecture in the classic era of American racing. However, over here it was not called "monobloc." In the USA, that term was used for cylinders and crank case cast in unit, as with the 1932 Ford V8.

[Can Am]

I have seen a method that the head and cylinder liners were one and was dropped into the block, MAY have been the supercharged GP Benz's of the late 30's.

One of the Hart Turbo's was similar to this too I think.

Grek Locock, think laterally, water and oil do not have to pass thru the head to block joint.

[desmo]

It appears to me that F1 heads are substantially oil-cooled anyway.

[Engineguy]

Quote

Originally posted by desmo
It appears to me that F1 heads are substantially oil-cooled anyway.

I happen to have a Suzuki oil-cooled cylinder head sitting on my desk as I type this, so I'm well aware it can be done, but of course Suzuki discontinued the practice years ago and went to fully water-cooled engines. I have not however, ever heard anything about current F1 engines (or anything else for that matter) having oil-cooled heads. Care to elaborate on your reason for thinking this?

[Greg Locock]

"Grek Locock, think laterally, water and oil do not have to pass thru the head to block joint. "

Way ahead of you if you reread my sandwich post

[desmo]

Quote

Originally posted by Engineguy


I happen to have a Suzuki oil-cooled cylinder head sitting on my desk as I type this, so I'm well aware it can be done, but of course Suzuki discontinued the practice years ago and went to fully water-cooled engines. I have not however, ever heard anything about current F1 engines (or anything else for that matter) having oil-cooled heads. Care to elaborate on your reason for thinking this?

Oil and fuel cooled perhaps would have been better worded. Looking at the small coolant passages in the cross sectional and cutaway drawings from Wright's book, being told that cooling from fuel evaporation in the intakes provides a susbstantial proportion of the cooling on the intake side, to the point where one manufacturer upon testing CF inserts in the IN ports traced a hard to diagnose incipient misfire to insufficient heat transfer to accomplish the fuel/vapor exchange at high rpm, and- again anecdotally- that the EX port areas once out of the immediate vicinity of the combustion chambers are designed to run very hot with no local coolant passages. It looks to me as well that from the tappet blocks up that rather a lot of oil is splashed about, perhaps more than is strictly necessary for purely lubrication purposes. These are just personal observations and surmises, which is why I said "appears". I'd love to see some data on what proportion of cooling is done by each.

The original Suzuki GSX-R design was a neat piece of work. With its extensive fine pitch finning it might be somewhat more accurate to call it primarily air cooled, but then again what engine isn't in the end air cooled?

[J. Edlund]

Quote

Originally posted by desmo
It appears both the classic Offy-Drake and the Hart 415T were indeed monobloc designs.

Offy Drawing

While the Hart 415T was a monoblock design I think the TAG Porsche Turbo used welded liners to the head, the turbocharged F1 engines, because of the high combustion pressures and the risk of knock, had some problems with the cylinder to head sealing.

[Engineguy]

Quote

Originally posted by desmo

The original Suzuki GSX-R design was a neat piece of work. With its extensive fine pitch finning it might be somewhat more accurate to call it primarily air cooled, but then again what engine isn't in the end air cooled?

The "fins" on the head I have are mostly just decorative, just a few millimeters deep over most of their area, and just on the endwalls of the head, where they certainly wouldn't do anything for cylinders 2 and 3. This isn't splash cooling; the head is filled with pressurized oil exactly as a water-cooled head is filled with water. If you can say a water-cooled aluminum head is 92% water-cooled, 6% oil-cooled (cam friction, valve springs and upper valve stems), and 2% air-cooled (outer surface area), I'd say this Suzuki head is 97% oil-cooled. The cylinders are another story.

[xkssFrankOpalka]

Did not Bugatti have head and block together.? I still have scars from that.