My porsche 944 S2 16 valve turbo project
#346
Drifting
Everything I've read advises fitting 944 turbo exhaust valves when using the 2.7 head. I know people have used original S2/968 exhaust valves, but I'd stump up for new uprated exhaust valves if I was building a 16v turbo.
#347
Nordschleife Master
My comments only apply to the four valve head with much smaller exhaust valves.
#348
Drifting
Cosworth specified sodium filled exhaust valves in Ford Sierra Cosworth 16v engines in the 80's. Would Ford have paid for many 1000's of them if they didn't have to?
#349
Nordschleife Master
My thinking is that every engine is different and has different requirements for the exhaust valves. My advice is to figure out those requirements and then meet them with the the valve you choose.
I’m not an engineer, obvious to many at this point, and I can’t claim that I know that solid valves from the right material are better or worse compared to sodium filled multi-piece valves. I think (but don’t know) that either can work, but I suspect that sodium filled valve fails less gracefully when overheated above their spec.
John Kuhn just took apart two turbo 928 engines, one with solid SS Ferrea exhaust valves and another one with stock 928 exhaust valves. Both engines look good, but if anything the stock 928 exhaust valves look better. And we got frustrated by our inability to break that stock engine even bolting the wastegates shut for some sweeps, so it saw some miles and load. The longest continuous test drive with the stock engine and turbos was 1000 miles with refueling breaks only, plus several years of regular driving plus some hundreds of dyno pulls.
Personally, I consider the original stock 1980’s stock exhaust valves great pieces. In my opinion, better than the non-super alloy regular stainless Ferrea valves. I’ve looked at the spare part porsche exhaust valves and they look different — so this is only about the old original valves.
Wider seats and higher seated loads are usually a good idea for exhaust valves when adding a turbo just from heat flow perspective.
Last edited by ptuomov; 12-27-2019 at 03:31 PM.
#350
Rainman
Rennlist Member
Rennlist Member
i can only offer conjecture, but i'd think that exhaust valve material is of less importance when the turbine (and pre-turbine plumbing) are low restriction, getting the hot gas out of there ASAP rather than lingering around the valve.
the sodium valves are as much about warranty durability as they are performance; GM uses them in some of their N/A big V8s (LS7 iirc)
the sodium valves are as much about warranty durability as they are performance; GM uses them in some of their N/A big V8s (LS7 iirc)
#351
Interesting the topic of sodium valves came up.
I was reading up on this very long Cayenne Turbo rebuild thread and there was a comment on the sodium filled valves.
If anyone has time take a read through it. Interesting info and the comments about the valves I had never heard before.
It was suggested that when the sodium gets super hot it turns into a gas rather than liquid making heat transfer worse.
I will do my own research but it caught my curiosity.
Boiling point of sodium is 883 Celsius
Melting point 97.8C
https://rennlist.com/forums/cayenne-...fr-7163-s.html
I was reading up on this very long Cayenne Turbo rebuild thread and there was a comment on the sodium filled valves.
If anyone has time take a read through it. Interesting info and the comments about the valves I had never heard before.
It was suggested that when the sodium gets super hot it turns into a gas rather than liquid making heat transfer worse.
I will do my own research but it caught my curiosity.
Boiling point of sodium is 883 Celsius
Melting point 97.8C
https://rennlist.com/forums/cayenne-...fr-7163-s.html
#352
Nordschleife Master
Interesting the topic of sodium valves came up.
I was reading up on this very long Cayenne Turbo rebuild thread and there was a comment on the sodium filled valves.
If anyone has time take a read through it. Interesting info and the comments about the valves I had never heard before.
It was suggested that when the sodium gets super hot it turns into a gas rather than liquid making heat transfer worse.
I will do my own research but it caught my curiosity.
Boiling point of sodium is 883 Celsius
Melting point 97.8C
https://rennlist.com/forums/cayenne-...fr-7163-s.html
I was reading up on this very long Cayenne Turbo rebuild thread and there was a comment on the sodium filled valves.
If anyone has time take a read through it. Interesting info and the comments about the valves I had never heard before.
It was suggested that when the sodium gets super hot it turns into a gas rather than liquid making heat transfer worse.
I will do my own research but it caught my curiosity.
Boiling point of sodium is 883 Celsius
Melting point 97.8C
https://rennlist.com/forums/cayenne-...fr-7163-s.html
#354
Nordschleife Master
I think that the “interwebs” overemphasize the valve material and under emphasize the cooling geometry when considering turbo exhaust valves.
In another thread and another time and place, someone made a highly perceptive point that, when considering valve seat widths from reliability perspective, one key measure is valve seat contact area divided by valve head area. What do you think is the right number to shoot for there for the intake valve, NA exhaust valve, and turbo exhaust valve? I did some archaeological digging and it turns out that the 1980's water-cooled Porsches had about 15% on the intake, 20% on the NA exhaust, and 25% on the turbo exhaust. Those are just round numbers.
The factory '87 Porsche 928 S4, for example, runs 1.8mm (0.071") seats on 33mm (1.299") exhaust valves normally aspirated. That's about 5.5%. It's a multi-angle seat from the factory, too, with very durable exhaust valve and very hard seat insert. I think (but do not know) that the seat width was set entirely because of cooling needs, the seat and valve themselves could (if kept at design operating temperature) live with much narrower seats.
For turbo versions (of other models, the prototype 928 turbo never was produced), they seemed to go 25% wider on the exhaust seat compared to the normally aspirated version. For a four-valve turbo head, that would mean going from 1.8mm to 2.25mm exhaust seat (or from 5.5% to 6.8% of the valve head diameter). A friend tried different seat widths on the exhaust and the flow on the flow bench didn't go down with wider exhaust seat (the stock exhaust throat diameter in the seat insert ring is about 85% of the valve head diameter). So I'm thinking why not go for reliability and ape the factory designs?
Interestingly, as they go with wider valve seats for the exhaust side for turbo models, they seem to also increase the spring loads on the exhaust side. Ballparking, the seated spring load is increased by about same factor as the valve seat area, making the pressure at the seat about the same for NA and turbo models. Does that make sense, keeping the exhaust valve seat pressure the same? I’m thinking that the seat needs certain amount pressure (load/area) to work well.
Another thing about geometry and exhaust valve cooling is that I think one should leave the valve guide boss in the port and not grind it out to chase more flow. Better the valve is covered by the guide and better the guide is covered by the boss, better the exhaust valve cooling. Found this ancient paper on exhaust valve guide boss shape and size and the exhaust valve temperature:
https://digital.library.unt.edu/ark:/67 ... 1931/m1/8/
Bigger boss can be designed without reducing flow and the bigger boss reduces exhaust valve temperature by 70F degrees. Of course, this might also be dated info or obvious.
In another thread and another time and place, someone made a highly perceptive point that, when considering valve seat widths from reliability perspective, one key measure is valve seat contact area divided by valve head area. What do you think is the right number to shoot for there for the intake valve, NA exhaust valve, and turbo exhaust valve? I did some archaeological digging and it turns out that the 1980's water-cooled Porsches had about 15% on the intake, 20% on the NA exhaust, and 25% on the turbo exhaust. Those are just round numbers.
The factory '87 Porsche 928 S4, for example, runs 1.8mm (0.071") seats on 33mm (1.299") exhaust valves normally aspirated. That's about 5.5%. It's a multi-angle seat from the factory, too, with very durable exhaust valve and very hard seat insert. I think (but do not know) that the seat width was set entirely because of cooling needs, the seat and valve themselves could (if kept at design operating temperature) live with much narrower seats.
For turbo versions (of other models, the prototype 928 turbo never was produced), they seemed to go 25% wider on the exhaust seat compared to the normally aspirated version. For a four-valve turbo head, that would mean going from 1.8mm to 2.25mm exhaust seat (or from 5.5% to 6.8% of the valve head diameter). A friend tried different seat widths on the exhaust and the flow on the flow bench didn't go down with wider exhaust seat (the stock exhaust throat diameter in the seat insert ring is about 85% of the valve head diameter). So I'm thinking why not go for reliability and ape the factory designs?
Interestingly, as they go with wider valve seats for the exhaust side for turbo models, they seem to also increase the spring loads on the exhaust side. Ballparking, the seated spring load is increased by about same factor as the valve seat area, making the pressure at the seat about the same for NA and turbo models. Does that make sense, keeping the exhaust valve seat pressure the same? I’m thinking that the seat needs certain amount pressure (load/area) to work well.
Another thing about geometry and exhaust valve cooling is that I think one should leave the valve guide boss in the port and not grind it out to chase more flow. Better the valve is covered by the guide and better the guide is covered by the boss, better the exhaust valve cooling. Found this ancient paper on exhaust valve guide boss shape and size and the exhaust valve temperature:
https://digital.library.unt.edu/ark:/67 ... 1931/m1/8/
Bigger boss can be designed without reducing flow and the bigger boss reduces exhaust valve temperature by 70F degrees. Of course, this might also be dated info or obvious.
Last edited by ptuomov; 12-28-2019 at 10:29 AM.
#355
Drifting
On the first page of this thread Tim said he had new exhaust valves made, if I was building a 4 cylinder 16v turbo I'd take notice of what the likes of Tim, Patrick and Duke have done.
#356
Nordschleife Master
Still, I am curious, so I have to ask: Is there a long history of failed 4-valve head original exhaust valves in 16v turbo 944s?
#357
Drifting
#358
Nordschleife Master
Is there a pattern of failure of the factory original 4V head exhaust valves in any use? You guys seem to make more power per cylinder whether it’s NA or forced induction, so I am curious if you’ve found the limits of the factory stock original 33mm exhaust valves.
#359
Porsche did some strenuous dyno testing with a 968 engine back in the days and the results suggest than none of the current high hp 4V engines such as Duke's would ever likely drop an exhaust valve.
Last edited by Thom; 12-29-2019 at 11:10 AM.